Chapter 24: Celotex RS5000 insulation

Chapter 24: Celotex RS5000 insulation

Introduction

¶24.1 The insulation used in the cladding system at Grenfell Tower was for the most part Celotex RS5000, a polyisocyanurate (PIR) thermosetting foam product. It was not a material of limited combustibility and was therefore not capable of conforming to the guidance on the construction of external walls contained in paragraph 12.7 of Approved Document B. However, RS5000 was said by Celotex to have been tested in accordance with BS 8414‑2 in May 2014 and to have met the criteria in BR 135, thereby conforming to the guidance contained in paragraph 12.5 of Approved Document B. RS5000 was launched in August 2014. It was marketed by Celotex as "the first PIR board to successfully test to BS 8414",^11491149 {CEL00000012}. and as "acceptable for use in buildings above 18 metres in height".^11501150 {CEL00000013}. In fact, however, the test on which Celotex relied as support for that claim had been rigged by the inclusion at critical points in the construction of the system under test of fire-resistant magnesium oxide boards. Needless to say, Celotex concealed the manipulation of the test in its marketing. Moreover, BS 8414 and BR 135 did not provide for the testing and classification of individual products; they applied only to a complete system, which, if it met the criteria in BR 135, could be regarded as suitable for use in buildings over 18 metres in height. Celotex deliberately tucked that information away in the small print of its marketing literature where it was unlikely to be spotted. This part of the report describes the development and testing of RS5000 and explains what led Celotex to market its new product in a dishonest way.

Background

Key personnel and management

¶24.2 At the relevant time, the management team within Celotex was known as the management action group.^11511151 O'Sullivan {CEL00010027/72-73} pages 72-73, paragraph 217. It was the senior day-to-day decision-making body within Celotex.^11521152 Evans {Day72/64:3-5}. Paul Evans (Celotex's head of marketing from 2013 to 2018,^11531153 Evans {CEL00010058/6} page 6, paragraph 18. and Jonathan Roper's line manager) was a member of the management action group, as were Craig Chambers and Dean O'Sullivan (managing directors of Celotex from 2013 to 2016 and from 2016 respectively). Other management bodies included the product life cycle and planning meeting, whose function was to discuss new ideas for products and projects.^11541154 Warren {CEL00010043/8} page 8, paragraph 29. From May 2013, it was replaced by the Service Product Innovation Group.^11551155 Warren {CEL00010043/8} page 8,paragraph 30.

¶24.3 The dishonest and cynical way in which RS5000 was tested and marketed reflected a culture within Celotex stretching back to at least 2009. There are a number of striking aspects of that history that deserve examination.

120

The acquisition by Saint-Gobain

¶24.4 Celotex was originally owned by AAC Capital Partners, the private equity arm of ABN Amro. In August 2012 it was acquired with effect from 31 December 2015 by BPB UK Ltd, subsequently re-named Saint-Gobain Construction Products UK Ltd.^11561156 O'Sullivan {CEL00010027/3} page 3, paragraph 9. The corporate culture of Celotex before its acquisition had been marked by a drive to create profits and increase the company's share price with a view to a sale,^11571157 Hayes {CEL00010154/6} page 6, paragraph 17; Hayes {Day74/30:9-15}; Evans {Day72/67:20-23}. and that continued to be its focus after the sale to Saint-Gobain.^11581158 Hayes {CEL00010154/6} page 6, paragraph 17; Evans {Day72/68:1-3}. As a consequence of the focus on sales, budgets and the business generally were very lean.^11591159 Evans {CEL00010058/5} page 5, paragraph 15. Following the acquisition, Saint-Gobain set an annual budget for increasing profits, of which at least 15% was expected to come from new products.^11601160 Hayes {CEL00010154/6} page 6, paragraph 17; Roper {Day70/124:11-16}.

Lambda values

¶24.5 As explained elsewhere,^11611161 Chapter 56. the lambda value of a material represents its thermal conductivity expressed in watts per metre kelvin (W/mK). The lower the lambda value, the better the performance of the product as an insulator. The relevant European and national harmonised standard applicable to PIR products is the standard for rigid polyurethane foam building products, BS EN 13165.^11621162 {CEL00001204/10} paragraph 3.1.1. It requires a manufacturer to declare the lambda value of a product as a limit value representing at least 90% of its production with a confidence value of 90%.^11631163 {CEL00001204/14} paragraph 4.2.1. The duty of ensuring conformity with the standard rests on the manufacturer, based on tests carried out at the factory.^11641164 {CEL00001204/24} paragraph 7.1. The standard requires the product to be tested at least once every 24 hours.^11651165 {CEL00001204/25} paragraph 7.3 read with Annex B; {CEL00001204/28}. It may be tested more often, but there is nothing in the standard to suggest that if the product is tested more than once every twenty four hours it is permissible to select from among the results thus obtained.

¶24.6 As well as a culture of maximising profits, there had been a history within Celotex of data manipulation and dissimulation. At least as early as 2009, it had adopted a practice of taking a number of measurements each day of the lambda value of its products and selecting for the purposes of the standards (and ultimately, of course, for the purposes of marketing) the most favourable measurement without declaring the rest.^11661166 {CEL00010498}. There was a drive within Celotex to produce products with a lower declared lambda value, both as a way of differentiating them from other PIR products on the market and as a way of competing with Kingspan in the market for insulation on buildings.^11671167 Hayes {Day74/34:3-24}; Evans {CEL00012233/2} page 2, paragraphs 10-11.

¶24.7 Even within Celotex there was an awareness at the time that the selection of data in that way^11681168 Evans {Day72/71:7-9}. was questionable. An internal memorandum dated 9 December 2009 described the process as involving a high degree of data management and manipulation in order to "avoid an auditor finding evidence that thermals are selected to give [a set lambda value] rather than recording the results regardless of their value".^11691169 {CEL00010272/1}. Not only did Celotex actively seek to conceal that practice from the British Board of Agrément (BBA), its auditors for this purpose,^11701170 Evans {Day72/74:17-21}; {CEL00010456}. but it also appears that the practice was not disclosed to Saint-Gobain during the acquisition and was mentioned only at a management action group meeting in November 2013.^11711171 {CEL00010286/2}.

121

¶24.8 The lambda story is revealing. First, it is indicative of a culture in which breaking the rules and misleading external bodies was tolerated long before the advent of RS5000. It demonstrates an historic inability or unwillingness to challenge the honesty and good faith of company practices.^11721172 Evans {Day72/83:3-4}; Evans {Day72/83:10-14}. Second, it shows the malign influence of the competition with Kingspan: Paul Evans explained that the drive to create a product with an improved lambda value was a response to a similar development by Kingspan.^11731173 Evans {CEL00012233/2} page 2, paragraphs 10-11.

The change of polyol

¶24.9 A further indication of the nature of the culture at Celotex relates to FR5000 itself. FR5000 was made on two manufacturing lines, known as Hipchen and Henneke. Products manufactured on the Hipchen line were reinforced with glass fibre. Those manufactured on the Henneke line were not.^11741174 O'Sullivan {CEL00010027/12} page 12, paragraph 41. The products were manufactured using a polymer resin, which was mixed with a polyol and various additives before the mix was allowed to rise on the production line before being cut into shape.

¶24.10 In August 2012, Celotex changed the polyol used on the Hipchen line from ElastoPIR 501 to ElastoPIR 503. ElastoPIR 501 continued to be used on the Hennecke line.^11751175 {CEL00009889}. No thought appears to have been given to whether the change in formulation might affect the fire performance. Although the BBA expressly warned Celotex that it should tell it about any change in the formulation of the product, Celotex failed to do so,^11761176 {CEL00009516}. thereby withholding potentially material technical data.^11771177 Evans {Day72/93:14-17}.

The development of RS5000

Genesis of the "above 18m project": 2008-2012

¶24.11 Celotex had manufactured and marketed FR5000 for some years before 2012.^11781178 {CEL00000723}. It was not a material of limited combustibility and therefore could not be used in the construction of an external wall of a building over 18 metres in height by anyone who wished to follow the guidance in paragraph 12.7 of Approved Document B. Nor had it been incorporated in a system that had been tested in accordance with BS 8414 and had satisfied the performance criteria in BR 135. That was all well known within Celotex.^11791179 Roper {Day70/139:13-21}; Hayes {Day74/37:7-12}.

¶24.12 In November 2011, FR5000 was tested in accordance with BS 476-6 and BS 467-7 and achieved a Class 0 classification.^11801180 {CEL00000382}. Thereafter, Celotex marketed FR5000 as having Class 0 fire performance "throughout". The reference to Class 0 was no more than a marketing device;^11811181 Hayes {Day74/35:10-17}. for reasons we explain later, Class 0 had come to be understood by many in the UK construction industry as meaning that a product could be used on buildings over 18 metres in height in compliance with the Building Regulations. Kingspan was marketing its phenolic foam products in that way, and achieving Class 0 fire performance was seen as a way of demonstrating that FR5000 was as good as Kingspan's K15.^11821182 Evans {Day72/108:2-5}; Hayes {Day74/34:13}-{Day74/35:17}. The description of it as being Class 0 "throughout"' was intended to enable it to challenge K15, which was being marketed as having a Class 0 core.^11831183 Hayes {Day74/36:1-8}.

122

¶24.13 That was a clear abuse of the regulatory regime. The expression "Class 0 throughout" was meaningless because Class 0 is concerned with the reaction to fire of the material forming the surface of a product. It is not concerned with the underlying material, although a product which is composed entirely of a material of limited combustibility will have a Class 0 surface (see Chapter 5). To market FR5000 as "Class 0 throughout" was not only false (because the product was combustible) but can only have been designed to mislead buyers with a poor understanding of fire classification standards. Regrettably, Celotex clearly knew that many potential customers were ignorant and credulous in relation to matters of fire safety. Its marketing of FR5000 in that way laid the ground for the more fundamental mis‑selling that was to follow with RS5000.

¶24.14 Celotex had considered developing a product for use on buildings over 18 metres in height in 2008, but had decided not to do so because the cost of testing was thought to be too high.^11841184 Warren {CEL00010043/8} page 8, paragraphs 31-32. By 2011, however, it had become aware that it was losing sales as a result of not having a product which could be used on high-rise buildings in accordance with the guidance in Approved Document B.^11851185 Warren {CEL00010043/9} page 9, paragraph 33. On 19 October 2011, Robert Warren (then a technical manager, but from 1 May 2013 head of technical)^11861186 Warren {CEL00010043/2} page 2, paragraph 8. sent a number of colleagues an email he had received from Adrian Friar, a specification manager at Celotex, asking them to track all enquiries it had received for insulation to be used on buildings over 18 metres in height.^11871187 {CEL00008564}.

¶24.15 By August 2012, the tracker was being maintained by Paul Evans.^11881188 {CEL00008570}. There was concern that market share was being lost to competitors and that unless action was taken soon the market for insulation suitable for use on buildings over 18 metres in height would be controlled by other manufacturers, particularly Kingspan.^11891189 {CEL00002892}. Mr Roper, who subsequently led the project to develop what eventually became RS5000, understood that the project was driven by a desire to compete with Kingspan in that section of the market.^11901190 Roper {Day70/120:10-14}.

Allocation to Jonathan Roper: 2012-2013

¶24.16 On 15 December 2012, Paul Evans, then a product manager at Celotex, met Tony Baker of BRE with a view to discussing the requirements for testing PIR insulation boards for use in external walls of buildings over 18 metres in height.^11911191 Evans {Day72/96:14-16}; {CEL00002544}. They discussed both the BS 8414-1 and BS 8414-2 test methods, as well as the fact that the results would have to be classified according to BR 135.^11921192 {CEL00002544}.

123

¶24.17 Paul Evans then assigned the task of carrying forward the project to Jonathan Roper.^11931193 Evans {Day71/117:15-23}. Mr Roper had joined Celotex in May 2012 straight from university,^11941194 Roper {CEL00010052/2} page 2, paragraph 2.1. having graduated with a degree in business management.^11951195 Roper {Day70/110:4-10}. He had no previous experience of the insulation industry and had no relevant scientific qualifications.^11961196 Roper {Day70/110:13-17}. Mr Roper told us that he had been closely managed by Mr Evans,^11971197 Roper {CEL00010052/2-3} pages 2-3, paragraph 2.4. and that he had not done anything or said anything, at least publicly, without Mr Evans' approval.^11981198 Roper {Day70/112:9-11}. Mr Roper's experience, as described to us, was echoed by Jamie Hayes, a colleague of Mr Roper, who said that Mr Roper had run everything past Mr Evans.^11991199 Hayes {Day74/44:6-9}. Mr Evans himself accepted that he had followed what Mr Roper was doing closely.^12001200 Evans {Day72/120:18-25}. Given Mr Roper's relative youth and inexperience, it would be surprising if he had not been subject to a high degree of supervision.

¶24.18 Despite his relative youth and inexperience, however, or perhaps because of it, Mr Roper was given the job of developing an insulation product which could be sold for use on buildings over 18 metres in height. Completion of the project was one of his formal goals for 2013.^12011201 {CEL00002900}. Mr Roper stood to gain a financial benefit from achieving that goal, since gaining accreditation for FR5000 was worth 20% of his potential bonus.^12021202 Evans {Day72/123:23-25}. However, his own understanding was that performance indicators were primarily used to create a focus on key areas for the business and we do not think that he was motivated primarily by the prospect of a bonus.^12031203 Roper {Day70/113:2-8}.

¶24.19 In an introductory briefing,^12041204 Roper {Day70/140:20}-{Day70/141:3}. Paul Evans and Robert Warren told Mr Roper that since PIR foam was not a product of limited combustibility, the only way that Celotex could satisfy the criteria in paragraph 12.5 of Approved Document B was to be tested in a system that met the performance criteria of BR 135 following a BS 8414 test.^12051205 Roper {Day70/140:13-18}. In March 2013, Mr Roper prepared a paper entitled "Above 18m Action Plan",^12061206 {CEL00001340} in which he set out the steps required to obtain accreditation for FR5000 for use on buildings over 18 metres in height. He noted that the records showed that between August 2012 and January 2013 Celotex had lost the opportunity of selling 70,000m² of insulation as the result of not having a product that could meet that requirement.

¶24.20 Mr Roper embarked on a programme of research, reporting the results to Paul Evans. He was assisted by Mr Hayes, a member of the Technical Support team,^12071207 Hayes {Day74/46:19-22}. who did not himself have any technical expertise or qualifications.^12081208 Hayes {Day 74/5:10-14}. As recorded in the Action Plan,^12091209 {CEL00001340/2}. it involved looking at the recently published third edition of BR 135, which he obtained and sent to both Mr Warren and Mr Hayes.^12101210 {CEL00011493}. He noted the criteria and although he did not read the whole document,^12111211 Roper {Day70/131:18}-{Day70/132:3}. he understood that it applied only to complete systems.^12121212 Roper {Day70/143:6-14}. He was also aware that any classification obtained under BR 135 applied only to the system tested.^12131213 {CEL00000584/33}; Roper {Day70/145:5-25}. As part of his research he and Mr Hayes also looked at Kingspan's publicly available marketing material.^12141214 Hayes {CEL00010154//12} page 12, paragraphs 36-37.

124

June 2013: business case

¶24.21 On 17 June 2013^12151215 {CEL00001856}. Mr Roper sent Mr Evans a project plan^12161216 {CEL00001857}. and business case^12171217 {CEL00001858}. for Celotex to undertake BS 8414 testing. The business case reiterated the point made in the earlier Action Plan that the lack of a product that was suitable for use on buildings over 18 metres in height was one of the main areas in which Celotex was unable to compete with Kingspan.

Celotex's investigations during 2013

¶24.22 In order to improve his understanding of testing under BS 8414 and classification in accordance with BR 135, Mr Roper obtained assistance from a number of external bodies. Some (NHBC, the BBA and BRE) were introduced to him by Mr Evans;^12181218 Roper {Day70/132:17-23}. others he found for himself.

Sotech and IFC: June to October 2013

¶24.23 Celotex sought advice on how to design a test rig in order to give it the best chance of a favourable result.^12191219 Roper {Day71/15:17-19}. On 10 May 2013 Mr Roper contacted a firm of consultants, International Fire Consultants, seeking advice and explaining that Celotex wanted to design a wall incorporating FR5000 which would have the best chance of succeeding in a full-scale test.^12201220 {CEL00000589}.

¶24.24 Mr Roper also sought assistance from Sotech Ltd, a designer and manufacturer of cladding with previous experience of the BS 8414 test that he had come across at a RIBA event in June 2013. By that time Sotech had already been involved in two BS 8414 tests using Kingspan insulation, each of which, Mr Roper was told, had failed.^12211221 {CEL00001851}. On 20 June 2013, Mr Roper and Mr Hayes met Sotech to discuss testing. A note of the meeting dated 24 June 2013 records that Sotech's previous experience of testing had shown that aluminium railing systems and cladding panels melted and allowed fire to enter the cavity, and that therefore what had proved crucial was the resistance to fire of the outer surface and the performance of the fire barriers.^12221222 {CEL00001863/1}. Mr Roper said that he had understood from the meeting that the success of a BS 8414 test may depend on the cladding material used.^12231223 Roper {Day71/5:2-4}. They also discussed K15 and Sotech expressed surprise that K15 was used so widely based only on a single test involving the use of a cement particle board to represent a typical cladding panel.^12241224 {CEL00001863/1}. Both Celotex and Sotech considered the Kingspan test to be unrepresentative,^12251225 Roper {Day71/7:2-10}. and thought that Kingspan had taken advantage of what they regarded as a carefully worded certificate from the BBA and cleverly designed marketing literature.^12261226 {CEL00001863/1}.

125

¶24.25 Mr Roper sent his note to Mr Evans on 24 June 2013.^12271227 {CEL00001865}. In response Mr Evans indicated that his preference at that stage was to test a commonly used cladding panel rather than simply copying Kingspan's use of a cement particle board.^12281228 {CEL00001865}. International Fire Consultants advised that it should be possible to test with a standard cladding panel using improved fire barriers.^12291229 {CEL00001864}.

¶24.26 Mr Roper then arranged a meeting with Peter Jackman, the technical director of International Fire Consultants, on 22 July 2013.^12301230 {CEL00000632}. Mr Roper could not recall what had been said at that meeting,^12311231 Roper {Day71/16:21-23}. but according to the handwritten notes taken by Robert Warren he was told that Approved Document B only contains guidance and was open to "interpretation".^12321232 {CEL00002743}. We understand that to mean that Mr Jackman thought there were ways of interpreting Approved Document B that would render the guidance less restrictive. For his part, Jamie Hayes recalled a meeting with International Fire Consultants (specifically Peter Jackman and Dr Parina Patel), although he could not recall when it had taken place. He did remember that they had told him and Mr Roper that it might be possible to obtain a field of application report, in other words, to carry out a BS 8414 test and, using the results obtained from it, produce a report which was wider in its scope of application and cover the use of some components other than those used in the test rig.^12331233 Hayes {Day74/72:13-17}. That reflects the view of some in the industry at the time that "desktop" assessments based on test data were a permissible way of broadening the range of components that could be used in external walls while satisfying Approved Document B.^12341234 See Chapter 26, and the note of the meeting on 15 November 2013 between Wintech and NHBC at {NHB00000604}.

¶24.27 Celotex then commissioned International Fire Consultants to conduct small-scale testing with a view to giving an opinion on the likelihood of meeting the criteria in BR 135 if they were to test FR5000 under BS 8414.^12351235 {CEL00000669}; {CEL00000670}. On 4 September 2013, Mr Jackman confirmed that he had reviewed BS 476-6 and BS 476-7 test data for FR5000 and concluded that there were grounds for thinking that FR5000 would perform well in a BS 8414 test.^12361236 {CEL00000673}. It is unclear on what basis he thought that it was possible to draw that conclusion from BS 476 data, but Celotex did not explore that question with him. On 24 September 2013 Celotex engaged International Fire Consultants to assist in the design of BS 8414 tests, having decided by that stage to test only to BS 8414-2 (i.e. a steel-framed construction).^12371237 {CEL00000679}. On 27 September 2013 David Cooper of International Fire Consultants acknowledged the proposal, commenting that the "fire barrier" at storey levels and around the opening would be critical.^12381238 {CEL00000688}.

126

¶24.28 On 3 October 2013, Mr Roper and Mr Hayes met representatives of International Fire Consultants and Sotech.^12391239 {CEL00011052}. Mr Hayes' note of that meeting referred to tests using Kingspan insulation having failed twice with standard cavity barriers.^12401240 {CEL00001195}. Mr Hayes also recalled (although it was not mentioned in the note) that there had been discussion of those tests.^12411241 Hayes {Day74/80:18-23}. Mr Hayes and Mr Roper both told us that following the meeting they had both been aware that, if tested with an aluminium panel, FR5000 was unlikely to pass a BS 8414 test.^12421242 Roper {Day70/194:2-3}; Hayes {Day74/80:24}-{Day74/81:2}.

¶24.29 On 31 October 2013 Mr Roper sent International Fire Consultants details of two European Class A2 products he had found during his researches, namely Alucobond A2 and Marley Eternit.^12431243 {CEL00000714}. He explained that he saw three options for testing: a. test with an ACM panel and an improved cavity barrier system (e.g. with the double

cavity barrier discussed at the meeting on 3 October),^12441244 Roper {Day71/24:12-16}. although he was not particularly confident in that system following the meeting with International Fire Consultants and Sotech;^12451245 Roper {Day71/24:9-10}.

¶b. test with an A2 panel and standard cavity barrier, which he noted was a common system which Celotex could support if challenged; and

¶c. test using a cement particle board as the rainscreen, as he understood Kingspan had done, which he viewed as a last resort

¶24.30 Mr Roper's preference was to test using an A2 panel with standard cavity barriers, but he observed that if this failed then it would be necessary to decide whether Celotex would then want to adopt Kingspan's method or pull out altogether.^12461246 {CEL00000714}. What he meant by adopting Kingspan's method can be seen from what Celotex chose to do later.

Discussions with BRE: May and October 2013

¶24.31 In May 2013, Jonathan Roper asked Tony Baker of BRE which of the BS 8414 tests Celotex should be considering.^12471247 {CEL00000585/2}. Mr Baker advised testing under BS 8414-2, adding that obtaining the classification was likely to be commercially advantageous to Celotex.^12481248 {CEL00000585/2}. Mr Roper contacted Mr Baker again on 21 June 2013, asking to arrange a call with him about the purpose of testing. He said that he was sceptical of how a particular competitor of Celotex (clearly a reference to Kingspan) could have satisfied the criteria in BR 135.^12491249 {CEL00000617}. Mr Roper went on to ask Mr Baker for advice about designing the test and whether preventing the flame from entering the cavity altogether would be acceptable,^12501250 {CEL00000617}. but he could not recall having received any response to his message.^12511251 Roper {Day70/170:19-21}.

127

¶24.32 On 18 October 2014, Mr Roper asked Stephen Howard of BRE about the possibility of extending the use of FR5000 to external wall systems similar to one that had been tested successfully under BS 8414.^12521252 {CEL00000708/3}. He again mentioned his concern about how the market was operating, asking how it was possible for competitors who had tested one particular system to sell its products into a variety of different systems used on buildings over 18 metres in height.^12531253 {CEL00000708/3}. Mr Roper confirmed to us that he was referring to Kingspan and Xtratherm (the only other company that offered a product for use on buildings over 18 metres in height at that time).^12541254 Roper {Day70/174:22-25}. He told us that the origin of his question lay in a controversy within Celotex between individuals, some of whom viewed BS 8414 as a system test and some of whom viewed it as what he called a "product in application" test.^12551255 Roper {Day70/171:16}-{Day70/172:2}. He put himself, Jamie Hayes and Robert Warren in the former category, and Paul Evans, Craig Chambers and the sales team in the latter,^12561256 Roper {Day70/172:13-25}. which he saw as a commercial view based on Kingspan's approach.^12571257 Roper {Day70/163:21-24}. The "commercial view" was not a tenable reading of the guidance in Approved Document B, not least since the third edition of BR 135 published earlier in 2013 had made it plain that the classification applied only to the system as tested and described in the classification report.^12581258 {BRE00005555/28} at A2, left column and {BRE00005555/32} at B2 right column. It followed that the classification could not apply to a single product forming part of the system tested when it was used in a different system.

¶24.33 Mr Howard responded a few days later on 21 October 2013, saying that the test report would set out exactly what had been tested, and that the way to extend it to other systems was by what he called an "assessment report", which he described as "fairly standard practice".^12591259 {CEL00000708/2}. Mr Roper understood him to mean that if he gave BRE details of a particular system they would be able to tell him whether it might pass a BS 8414 test.^12601260 Roper {Day70/176:8-16}.

The Roper-Evans exchange: 1 November 2013

¶24.34 At the end of October and early November 2013, there was an important internal exchange between Jonathan Roper and Paul Evans. On 31 October 2013, Mr Roper sent a message to Mr Evans saying he needed to speak to him about the "Above 18m" project, and that he would send an invitation to Mr Evans, Mr Warren, Mr Hayes and Mr Chambers because they had reached a point at which a decision had to be made.^12611261 {CEL00000716/2}.

¶24.35 The following day, 1 November 2013, Mr Roper sent another message to Mr Evans setting out the matters for discussion.^12621262 {CEL00000716/1}. He explained that they had originally contemplated testing what they regarded as a "worst case scenario with improved fire barrier" (i.e. using a less fire-resistant panel) that could then be supported by an assessment report which would extend the scope of application. However, he no longer thought that could be done because it was not possible, as he put it, "to find or design a suitable barrier in which we have enough confidence that it can be used behind a standard ACM panel which we know will melt and allow fire into the cavity." His comments about how an ACM panel would perform in a full-scale test tells us much about how well known its poor fire performance was at the time, at least among manufacturers of insulation.

128

¶24.36 In his email Mr Roper made a number of important observations about the poor level of understanding in the market of the guidance in Approved Document B and how Kingspan had taken advantage of it. He doubted that Kingspan had a report which would support the use of K15 on buildings over 18 metres in height. His view was that it was being used in those circumstances because of widespread ignorance in the market, including among architects, about the applicable requirements^12631263 Roper {Day71/35:18-24}. and a failure to think about fire performance as opposed to thermal performance.^12641264 Roper {Day71/36:5-6}; {Day71/36:22}-{Day71/37:1}. He thought that the only person who might possibly challenge a product's use on buildings over 18 metres in height was the building control officer^12651265 {CEL00000716/1}. and that Kingspan had effectively neutralised that risk through the wording of their BBA and Local Authority Building Control (LABC) certificates.

¶24.37 The timing of his observations is uncanny. They were made exactly at the time that Kingspan was using a BBA certificate revised in July 2013 to claim that K15 complied with paragraph 12.7 of Approved Document B (i.e. that it was a material of limited combustibility), a statement that was not only self-evidently incorrect but which was causing concern at NHBC.^12661266 See Chapter 26. It seems that Mr Roper had tapped into a growing awareness in the cladding industry that Kingspan was taking advantage of an erroneous BBA certificate for K15, thereby mis-selling it.

¶24.38 Mr Roper concluded his report to Mr Evans with words that were as prescient as they were frank: "Trying to do the right thing requires a complete re-education of the market and this would require a huge campaign and probably a lawsuit". He told us that by "doing the right thing" he had meant selling a product that had been part of a system that had met the performance criteria in BR 135 for use only in a system which was the same as that tested, and not exploiting market ignorance in a commercially more advantageous way.^12671267 Roper {Day71/37:24}-{Day71/38:5}. He explained that the reason he considered re-education of the market would be necessary was because he considered it to have been distorted by Kingspan. Doing things in the right way would mean explaining to specifiers that they had been using Kingspan K15 inappropriately.^12681268 Roper {Day71/38:10-19}.

¶24.39 His report to Mr Evans then proceeded to set out various options: testing with an A2 panel, adopting Kingspan's device of using a cement particle board, or taking the view that, as he put it, "realistically our product shouldn't be used behind most cladding panels because in the event of a fire it will burn".

¶24.40 Mr Evans responded the same day 1 November 2013. His words are, again, worth quoting: "Great summary and shows the real merit of good research and talking to the market. We are trying not to create a 'me too' here but if we do, it will be for the right reasons".^12691269 {CEL00000718}. Mr Evans was unable to recall what he had meant by "the right reasons".^12701270 Evans {Day72/166:11-15}. Mr Roper said that he had understood "the right reasons" to be commercial reasons.^12711271 Roper {Day71/44:5-15}. We accept his evidence and agree that that is indeed what Mr Evans meant. Mr Evans asked Mr Roper to put the contents of his email into some slides for discussion at the meeting scheduled for 4 November 2013.

129

¶24.41 It is clear that by the end of 2013, Celotex had begun to come to terms with the idea that it might follow Kingspan's lead in relation to the testing and marketing of an insulation product for use on buildings over 18 metres in height. Mr Roper's preference was still to test a system incorporating an A2 rainscreen panel rather than a cement particle board. He understood that to reflect a preference within Celotex as a whole to test a more representative system.^12721272 Roper {Day71/25:4-7}. However, the use of a cement particle board had not been entirely ruled out.^12731273 Roper {Day71/11:4-16}. There remained a general desire within Celotex to enter the market for buildings over 18 metres in height in any event and if copying Kingspan was the only way of obtaining a test result that would enable it to do so, it would act accordingly.^12741274 Roper {Day71/25:20-25}. In the end, Celotex did indeed adopt the same strategy as Kingspan. It neither tested a representative system nor was truthful in its marketing literature.

4 November 2013 meeting – the 5 options

¶24.42 On 4 November 2013, Mr Roper met Mr Evans, Mr Hayes and Mr Warren. The electronic invitation suggests that Craig Chambers, then managing director of Celotex, was also invited.^12751275 {CEL00003011}. Mr Roper,^12761276 Roper {Day71/52:15-23}. Mr Roper's witness statement to the effect that he himself had not attended this meeting was corrected by him in oral evidence. Mr Hayes^12771277 Hayes {Day74/46:11-16}; {Day74/96:1-5}. and Mr Evans^12781278 Evans {Day72/170:6-10}. all recalled his being there and we are satisfied that he was.

¶24.43 Mr Roper prepared slides for the meeting.^12791279 {CEL00011199}. The first slide set out paragraph 12.5 of Approved Document B highlighted in yellow. No one who saw it can have been under any illusion about what the guidance was or what it required.^12801280 Roper {Day71/47:5-9}.

¶24.44 Slide 14 set out the options available to Celotex.^12811281 {CEL00011199/14}. The first was "Worst case scenario with field of application report", which referred to testing a system that would enable a desktop assessment to be produced to allow FR5000 to be used in other systems.^12821282 Evans {Day72/172:12-21}. Mr Roper envisaged that it would incorporate ACM rainscreen panels and improved cavity barriers,^12831283 {CEL00000718} and see paragraph 23.29 above. but he had little confidence that he could find such a product.

¶24.45 The second was "System Route (Limits Scope – Requires Re-education)", which referred to testing a representative system and making it clear that any BR 135 classification applied only to the system tested.^12841284 Roper {Day71/49:6-10}. The "re-education" mentioned referred to the need to explain to the market that the industry had up to that point misunderstood the effect of full-scale testing and classification in accordance with BS 8414 and BR 135 and had wrongly allowed K15 to be used in a wider set of applications than was justified by the tests.^12851285 Evans {Day72/173:12-18}.

¶24.46 The third and fourth options were to test and launch FR5000 with, and without, certification by the BBA and LABC.

130

¶24.47 The final option was to abandon the market for insulation suitable for use on buildings over 18 metres in height altogether. It was not discussed at any length.^12861286 Evans {Day72/174:9-11}.

¶24.48 The next slide assessed the prospects of successfully testing FR5000 with various combinations of other products. A system incorporating an ACM panel and an improved cavity barrier was judged to have less than a 50% chance of success, a system incorporating an A2 rainscreen panel with a standard cavity barrier an 80% chance of success and a system incorporating a cement particle board rainscreen with a standard cavity barrier a 90% chance of success.^12871287 {CEL00011199/15}.

The decision to proceed

¶24.49 At the meeting a decision was taken to proceed with BS 8414 testing^12881288 Hayes {Day74/72:15-22}. using an A2 rainscreen panel and then to apply for LABC approval.^12891289 Roper {Day71/54:17-20}. Mr Roper said that there had been a difference of opinion between Mr Evans and Mr Chambers on the one hand and Mr Warren on the other over what should be tested. Both Mr Roper^12901290 Roper {CEL00010052/10} page 10, paragraph 5.15; Roper {Day71/54:10-15}. and Mr Hayes^12911291 Hayes {Day74/100:6-12}. recalled that Mr Warren had thought that Celotex should adopt the "systems route" or opt out, while Mr Chambers and Mr Evans were concerned that that approach would unduly limit the commercial application of the product. Mr Evans denied that that had been his view and told us that, if the product were to be taken to market, Celotex should be clear that it had tested a particular system.^12921292 Evans {Day72/177:20-22}. He accepted that people might want to make use of the results to demonstrate that the product could be used in different systems, but did not believe that his and Mr Chambers' view was entirely different from that of Mr Warren.^12931293 Evans {Day72/177:23}-{Day72/178:2}. We have doubts about the reliability of Mr Evans' evidence on this point because, in the end, we know that Celotex chose not to be clear about what it had tested. On the contrary, it deliberately concealed from end users a vital component of the system tested, namely the magnesium oxide boards, a decision which we examine below. Mr Chambers did not recall the meeting, but in his written evidence he said that he understood that an insulation material had to be either non-combustible or pass a BS 8414 test in order to achieve compliance with the Building Regulations.^12941294 Chambers {CEL00010056/4} page 4, paragraph 20.

Designing the test

¶24.50 On 30 October 2013 Mr Roper was introduced by a colleague at Celotex to Graham Smith of Simco External Framing Solutions Ltd (Simco), a company specialising in the production of external framing systems.^12951295 Roper {CEL00010052/8} page 8, paragraph 5.7. Simco was engaged by Celotex to help design the rig for the BS 8414 test.^12961296 Roper {Day70/190:4-9}.

¶24.51 Following the meeting on 4 November 2013, Celotex decided to use a Marley Eternit panel classed A2. Mr Roper explained that Celotex had chosen the Marley panel because Simco had suggested that it was similar to a cement particle board, although providing a decorative facade, and because Marley was a relatively well-established brand.^12971297 Roper {Day71/67:9-13}.

131

¶24.52 Simco produced drawings for the test rig on 22 January 2014.^12981298 {CEL00000806}; {CEL00000807}; {CEL00000808}; {CEL00000809}. The system to be tested comprised a steel frame, a 12mm magnesium oxide sheathing board, 100mm Celotex FR5000, Lamatherm fire barriers and 8mm Marley Eternit cladding. The components appear to have been delivered to BRE's testing centre at Watford. BRE was unable to produce a complete test file for the first Celotex test, but delivery notes for some of the components were found in the file relating to the second test carried out in May 2014, to which we return below.^12991299 {BRE00032372/4} delivery note for 12mm magnesium oxide sheathing board used in first test; {BRE00032372/5} delivery note for FR5100 used in first test.

The first test: 14 February 2014

¶24.53 The test was carried out on 14 February 2014. Mr Roper, Mr Hayes, Mr Warren and two representatives of International Fire Consultants, Dr Parina Patel and David Cooper, attended.^13001300 Roper {Day71/68:18}-{Day71/69:6}. It was overseen by Philip Clark of BRE. The test was stopped after 26 minutes when flames overtopped the rig,^13011301 Roper {Day71/69:7-9}. but Mr Clark allowed the rig to continue burning for a further period to enable Celotex to obtain more data.^13021302 {INQ00014136/3}. BRE subsequently confirmed that the test had failed and that no classification report would be issued.^13031303 Hayes {CEL00010154/18} page 18, paragraph 56.

¶24.54 Mr Roper said that Mr Clark had commented favourably on the performance of the system, although he understood that it had failed the test.^13041304 Roper {Day71/77:1-3}. In particular, he recalled Mr Clark's saying that he had "seen worse fails" and suggesting that Celotex might want to strengthen the outside of the rig in order to prevent the Marley Eternit panels cracking.^13051305 Roper {CEL00010052/12} page 12, paragraph 5.29. Mr Roper said that he had suggested that the panels were available in 12mm thickness and that Mr Clark had said that he thought that might suffice.^13061306 Roper {Day71/71:17-24}. Mr Roper thought that (possibly on a later occasion)^13071307 Roper {Day71/74:16-20}. he had discussed with Mr Clark the possibility of using two rainscreen panels at the level of the cavity barrier to prevent cracking and allowing the fire to bypass the barrier.^13081308 Roper {CEL00010052/13} page 13, paragraph 5.31. He was unable to remember the details of the conversation, however.^13091309 Roper {Day71/74:6-11}. Mr Clark denied that he had made any comments either about the performance of the system or possible changes to it.^13101310 Clark {BRE00005768/50} page 50, paragraphs 203-204; Clark {BRE00005768/66} page 66, paragraph 269.

¶24.55 Mr Roper was an exceptionally candid witness, whose recollection we generally consider to have been reliable. Moreover, a recording made by the head camera worn by Mr Clark during the test reveals that he did say that he had "seen worse"^13111311 {INQ00014137/2}. and that "for a first attempt it's not bad, actually".^13121312 {INQ00014137/3}. Mr Clark then turned off the camera. No recording was disclosed of any conversation in which Mr Clark made recommendations about the design of any future test rig,^13131313 Clark {Day95/101:8}-{Day95/102:15}; {Day95/104:7-16}. but the recording that was disclosed supports that part of Mr Roper's evidence. We therefore accept that Mr Clark did discuss possible changes to the design of the second test rig with Mr Roper, but there is no evidence to suggest that those discussions touched on the use of magnesium oxide boards.

132

¶24.56 This episode caused us to doubt the reliability of Mr Clark's evidence where it was not supported by the contemporaneous documents. For example, after the test Mr Roper sent Parina Patel and David Cooper of International Fire Consultants thermocouple data from the test,^13141314 {CEL00000843}. saying that he had spoken to Mr Clark, who saw no reason why a classification report could not be issued as, in his opinion, extending the test to the full duration of 30 mins would have made little difference.^13151315 {CEL00000842}. Mr Clark strongly denied that any such discussion had taken place, to the point of accusing Mr Roper of having invented it at the time,^13161316 Clark {Day95/120:4}-{Day95/121:25}. but there was no reason for Mr Roper to have made up the conversation for the benefit of International Fire Consultants and no plausible explanation of how he had obtained the thermocouple data if not from Mr Clark. We prefer the evidence of Mr Roper.

Reaction to the failed first test

¶24.57 Mr Roper considered that the failure of the first test demonstrated the impossibility of meeting the requirements of BR 135 with a representative system^13171317 Roper {Day71/43:6-12}. and in view of the cost of the first test, he did not expect to be given an opportunity to run a second.^13181318 Roper {Day71/78:18-23}. By contrast, Mr Evans said that although it had been a disappointment, he had not considered it to have been a significant problem for Celotex.^13191319 Evans {Day72/182:10-14}. In the light of the importance of the project and the fact that further testing required budgetary approval,^13201320 Roper {Day71/82:6-22}. we do not think that can have been correct. Without passing a BS 8414 test and meeting the BR 135 criteria Celotex could not break into the market for insulation suitable for use on high-rise buildings, and there was at that time no guarantee that further testing would be possible. In reality, the failure of the test in February 2014 was a significant setback.

Decision to retest

¶24.58 After the test, there were further discussions between Mr Roper, Mr Hayes, Mr Evans and Mr Warren.^13211321 Roper {Day71/80:7-11}. Mr Evans could not recall them but did not dispute that they had taken place.^13221322 Evans {Day72/190:12-19}. Mr Mahoney (then Celotex's Head of Research and Development) was also consulted, as he had responsibility for the development budget from which a second test would be funded.^13231323 Roper {Day71/82:14-24}. Mr Evans indicated that the cost was such that it would also have required the approval of Craig Chambers, who was in any case interested in progress, given that it involved the first launch of a (supposedly) new product since the company's acquisition by Saint-Gobain.^13241324 Evans {CEL00010058/29} page 29, paragraph 102.

Design changes

¶24.59 The discussions centred on how to ensure that any further test could be designed to be successful. Mr Roper said that two main changes had been proposed: the use of thicker rainscreen panels (12mm in place of 8mm) and the use of a 6mm magnesium oxide board in the area of the level 2 thermocouples.^13251325 Roper {Day71/81:14-19}. Magnesium oxide is a non-combustible material and the purpose of including it was to reduce the chance that the temperature at that level would exceed 600°C, thereby failing to meet the performance criteria in BR 135.^13261326 Roper {Day71/97:7-12}. It was also suggested that a layer of 6mm magnesium oxide board should be used at the top of the rig to provide a firm surface against which the cavity barriers could expand to prevent flames from overtopping the rig.^13271327 Roper {Day71/97:13-17}. The 6mm magnesium oxide board was to be covered with 8mm Marley Eternit, rather than the 12mm used elsewhere, in order to ensure as much continuity of the rainscreen surface as possible.^13281328 Hayes {CEL00010154//20} page 20, paragraph 61. Celotex also decided to reduce the ventilation gaps between the rainscreen panels.^13291329 Roper {Day72/2:8-9}. Mr Roper believed there were some gaps but he accepted that they were minimal and less than 10mm.^13301330 Roper {Day72/3:10-15}. Taken collectively, those decisions reflected a conscious decision by Celotex to manipulate the design of the rig in order to ensure as far as possible that the performance criteria in BR 135 were met.^13311331 Roper {Day71/98:2-3}.

133

¶24.60 The proposal to use a 6mm magnesium oxide board was originally made by Mr Hayes;^13321332 Roper {Day71/86:4-16}. Mr Roper and Mr Evans then suggested using the thinner Marley board to cover it.^13331333 Hayes {Day74/127:13-22}. Mr Evans told us that he could recall only the decision to revise the rig design to incorporate a thicker cladding panel.^13341334 Evans {Day72/192:13-15}. He was adamant, despite repeated questioning, that at no point had he known of the presence of a 6mm magnesium oxide board until he had found a reference to it on his phone in 2017, and that the only variation in design that he had been aware of at the time was the change to the thickness of the panel.^13351335 Evans {Day72/193:9}-{Day72/195:10}; Evans {CEL00010058/31} page 31, paragraph 114. He was equally adamant that he had not taken part in any meeting at which the use of a 6mm magnesium oxide board had been discussed,^13361336 Evans {Day73/2:15}-{Day73/3:5}. or that he had known that it had been used in the area of the cavity barriers.^13371337 Evans {Day72/199:1-13}. We did not find his evidence on those matters credible and indeed Mr Evans ultimately accepted that he had become aware of the use of the 6mm magnesium oxide boards at the latest by the time of a management action group meeting on 14 May 2014 although he maintained that he had not made the connection even by that time.^13381338 Evans {Day73/8:15-23}; {Day73/9:16}-{Day73/10:15}. Given his faulty memory on this point, we regard the evidence of Mr Hayes and Mr Roper as far more reliable and we conclude that Mr Evans was indeed aware of the changes to the design of the second test.

No new drawings

¶24.61 Celotex did not commission any drawings of the second test rig.^13391339 Roper {Day71/101:1-6}. Mr Roper said that that had been in part a financial decision.^13401340 Roper {CEL00010052/13}page 13, paragraph 5.35. He had also formed the view that Simco did not particularly wish to assist further.^13411341 Roper {Day71/100:9-20}. Although Mr Roper did not think that the decision not to produce revised drawings had been motivated by a desire to conceal the use of the 6mm magnesium oxide board,^13421342 Roper {Day71/101:20-25}. its failure to do so was entirely consistent with Celotex's overall desire to conceal as far as possible the introduction of that layer. The only drawings available to BRE before the test on 2 May 2014, therefore, contained the same design as had been tested in February. Moreover, although revised drawings were subsequently commissioned by Celotex at the end of June 2014,^13431343 {CEL00000980}. they also omitted anything to indicate the use of 6mm magnesium oxide boards.^13441344 {CEL00000997}.

134

Source of materials

¶24.62 Cavity barriers were ordered from Siderise Insulation Ltd and were delivered to BRE on 7 April 2014.^13451345 {BRE00032372/2}. The cladding, which was Marley Eternit in thicknesses of 12mm and 8mm, was obtained from a company called FGF Ltd and delivered to BRE on 14 April 2014.^13461346 {BRE00032372/3}. The delivery note for the 8mm ruby Eternit boards that were intended to cover the 6mm magnesium oxide boards was signed by Mr Clark.^13471347 Clark {Day96/3:2-23}. The FR5000 was delivered on 1 April 2014, and again was signed for by Mr Clark.^13481348 {BRE00032372/6}.

¶24.63 The 6mm magnesium oxide boards were supplied by Euroform Products Ltd, which also supplied the 12mm sheathing board.^13491349 Roper {Day71/98:4-13}. Although the invoice shows the delivery as being to Celotex,^13501350 {CEL00001985}. Mr Roper asked for it to be delivered to BRE.^13511351 {CEL00001981}. Although BRE disclosed a delivery note for the sheathing board from the first test (addressed 'FAO Phil Clark', although signed for by another BRE employee),^13521352 {BRE00032372/4}. no delivery note for the sheathing board used in the second test has been disclosed by BRE. We return to the significance of that below.

The second test – May 2014

¶24.64 The second test was carried out on 2 May 2014. As set out above, the system comprised a steel-framed rig, 12mm magnesium oxide sheathing board, 100mm Celotex FR5000, Lamatherm fire barriers, 8mm Marley Eternit cladding over 6mm magnesium oxide boards over the level 2 thermocouples and 12mm Marley Eternit cladding over all other areas. The results met the criteria in BR 135.

BRE's knowledge of the design changes

¶24.65 As in the case of the test on 14 February 2014, the officer at BRE in charge of the test on 2 May 2014 was Phil Clark. There was a serious conflict of evidence about what, if anything, he knew about the changes in the design between the rig tested in February and the rig tested in May. Mr Clark said that he had known nothing of the changes in design other than the use of thicker external cladding,^13531353 Clark {BRE00005768/70} page 70, paragraph 287. but his evidence was contradicted by Mr Hayes and Mr Roper, both of whom had no doubt that he had been well aware not only of the thicker Marley panels but also of the use of the 6mm magnesium oxide boards.^13541354 Roper {Day71/99:11-19}; Hayes {Day74/148:11-17}. Mr Roper told us that he remembered having referred to the additional materials in the course of a telephone conversation with Mr Clark when the second test was being planned.^13551355 Roper {Day71/98:19-22}. Mr Hayes remembered a conversation with Mr Clark in the course of which he had referred to his experiences of working during the BSE outbreak.^13561356 Hayes {CEL00010154/22} page 22, paragraph 67. Mr Hayes said that during the conversation, Mr Clark had expressed the view in response to a question from Mr Roper that the test would have been successful even without the addition of the 6mm magnesium oxide boards. Mr Clark denied the conversation and sought to cast doubt on Mr Hayes' recollection challenging the description he had given of his office,^13571357 Clark {Day96/27:1-22}. but the description given by Mr Hayes was similar to that given by Patrick Jones of Simco, who installed the rig for both tests and we accept what he told us.^13581358 Jones {Day99/129:2-20}.

135

¶24.66 BRE's standard procedures required the project officer to compare the drawings of the test rig with the structure handed over and if any discrepancy were noted to ask for revised drawings.^13591359 Clark {BRE00005768/9} page 9, paragraph 41. The senior member of staff who signed the test report was also expected to check the photographs and drawings.^13601360 Clark {Day95/56:4-9}. BRE's standard operating procedures dated May 2013 required all components to be checked and, if necessary, samples to be taken.^13611361 {BRE00005773/3}. BRE produced forms for that purpose,^13621362 {BRE00005773/3}. but no checklists were disclosed for the second test, and Mr Clark explained that they had not normally been used, otherwise, perhaps, than as an aide memoire.^13631363 Clark {Day95/156:19-25}. Had the drawings been compared to the rig as it was built, or even to the components being delivered to BRE, the discrepancy between the drawings and the rig as built would have been obvious. Indeed, Mr Clark knew that the drawings had not been revised to show the thicker Marley panels and therefore knew that the Simco drawings did not accurately describe the test rig.^13641364 Clark {Day95/158:12-16}.

¶24.67 In coming to a decision about whether Mr Clark knew that the second test rig contained the additional 6mm magnesium oxide boards we have carefully considered the detailed submissions made by Celotex and BRE. Having done so, we have come to the conclusion that the evidence, taken as a whole, points strongly to the conclusion that Mr Clark was aware of the presence on the test rig of the 6mm magnesium oxide boards and the 8mm Marley Eternit boards that were used to cover them. First, it is right to say that we found Mr Roper and Mr Hayes more reliable as witnesses than Mr Clark. If the question had turned on the evidence of the witnesses alone we should have accepted their accounts of their conversations with Mr Clark. However, their evidence did not stand alone.

¶24.68 The delivery notes indicate that Mr Clark took delivery of the 12mm Marley panels together with the thinner 8mm ruby panels which were used to cover the 6mm magnesium oxide boards.^13651365 {BRE00032372/6}. There was no rational explanation for including 8mm and 12mm Marley panels in the components unless the thinner panels were to cover the 6mm magnesium oxide boards, and Mr Clark could not explain why the different thicknesses had not been noted other than as the result of an oversight.^13661366 Clark {Day96/7:1-9}; {Day96/8:3-11}. Mr Roper had arranged for the delivery of all of those materials and expected Mr Clark to sign for them because, as he understood it, he took delivery of most of the materials used on the test rigs.^13671367 Roper {Day71/98:14-18}; Hayes {Day74/145:2-4}. It is inherently unlikely that Mr Roper would have openly arranged for the delivery of those materials to BRE, which had a proper system for recording deliveries, if he had been intending to conceal their presence from BRE, and unless Mr Clark was aware of the plan, concealment would have been essential, probably quite difficult and certainly risky. There is no reason to think that BRE's system for recording deliveries was so inefficient that Mr Roper can have expected to get the 6mm magnesium oxide boards onto the test rig without anyone's noticing.

136

¶24.69 The absence from BRE's disclosure of a delivery note covering the materials supplied by Euroform has not been explained by BRE and no credible explanation for it has been provided. That it was simply lost is too much of a coincidence and does not do justice to BRE's systems. It could have been removed or destroyed deliberately, but no one at BRE other than Mr Clark was involved in the Celotex tests in a way that would provide a motive for doing that. Mr Clark initially told us that he had not had access to either the paper or electronic versions of the test files other than documents that had been provided to him by BRE,^13681368 Clark {Day95/182:14-21}. but we subsequently learnt that he had attended BRE's offices over the course of several days when preparing his witness statement for the Inquiry and been given unsupervised access to any documents he wished to review. Mr Clark was unable to explain satisfactorily why he had failed to mention that when he made his statement.^13691369 Clark {Day96/110:2-12}.

¶24.70 The use of the ruby Eternit board meant that the rig had a distinctive red band in the area of the level 2 thermocouples and at the top of the rig. Mr Clark said he had understood that the ruby board was used because there had been a shortage of panels in the other colour,^13701370 Clark {BRE00005768/66} page 66, paragraph 269. but that was not very plausible and he had not asked whether it had any technical significance.^13711371 Clark {Day95/190:18-24}. However, the absence of any attempt on the part of Celotex to divert attention from the use of that material and to forestall questions of that kind is inconsistent with any plan to deceive BRE. Clearly, any attempt to smuggle the ruby board past BRE would have been risky.

¶24.71 Patrick Jones of Simco, who constructed the test rig for Celotex, was clear that he had been given no instructions to conceal the 6mm magnesium oxide board from BRE.^13721372 Jones {Day99/151:19}-{Day99/152:3}. His impression was that BRE was carrying out its normal checks.^13731373 Jones {Day99/155:22}-{Day99/156:5}. The rig was assembled over several days and there was no attempt to prevent BRE from seeing the magnesium oxide boards, and no reason to think that they had not been observed.^13741374 Jones {Day99/151:19}-{Day99/152:3}. He told us that there was nothing underhand going on and that BRE was fully aware of what was on the test rig.^13751375 Jones {Day99/144:8-16}; {Day99/150:23}-{Day99/152:3}; {Day99/164:9-25}. Mr Jones was a patently honest witness who had no reason to be anything other than candid in his evidence.

¶24.72 Following the test, the rig remained standing for some 17 days until it was stripped down on 19 May 2014.^13761376 {BRE00005548/6}. That provided a further opportunity for anyone at BRE to ask questions about the use of the ruby board.

¶24.73 On 15 May 2014, during the stripping down of the rig, which was also carried out by Mr Jones, Mr Roper asked Mr Clark by email to ensure that photographs were taken of the fire barrier at the level 2 thermocouples.^13771377 {BRE00005548/6}. Again, that is wholly inconsistent with an intention on the part of Celotex to deceive BRE or to gamble on its staff's failing to notice the presence of the 6mm magnesium oxide boards. If Mr Roper had hoped that Mr Clark might fail to notice them, it is unlikely that he would have directed his attention to the very place in which they were being used. The photographs on BRE's files clearly show their presence.^13781378 {BRE00030681}.

137

¶24.74 The recording made by the head camera worn by Mr Clark during the second test does not of itself demonstrate conclusively that Mr Clark was aware of the presence of the magnesium oxide board, but it contains several comments made by Mr Clark that strongly suggest that he was: a. Mr Clark referred to "changing two things at the same time".^13791379 {INQ00014220/2}. His explanation that

that had been a reference to the scientific principle that only one variation should be made to a system at a time was unconvincing. He was unable to explain satisfactorily why, if he thought that only one change had been made, he had referred to changing two things.^13801380 Clark {Day96/55:5-21}. Celotex had in fact made two significant changes to the design: the use of an increased thickness of cladding overall (12mm Marley Eternit as opposed to 8mm); and the use of the 6mm magnesium oxide board.

¶b. Mr Roper at one point drew Mr Clark's attention to the ruby board, (which he described as "the rose-coloured panel"),^13811381 {INQ00014220/2}. being aware, he said, that Mr Clark knew what material was behind it.

¶c. At one point Mr Clark commented, "See how that flame seems to have ceased now that the board is there, because you're losing a lot of the energy from behind it".^13821382 {INQ00014218/3}. Unless he was aware of the presence of the magnesium oxide board, it is difficult to understand which board he was referring to.

¶d. In response to a comment by Mr Clark that the board at the level of the fire barrier had not deformed, Mr Roper said that it was "a shame we put calcium silicate behind it" and subsequently "like I say, it's a shame commercially."^13831383 {INQ00014220/2}; {BRE00035419} Mr Roper said that he tended to refer to calcium silicate interchangeably with magnesium oxide, because their fire resistance properties are very similar.^13841384 Roper {Day71/170:12-14}. It is likely that they were referring to the 6mm magnesium oxide board because its use in that location would not be adopted commercially.

¶24.75 The way in which Celotex behaved in arranging the test was wholly inconsistent with any plan to deceive BRE. It could have acted as it did only with the knowledge of Mr Clark. We are satisfied that Mr Clark was aware of the design of the rig for the second test, including the use of the 6mm magnesium boards and the 8mm ruby Eternit boards, and the reasons for it.

¶24.76 The question remains why Mr Clark, a senior BRE burn hall technician, was willing to lend himself to a plan to a test rig that was so obviously unrepresentative and designed only to meet the performance criteria, since there was no evidence that he had anything to gain personally from a successful test. He might have thought that a customer should be allowed to test any design of rig it liked, provided the subsequent report accurately reflected its structure and the outcome of the test, but if that was his view, it is difficult to understand why he subsequently denied any knowledge of the additional material. In essence, over the years leading up to 2014 BRE had by degrees lost sight of the importance of maintaining a proper distance between itself and clients and of the need for scientific rigour and independence. Its internal controls were not the subject of regular and robust training and there was no independent supervision that ensured that conflicts of interest were properly managed. It had gradually become much closer to its clients and wanted to work with them to find solutions to their problems. Scientific rigour and independence were sacrificed in favour of financial sustainability.

138

Decisions following the second test

¶24.77 On 12 May 2014, Mr Roper and Mr Evans met to discuss the outcome of the test. The meeting appears to have been held in Mr Chambers's office,^13851385 {CEL00010627}. and Mr Roper thought it was likely the purpose of the meeting was to brief him.^13861386 Roper {Day71/106:18-20}. Mr Chambers did not believe he had attended such a meeting and could not recall how or by whom he had been advised that the system had passed.^13871387 Chambers {CEL00010056/6} page 6, paragraph 31.

The slide show

¶24.78 At or just after that meeting Mr Roper was asked by Mr Evans to prepare a presentation which could be used to report to the management action group at their next meeting, which was scheduled for 13 and 14 May 2014. Mr Chambers, Mr Warren and Mr Evans were present on 13 May 2014;^13881388 {CEL00009565}. Mr Chambers and Mr Evans were present on 14 May but Mr Warren probably was not. Mr Roper appears not to have attended on either day.

¶24.79 The slides created by Mr Roper described the requirements for use of combustible insulation on buildings over 18 metres in height, referring to the text of paragraph 12.7 of Approved Document B.^13891389 {CEL00000933/2}. A further slide set out the conclusions drawn by Mr Roper from the results of his market research, namely,^13901390 {CEL00000933/11}. a. that people used K15 because there was no alternative available; b. that no one understood the BS 8414 test requirements; c. that levels of understanding among building control bodies varied widely; and d. that the market would be happy to buy an alternative to K15 if one were available.

¶24.80 Although Mr Evans denied that the slide showed that Celotex intended to capitalise on the confusion in the market,^13911391 Evans {Day72/206:9-16}. that is the only reason why it could have been relevant for the management action group to know that nobody understood the testing requirements or that building control officers had widely differing levels of understanding on the subject. Mr Evans could not explain those statements, other than by saying that the purpose of the slides was to acquaint everyone with the research that had been carried out,^13921392 Evans {Day72/207:2-14}. but that hardly explains why they were thought to be of any interest to senior management. The simplest, and we think the most likely, explanation is that the slides were put before the meeting precisely to enable it to decide whether to take advantage of that confusion, and if so, how.

139

¶24.81 The slides also described the two BS 8414 tests that Celotex had carried out, including the systems tested. The description of the February test referred to 8mm Eternit cladding and Lamatherm cavity barriers.^13931393 {CEL00000933/12}. In the case of the May test, the description listed 12mm Eternit cladding and Lamatherm cavity barriers with 6mm magnesium oxide.^13941394 {CEL00000933/14}. When the slides are compared, differences are quite apparent, although there is no mention of the use of the 8mm Eternit panels that had also been used in the May test. It follows that all those present were, or should have been, aware of the presence of the 6mm magnesium oxide boards used in the May test, as well as the use of the thicker 12mm Eternit boards. Indeed, Mr Evans accepted that the presentation had provided that information.^13951395 Evans {Day72/211:19-21}. Mr Evans said he did not remember whether there had been any discussion about the 6mm magnesium oxide boards at the meeting,^13961396 Evans {Day72/214:8-10}; {Day73:4/11-16}. or whether anyone had asked how the test had passed in May after it had failed in February.^13971397 Evans {Day72/214:21-25}. In our view it is hard to believe that no one at the meeting asked that question, if they had not been told directly, since it was critical to their decision about what to do next. The slides show that the use of the 6mm magnesium oxide boards was openly acknowledged at that meeting.

¶24.82 The slides were presented at the meeting on 14 May 2014, at which Mr Evans was scheduled to lead a discussion on the subject at 11.30 am.^13981398 {CEL00009566}. There was a disagreement about who presented the slides, but it makes little difference who did so. What is clear is that they show that those attending the meeting were or should have been aware of the presence of the 6mm magnesium oxide boards and that Mr Roper did not seek to keep their use a secret from senior management.

¶24.83 The slides may also have been sent to Saint-Gobain. An email from Mr Evans to Mr Roper on 16 May 2014 about the project refers to a meeting with Jonathan Cheeseman of Saint- Gobain's legal team, which appears also to have been attended by Craig Chambers.^13991399 {CEL00003121}. In that email, Mr Evans told Mr Roper that his presentation could provide all the background needed to explain Celotex's approach. Mr Evans was unable to recall which set of slides he had been referring to,^14001400 Evans {Day73/121:5-8}. but given that this set had been used for reporting to the management action group (and that the metadata relating to the short set of slides discussed below suggests that they were not created before 30 May 2014),^14011401 {CEL00000961}. it is likely that Mr Evans was referring to the slides shown to the meeting on 14 May 2014.

Instruction to produce slides "for general business use"

¶24.84 Mr Roper said that after the meeting, Paul Evans had asked him^14021402 Roper {Day71/113:16-19}. to create "for general business use" another version of the slides which did not refer to the February test or to the 6mm magnesium oxide boards.^14031403 Roper {CEL00010052/14-15} pages 14-15, paragraph 6.3 He told us that Mr Evans had suggested that after the meeting a decision had been taken to launch the product and to say nothing further about the use of 6mm magnesium oxide boards.^14041404 Roper {Day71/114:6-10}. Mr Roper subsequently produced such a presentation.^14051405 {CEL00000961}. Mr Evans did not remember having given any such instruction, but was unable to provide any other explanation for Mr Roper's action.^14061406 Evans {Day73/13:12-19}. He was also unable to think of any reason why anyone else on the project team might have instructed Mr Roper to do that without his knowledge.^14071407 Evans {Day73/19:21}-{Day73/20:1}. We have little doubt that Mr Roper created the slides because he was asked to do so.

140

¶24.85 Mr Roper's evidence strongly supports the conclusion that the management action group made a decision at the meeting on 13 and 14 May 2014 to market FR5000 without disclosing the fact that 6mm magnesium oxide boards had been used to improve the performance of the test rig. There was a risk that BRE might disclose the use of the 6mm magnesium oxide and the 8mm Marley Eternit panels in its test report, but it was Celotex's general practice not to release such reports^14081408 Roome {Day70/73:17-22}; Roper {Day71/135:19-25}. and it might therefore have hoped to maintain the deception indefinitely. It was, we think, a risk that Celotex was prepared to run, having regard to the ignorance of a substantial section of its potential customers, a point to which we return in Chapter 56.

BRE draft report

¶24.86 Mr Clark produced a first draft of the test report,^14091409 Clark {Day95/171:4-7}. which he sent to Mr Roper on 19 June 2014.^14101410 {CEL00003177}. Mr Roper was about to go into a meeting with David White and Graham Perrior of NHBC when it arrived.^14111411 Roper {Day71/120:21}-{Day71/121:3}. It was dated 2 June 2014^14121412 {CEL00003178/1}. and described the system that had been tested as comprising 100mm RS5000 with 8mm Marley Eternit, but made no reference to the 6mm magnesium oxide boards.^14131413 {CEL00003178/6}. Mr Roper denied having asked Mr Clark to omit any reference to the 6mm magnesium oxide boards; he thought it was possible that Mr Clark had done so of his own initiative because of his view that they would not have made any difference in the light of the increase in the thickness of the cladding panel.^14141414 Roper {Day71/128:8-13}. Nonetheless, he said that he had been surprised by the omission.^14151415 Roper {Day71/131:15-24}. Mr Clark denied having consciously decided to omit any reference to the 6mm magnesium oxide boards; he said that he had simply been unaware of them, although he was unable to explain how he had overlooked them.^14161416 Clark {Day95/179:25}-{Day95/180:20}. He accepted that he must have seen the photograph at Figure 18 of the draft report showing the magnesium oxide board, as he had selected it for inclusion, but said that he had not studied it at the time.^14171417 Clark {Day96/67:23}-{Day96/68:10}.

¶24.87 It was clear from the photographs in the draft test report that at the level 2 thermocouples and at the top of the rig, exterior panels had been used that differed from the rest of the exterior of the rig. Figure 15 showed the ruby Marley Eternit panels at those levels, although it is not possible to tell from the photograph that the thickness of the ruby layer was different from that of the rest of the panels.^14181418 {CEL00003178/25}. However, the fact that those boards were at least of a different colour was not mentioned anywhere in the report. The report also contained a photograph of the rig in the process of being dismantled (Figure 18), which showed the white 6mm magnesium oxide boards quite clearly.^14191419 {CEL00003178/28}. The photograph shows the insulation layer, with one panel of the 6mm magnesium oxide still present at level 2 and the whole of it present at the top of the rig. Those were the levels which had been covered by the ruby cladding panels shown in Figure 15. A simple comparison between the two photographs would have alerted a careful reader to the fact that there was white material behind the ruby cladding but on top of the insulation layer, which called for an explanation.

141

Discussion of the draft report

¶24.88 On 19 June 2014, following receipt by Celotex of the draft report earlier that morning and after the NHBC representatives had left, a discussion ensued between Mr Roper, Mr Evans and another colleague at Celotex, Paul Reid (although it appears that Mr Reid was not present throughout).^14201420 Roper {CEL00010052/16} page 16, paragraph 7.4. Notes were made on a whiteboard of the concerns raised by NHBC, and the conversation which had subsequently taken place about the BRE draft test report.^14211421 {CEL00002517}. A photograph of the whiteboard was taken by Mr Evans.^14221422 Evans {Day73:36/6-12}. The first five points reflected concerns that had been raised by NHBC at the meeting that morning (one of which had been that the rig had been deliberately over-engineered through the use of large 12mm cladding panels with minimal gaps)^14231423 Roper {CEL00010052/16} page 16, paragraph 7.4 and Celotex's potential responses. Some of the proposed responses were less than frank. In relation to NHBC's concern that the test had been over-engineered, the proposed response was to note that cladding panels were commonly available in 8mm or 12mm but that only 12mm had been available for testing, whereas in fact Celotex had deliberately chosen to use large 12 mm panels. The final entry was in different handwriting, having been added by Mr Roper.^14241424 Evans {Day73:37/8-17}. It recorded a concern relating to the use of a "calcium silicate board" at level 2, although, as we noted earlier, Mr Roper did not distinguish in his own mind between calcium silicate and magnesium oxide.^14251425 Roper {Day71/170:12-14}. If NHBC were to become aware of the existence of the 6mm magnesium oxide board, Celotex proposed to say that it had confidence in the fire barriers to expand. When examined about that Mr Roper admitted that it was a poor attempt to hide the fact that 6mm magnesium oxide boards had been used.^14261426 Roper {Day71/171:22}-{Day71/172:1}. He said that Mr Evans and Mr Reid had suggested that they remove Figure 18 from the report.^14271427 Roper {CEL00010052/18} page 18, paragraph 7.16; Roper {Day71/137:21-24}. Mr Reid denied having been involved in the conversation at that point or having had any knowledge of the discussion about the removal of Figure 18,^14281428 Reid {CEL00011267/11} page 11, paragraphs 57-59. and we are not able to reach any firm conclusion about the extent of his involvement.

Decision not to correct the report

¶24.89 The upshot of the discussion was that it was decided that the omission from the test report of any reference to the 6mm magnesium oxide boards would not be corrected.^14291429 Roper {CEL00010052/18} page 18, paragraph 7.14 Mr Evans did not recall any discussion with Mr Roper about that,^14301430 Evans {Day73/24:18-23}. but he did accept that the white board reflected a discussion which had taken place about removing Figure 18.^14311431 Evans {Day73:38/19-23}. Although he recalled its having been discussed,^14321432 Evans {Day73/29:3-10}. he claimed to be unsure whether he had looked at the photograph.^14331433 Evans {Day73/29:17}. We do not accept that. Figure 18 clearly showed the 6mm magnesium oxide board and it could not have been discussed without Mr Evans' having appreciated that fact. It is also not credible that he could have taken part in a discussion about the removal of a photograph from the report without having looked at the photograph in question.

142

¶24.90 Similarly, we are not convinced by Mr Evans' assertion that he did not ask about the two orange layers shown in the photographs of the rig.^14341434 Evans {Day73/66:9-21}. The test represented the second attempt to succeed in what was an important project for Celotex. We do not think that Mr Evans felt as little concern as he suggested.^14351435 Evans {Day73/68:2-7}.

Request to remove figure 18

¶24.91 On 1 July 2014, Mr Roper replied to Mr Clark's email attaching the draft report with a number of suggested changes, including the removal of Figure 18 and its substitution by some photographs showing the condition of the insulation at the conclusion of the test.^14361436 {CEL00001350/2}. The email refers to a previous discussion between Mr Roper and Mr Clark about the removal of Figure 18, but Mr Clark could not remember what had been said.^14371437 Clark {Day96/65:19-24}. Mr Roper could not remember having specifically told Mr Clark that Celotex did not want any reference to be made to the areas of the rig where the 6mm magnesium oxide board had been used; the focus, as he recalled it, was simply on getting Figure 18 removed.^14381438 Roper {Day71/178:20}-{Day71/179:8}.

¶24.92 The reason Mr Roper gave for asking Mr Clark to remove Figure 18 from the test report was not the true reason, as he admitted. The truth was that Celotex wanted to ensure that a picture showing the 6mm magnesium oxide board did not appear anywhere in the report.^14391439 Roper {Day71/141:22}. Mr Roper said that Mr Evans and Mr Reid had told him to send the false explanation,^14401440 Roper {Day71/141:23}-{Day71/142: 4}. but that he had thought that Mr Clark was aware of the real reason.^14411441 Roper {Day71/142:20}-{Day71/143:3}. Mr Evans did not remember having asked Mr Roper to send the email,^14421442 Evans {Day73/33:4-13}. and he may not have done, but he was copied into it and did not express any surprise at the request. We are satisfied that he knew about it, but cannot say whether Mr Reid also knew.

¶24.93 Mr Roper also sent Mr Clark some revised drawings, which showed a ventilation gap that had not existed on the rig as built but still did not show the 6mm magnesium oxide boards.

Internal review and issue of the report

¶24.94 BRE did not reply to those requests. Internally, the final version of the report was checked by Tony Baker^14431443 {BRE00005851}; Baker {Day100/143:4-11}. and Stephen Howard,^14441444 {BRE00030677}; Howard {Day99/41:2-12}. each of whom failed to identify the discrepancy between the description of the system tested and the presence of the white layer shown in Figure 18 or the ruby Marley Eternit boards shown in Figure 15.

143

¶24.95 The final test report was issued by BRE on 1 August 2014. It included the additional photographs but also Figure 18, which became Figure 19 in the final version.^14451445 {CEL00001109/29}. As with earlier versions of the draft, the description of the system tested contained no mention of either the 8mm ruby Marley Eternit boards or the 6mm magnesium oxide boards.^14461446 {CEL00001109/6}. Mr Roper said that he, Mr Evans and Mr Warren had discussed the report and had decided that it would be risky to press BRE to remove Figure 19, as it might result in their correcting the description of the system.^14471447 Roper{CEL00010052/18} page 18, paragraph 7.16; Roper {Day71/144:25}-{Day71/145: 7}.

Mr Clark's involvement in the BRE test report

¶24.96 As set out above, we are satisfied that Mr Clark was aware of the changes that had been made to the design of the rig tested on 2 May 2014, including the use of the 8mm ruby Marley Eternit boards and the 6mm magnesium oxide boards at two levels. We also think that he realised that the presence of the 6mm magnesium oxide board might have enabled the system to pass the test, although it would not be used in a commercial context. That being so, we have to ask ourselves why he failed to include them in the list of components in the test report, not least since the delivery note for the ruby 8mm Marley Eternit was available to him on the test file.^14481448 {BRE00032372/3}.

¶24.97 Mr Roper denied having specifically asked him to do so, although, as can be seen from the attempt to suppress Figure 18 in the draft report, we have little doubt that he would, if necessary, have done so. It is likely, in our view, that Mr Clark was aware of Celotex's wish to avoid any reference to the magnesium oxide boards (and therefore the 8mm ruby Marley Eternit boards) and did not need to be asked to omit any reference to them. Mr Clark denied having deliberately omitted any reference to either of those components.^14491449 Clark {Day96/75:6-13}.

¶24.98 In view of the length of time it took to issue the draft report (just over 6 weeks), it is possible that Mr Clark simply relied on the original drawings when drafting the description of the system. That might explain why the draft referred to 8mm Marley Eternit panels as shown in the drawings for the first test.^14501450 {CEL00000807}. It might also explain the omission of the 6mm magnesium oxide boards, which had also not been used in the first test. However, we do not think that is the most likely explanation. We think he was willing to leave out any reference to them in order to assist Celotex, though how he expected the report to escape the scrutiny of those who would need to approve it before it was issued is less clear. Perhaps he knew that the system for reviewing reports was not very rigorous. Mr Clark said he had not looked at Figure 18 again in light of the request to remove it, because BRE did not remove material in response to a request from a client.^14511451 Clark {Day96/71:17}. We think that is probably what happened.

¶24.99 Although Mr Clark did not remove Figure 18 from the report, that does not mean that his omission of a reference to the 6mm magnesium oxide boards from the list of components was an oversight. He knew, or at least suspected, that the system had only met the BR 135 performance criteria at least in part because of the presence of the 6mm magnesium oxide boards, and that such boards were not representative of a system in actual use on a real building. He therefore knew that referring to the presence of those boards in the report would significantly undermine the commercial viability of the product and therefore deliberately omitted any reference to them.

144

¶24.100 The test report was subsequently peer reviewed by Stephen Howard and Mr Baker, neither of whom inquired about the white layer of magnesium oxide board or the ruby Eternit boards. However, that merely suggests a lack of curiosity and a review system that was less than robust. There is no evidence that in conducting their review either of them realised that Celotex had manipulated the test by adding a layer of magnesium oxide board behind the ruby Eternit boards, still less that they realised that Mr Clark was aware of the fact.

The RS5000 classification report

¶24.101 The classification report was issued by BRE on 4 August 2014 and reissued following typographical corrections on 11 August 2022.^14521452 {CEL00002133/1}. It ran to 12 pages and described the test in the same terms as the test report, omitting any reference to the 6mm magnesium oxide boards or the 8mm Marley Eternit boards.^14531453 {CEL00002133/4}. The classification report did not contain any photographs. Indeed, there is nothing in it to alert even a scrupulously careful reader to the possibility that the description of the components of the rig was incomplete.

The abridged report

¶24.102 The classification report contained a prominent notice that it was only to be distributed in its entirety, without amendment and that any reference to the results contained in it was to be accompanied by a copy of the full report, or a link to a copy of the full report.^14541454 {CEL00002133/1}. Despite that, on 11 August 2014, Mr Roper sent Paul Evans copies of the test and classification reports, asking for a discussion about which they were going to use as proof that Celotex had satisfied BR 135.^14551455 {CEL00009724/1}. Although his recollection of that discussion was hazy,^14561456 Roper {Day71/153:17-19}; {Day71/152:24}-{Day71/153:8}. he recalled that there was some reluctance to provide customers with any reports, and that they had decided that an abridged version of the classification report should be produced for customers to avoid having to provide them with details of the test rig.^14571457 Roper {Day71/154:3-6}. Mr Roper understood that a similar approach had been adopted by Kingspan.^14581458 Roper {Day71/154:14-23}.

¶24.103 The abridged report was prepared on Mr Roper's instructions.^14591459 Roper {Day71/155:11-18}. It contained no diagrams or photographs showing the system tested, but it did contain a list of components which omitted the 6mm magnesium oxide boards and the 8mm ruby Eternit boards.^14601460 {CEL00011125/2}. On 15 August 2014 a copy was sent to Jonathan Roome, again at Mr Roper's request.^14611461 {CEL00009725}. It was then sent by him to clients, including Harley on 27 August 2014.^14621462 {CEL00011960}.

¶24.104 As early as 29 August 2014, Anthony Harris of Celotex contacted Mr Roper and Mr Roome because a client to whom Celotex had provided the abridged report had complained that it was incomplete and that sections were missing.^14631463 {CEL00010805}. Mr Roper replied the same day providing the full classification report, but not the full test report. He told Mr Harris that the salient information was contained in the BR 135 classification report.^14641464 {CEL00010805}. As Mr Roper accepted, the salient information was not in fact contained in that report, not least because it did not include the photograph of the 6mm magnesium oxide boards.^14651465 Roper {Day71/162:12-23}.

145

The marketing of RS5000

¶24.105 Preparatory work on marketing materials to support the launch of RS5000 started in July 2014. On 1 July 2014, Jonathan Roper sent Paul Evans a marketing plan^14661466 {CEL00001213}. which set out the objectives for RS5000, including being presented as Celotex's "primary rainscreen application offering to compete directly with Kingspan K15 & Rockwool Duo-Slab".^14671467 {CEL00001214/2}. It also contained a phrase that would become something of a mantra in the marketing material, "the first PIR insulation board tested and approved to BR 135 and therefore acceptable for use in buildings above 18m in height". Mr Roper and Mr Hayes then started work preparing the marketing literature which was intended to accompany the launch of RS5000.^14681468 {CEL00003179/4}; Roper {Day71/194:13-23}.

¶24.106 On 15 July 2014, Mr Evans replied to an email from another colleague in the marketing team, Bex Hartlebury, about a press launch for RS5000, complaining that the message that RS5000 was acceptable for use on buildings over 18 metres in height was not always prominent enough, but saying that it was necessary to exercise caution about how that message was given and how often.^14691469 {CEL00009674}. Mr Evans accepted in evidence that there was a risk that without qualification it might mislead people into thinking that RS5000 was generally suitable for use on buildings above 18 metres in height,^14701470 Evans {Day73/86:5-10}. but he denied that Celotex was trying to walk a fine line between the headline and the small print.^14711471 Evans {Day73/86:18-24}. In our view, however, that is exactly what it was trying to do.

The RS5000 marketing literature

¶24.107 Celotex launched RS5000 on 5 August 2014. A number of documents were prepared to assist in its marketing:

¶a. a Rainscreen Cladding Compliance Guide ("the Compliance Guide");^14721472 {CEL00000012}.

¶b. a Rainscreen Cladding Specification Guide ("the Specification Guide");^14731473 {CEL00000013}.

¶c. a Handy Guide;^14741474 {CEL00000437}. and

¶d. a Data Sheet.^14751475 {CEL00007961}.

¶They were drafted by Mr Roper with the assistance of Mr Hayes (in relation to the

¶Specification Guide)^14761476 Hayes {Day74/173:11-20}. and others in the marketing and technical teams^14771477 Evans {Day73/81:16-22}. but they were reviewed and ultimately approved by Paul Evans.^14781478 Evans {Day73/82:1-3}. Mr Evans told us that Celotex also sought guidance on them from Saint-Gobain, in that they had been passed to the legal team, but he was unsure whether they had been reviewed by anyone on the Saint-Gobain board.^14791479 Evans {Day73/82:3-17}. Mr Roper also told us that Saint-Gobain's internal lawyers had reviewed the documents and that it had been as a result of that review that the disclaimers had appeared in the documents.^14801480 Roper {Day71/182:5}-{Day71/183: 6}. We accept their evidence about that.

146

¶24.108 We therefore turn to examine those documents.

The Compliance Guide

¶24.109 The front cover of the Compliance Guide stated that it was for use "when specifying Celotex RS5000 in buildings above 18m",^14811481 {CEL00000012/1}. explaining on the first inside page that it provided a step by step guide to the alternative route to compliance under Approved Document B by meeting the performance criteria in BR 135 following testing to BS 8414-1:2002 or BS 8414-2:2005.^14821482 {CEL00000012/2}. That was not accurate, if only because RS5000 had not been tested in accordance with BS 8414-1:2002 (masonry wall construction).^14831483 Evans {Day73/51:9-11}. Mr Evans was unable to explain that inaccuracy.^14841484 Evans {Day73/52:13-19}.

¶24.110 On page 3, it contained the following disclaimer:

"The classification applies only to the system as tested and detailed in the classification report. The classification report can only cover the details of the system as tested. It cannot state what is not covered. When specifying or checking a system it is important to check that the classification documents cover the end-use application".^14851485 {CEL00000012/3}.

¶24.111 The next page of the Compliance Guide contained a description of the system tested together with a diagram, Figure 4. The following components were listed:

¶a. 12mm Fibre Cement Panels;

¶b. Supporting aluminium brackets and vertical rails;

¶c. 100mm Celotex RS5000;

¶d. 12mm non-combustible sheathing board;

¶e. 100mm SFS system; and

¶f. 2 x 12.5mm plasterboard.^14861486 {CEL00000012/4}.

¶24.112 The description did not disclose either the presence of the 6mm magnesium oxide boards nor the 8mm Marley Eternit boards. Although the text goes on to warn that "the fire performance and classification report issued only relates to the components detailed and constructed in Figure 4" and that "any changes to the components listed and construction method set out in Figure 4 will need to be considered by the building designer",^14871487 {CEL00000012/4}. Figure 4 also omitted any reference to those elements. They had been omitted because, as

147

¶Mr Evans admitted, any reference to them would have revealed that the system tested had been entirely unrepresentative of a commercial construction.^14881488 Evans {Day73/62:15-22}.

¶24.113 The failure to refer to the 6mm magnesium oxide boards or the 8mm Eternit boards completely undermined the disclaimers on pages 3 and 4. If the description of the system tested was materially incomplete, as it was, even the most rigorous attempt to reproduce it was bound to fail. In its statements to the Inquiry Celotex was at pains to point out that the statements that RS5000 was suitable for use on buildings over 18 metres in height should be read in the context of the disclaimer on page 3 of the Compliance Guide, but the disclaimer was effectively meaningless, as Celotex well knew.

The Specification Guide

¶24.114 The introduction to the Specification Guide described RS5000 as

"a premium performance PIR solution for use in rainscreen cladding applications and suitable for use in building (sic) above 18 metres in height".^14891489 {CEL00000013/3}.

¶and in a section headed "Fire Performance", it went on to describe RS5000 as having been

"successfully tested to BS8414-2:2005 […], meets the criteria set out in BR 135 and is therefore acceptable for use in buildings above 18m in height".^14901490 {CEL00000013/5}.

¶This slogan was repeated throughout Celotex's other marketing material for RS5000.

¶24.115 Almost identical wording could be found in a Technical Bulletin issued by Kingspan in 2011, in which it described K15 as "successfully tested to BS8414:2002, can meet the criteria within BR 135 and is therefore acceptable for use above 18 metres".^14911491 {CEL00008510}. The similarity was not coincidental. As Mr Roper confirmed, Celotex took the wording directly from Kingspan's literature.^14921492 Roper {Day71/184:16-17}. Celotex had identified the device by which Kingspan had for years succeeded in selling an insulation product tested in one particular configuration for general use on buildings above 18 metres in height and had decided to adopt it.

¶24.116 The Specification Guide then described the system tested in the same terms as the Compliance Guide (i.e. without mentioning the 6mm magnesium oxide or 8mm Marley Eternit boards), together with the qualification that "the fire performance and classification report for Celotex RS5000 only relates to the components detailed above".^14931493 {CEL00000013/5}. Unlike the Compliance Guide, there is no diagram identifying that the listed "fibre cement panels" were in fact Marley Eternit. It did refer readers to the Compliance Guide for "full" specification details and provided a draft specification clause which said that "RS5000 has been successfully tested to BS8414-2:2005 and meets the performance criteria of BR 135".^14941494 {CEL00000013/8}.

¶24.117 Mr Roome accepted with hindsight that the Specification Guide was a thoroughly misleading document,^14951495 Roome {Day69/99:10-12}. and that a reader might have interpreted the claim that RS5000 was suitable for use on buildings over 18 metres in height as meaning it could be used generally above that height.^14961496 Roome {Day69/100:1-9}. The document was indeed thoroughly misleading and, although Mr Roome himself may not have realised it, had been deliberately crafted to mislead the reader into buying RS5000 for use on buildings over 18 metres in height without worrying about whether the external wall of which it was to form part was the same as that which had been tested. The disclaimer on page 5 that the fire performance and classification report related only to the components listed was not only buried away in the small print but was, as we have explained, itself disingenuous because the components of the system tested had not been fully described.

148

Other marketing documents

¶24.118 The Handy Guide repeated the expression "the first PIR board to meet the criteria set out in BR 135 and therefore is acceptable for use in buildings above 18m in height", without qualification, although it advised that the product should be installed in accordance with instructions supplied by Celotex.^14971497 {CEL00000437/8}.

¶24.119 The Product Datasheet contained the same claim prominently on the first page.^14981498 {CEL00007961/1}. Headers on each page included, without qualification, the words "suitable for buildings above 18m in height". Mr Roper explained that the reason for the disclaimer being placed on the third page was that certain people at Celotex wanted "to dilute that system message as much as they could".^14991499 Roper {Day71/205:8-10}. Mr Evans did not agree; his view was that it was not intended to be a free-standing statement but was to be read in conjunction with the description of the system that had been tested,^15001500 Evans {Day73/70:23}-{Day73/71:2}. but the description given on the third page of the Product Datasheet did not mention the magnesium oxide boards or the use of 8mm Marley Eternit. The section did contain a warning that the fire performance and classification report related only to the components described above,^15011501 {CEL00007961/3}. but that was itself misleading because the components of the system tested had not been fully described.

¶24.120 A press release prepared for the launch of RS5000 described it as a "new premium performing PIR solution for rainscreen cladding applications."^15021502 {CEL00009599/1}. That claim was itself false, because RS5000 was merely FR5000 under a different name. The press release repeated the claim that RS5000 was "acceptable for use in buildings above 18m in height" without qualification.

¶24.121 Taken in the round, the marketing literature for RS5000 contained two messages, both of which were false and deliberately misleading. The first, aimed at the designer or cladding specialist who did not read beyond the first page, was that RS5000 was suitable for use on any building over 18 metres in height, whereas it was not. The second was the warning aimed at the more conscientious reader that the fire performance and classification for RS5000 related only to the components described as making up the system tested, when in fact the components were misdescribed.

149

Internal launch of RS5000

¶24.122 RS5000 was launched on 5 August 2014, at a regional sales meeting at which Jonathan Roper gave a presentation to the sales team about the product.^15031503 Roper {Day71/208:16-22}. His slide presentation^15041504 {CEL00008668}. included a quotation from paragraph 12.7 of Approved Document B^15051505 {CEL00008668/5}. as well as the historical context provided by the fires at Knowsley Heights,^15061506 {CEL00008668/11}. Garnock Court,^15071507 {CEL00008668/13}. and in Dubai.^15081508 {CEL00008668/14}. Although he referred to test method BS 8414 and explained in terms that BR 135 was a system classification,^15091509 {CEL00008668/15-16}. he repeated the assertion that RS5000 was the first PIR board to meet successfully the performance criteria in BR 135 for insulated cladding systems and was therefore acceptable for use in buildings above 18 metres in height.^15101510 {CEL00008668/26}.

¶24.123 Jonathan Roome was present at the launch. Although he understood from the presentation that a BR 135 classification applied only to the system as tested,^15111511 Roome {Day69/52:1-9}. he accepted that the words "therefore acceptable for use above 18 metres in height" gave the impression that RS5000 was generally suitable for use on high-rise buildings and could be read as a representation that it was suitable for use in any rainscreen system on buildings over 18 metres in height.^15121512 Roome {Day69/53:2-9}. That is consistent with the evidence of Mr Roper, who said that, although he had wanted to make it clear that the test applied to a system as a whole, he had felt that the members of the management action group present at the launch had not wanted that message to be too prominent. He agreed that the combination of an early reference to the nature of the classification and the subsequent emphasis on the sales message had been a way of satisfying the technical requirements while giving a strong message to the sales teams that RS5000 was generally acceptable for use on buildings over 18 metres in height.^15131513 Roper {Day71/213:2-13}.

¶24.124 The launch presentation did not describe in any detail the system that had been tested but it did contain a diagram showing the rainscreen panel as 12mm Marley Eternit over 100mm of insulation.^15141514 {CEL00008668/27}. Mr Roome confirmed that he understood that to have been the system tested because it matched the depiction in the Compliance Guide and Specification Guide.^15151515 Roome {Day69/59:1-5}. That was all that he and the rest of the sales team were given by way of information about the system tested.^15161516 Roome {Day69/97:2-9}. He was not aware that 6mm magnesium oxide boards had been used in the test,^15171517 Roome {Day69/60:2-6}. and never became aware of it during the remainder of his time at Celotex.^15181518 Roome {Day69/60:19-23}.

¶24.125 The failure to identify the true components of the test in May 2014 was only one of a number of ways in which the presentation was misleading. It also implied that RS5000 was a completely new product and made no mention of the first, failed, test in February 2014. Mr Roome confirmed that he had not been aware of either fact. That was the result of a deliberate choice by Celotex.^15191519 Roper {Day71/220:19}-{Day71/221: 9}. In similar vein, Mr Roome was provided with the 12-page classification report only when he requested it following a direct enquiry from Harley,^15201520 Roome {Day70/73:12-16}. and during his time at Celotex he never saw the full 32 page test report.^15211521 Roome {Day70/73:17-22}.

150

Aftermath of the launch

Relations with NHBC

¶24.126 As we have already explained, Celotex had been in contact with NHBC since November 2013, when Mr Roper sought advice from Graham Perrior about his interpretation of the requirements for use of insulation on buildings over 18 metres in height.^15221522 {CEL00000748}. His approach was prompted by concerns about apparent inconsistencies in the approach being taken by NHBC.^15231523 Roper {Day70/198:5-15}. That was again motivated by a desire to understand how Kingspan was being widely used despite having carried out only one BS 8414 test.^15241524 Roper {Day70/201:20-22}.

¶24.127 On 15 May 2015 Jonathan Roper told Graham Perrior and David White of NHBC that Celotex had passed a BS 8414 test.^15251525 Roper {Day71/166:9-18}. We have already touched briefly on the meeting between Celotex and NHBC on 19 June 2014 which was held to discuss that test.^15261526 Roper {CEL00010052/15} page 15, paragraph 7.1. At that meeting NHBC raised a number of concerns which Mr Evans recorded on a white board.^15271527 {CEL00002517}. Celotex did not tell NHBC about the test in February 2014 which had failed or about the use of the 6mm magnesium oxide boards used in the successful test in May 2014.^15281528 Roper {CEL00010052/16} page 16, paragraph 7.4 Celotex deliberately misled NHBC about the true nature of the system that had been tested in order to obtain its approval of RS5000, which was seen as critical to the success of the product.^15291529 Roper {Day71/172:2-11}.

¶24.128 On 30 July 2014, Mr Roper sent the final draft of the test report to David White and Graham Perrior,^15301530 {CEL00000992}. possibly at the request of NHBC.^15311531 Roper {Day71/181:9-22}. That involved a certain level of risk, because the report contained the photograph showing the presence of the magnesium oxide boards. Mr Roper gave NHBC the impression that the report described exactly what had been tested,^15321532 {CEL00000992}. but that was not the case, since it omitted any reference to the magnesium oxide boards and the 8mm ruby Eternit panels.

¶24.129 By September 2014, NHBC was still expressing reservations about RS5000. On 5 September 2014, Jonathan Roper sent an email to Paul Evans explaining that he was aware of instances in which building control officers had refused to accept RS5000 on the basis that it had not been approved by NHBC.^15331533 {CEL00001020}. On the same date, a Celotex area sales manager, Michael Healey, asked NHBC to explain what problems it had with the test data.^15341534 {CEL00001022/3}. His request was made in response to a message from a designer on a project in which the NHBC project manager (Jon Behan) had said that there were questions about the tests that had been carried out and that the matter had been raised with Celotex.^15351535 {CEL00001022/4}.

151

¶On 22 September 2014, Mr Behan replied to Mr Healey, telling him that at present Celotex had no relevant test certificate to show that RS5000 was suitable for use on buildings over

¶18 metres in height.^15361536 {CEL00001022/2}. When Mr Healey referred to the test that Celotex had undertaken,

¶Mr Behan said that NHBC had "issues" with it and that the Celotex technical department was well aware of them.^15371537 {CEL00001022/4}. Mr Healey sent that exchange to Mr Roper who in turn sent it on to Mr Evans.^15381538 {CEL00001022/1}. It was understood as an indication that NHBC would not accept the use of RS5000 on buildings over 18 metres in height.^15391539 Evans {Day73/126:2-6}.

¶24.130 The importance of NHBC approval to Celotex was reflected in its 2015 budget presented at Saint Gobain's headquarters at Les Miroirs in Paris on 14 October 2014, which included £20,000 for "NHBC 'buy-in' for above 18m applications".^15401540 {CEL00003342/21}. It is unclear what the money was actually allocated for. Paul Evans thought that it might have been for additional testing but could not recall.^15411541 Evans {Day73/176:17-19}. He denied, however, that it was for lobbying.^15421542 Evans {Day73/176:20-23}. Despite its importance to Celotex, NHBC's approval of RS5000 for use on buildings over 18 metres in height was not forthcoming. By 7 January 2015, Celotex had learned through an email sent by a client (Nigel Shields of Durkan) to another Celotex area sales manager (Nigel Waring) that NHBC would not accept Celotex fire test data because it did not represent a true test of the product in all its applications. Worse still, NHBC was very nervous of its use on high-rise buildings and had taken steps to consult fire and rescue services.^15431543 {CEL00003445}. Although, as Mr Evans accepted, that was an indication that NHBC regarded RS5000 as dangerous,^15441544 Evans {Day73/127:17-25}-{Day73:128:1-3}. it did not cause Celotex to reconsider its marketing strategy.

Problems in the market

¶24.131 At the same time as NHBC was making its disquiet clear, difficult questions were being asked by clients about the breadth of application of Celotex's BR 135 classification. On 16 January 2015, Daniel Anketell-Jones of Harley sent an email to Jonathan Roome in relation to another project,^15451545 {CEL00000019/2}; Roome {Day70/63:14-18}. in which he explained that a customer had asked exactly how RS5000 had been installed when it was tested, who had carried out the testing, how it had been fixed, what it had been covered with, what rainscreen had been used, what support structure, and, most importantly, what had been the results. Mr Anketell-Jones described the request as a headache for Celotex, which indeed it was, because it was seeking detailed information that Celotex did not wish to make public.^15461546 Roome {Day70/61:16-22}.

¶24.132 The challenges faced by Celotex from the more astute section of the market were most clearly illustrated by a complaint made by Ardmore Construction Ltd in March 2015 after NHBC had rejected the use of RS5000 that had been installed on a high-rise building at Octavius Street in Deptford, London.^15471547 {CEL00002193}. Ardmore's Technical Director had complained about a lack of clarity in the marketing literature and asked for evidence of the fire testing on which Celotex relied to support its claim that RS5000 was suitable for use on buildings over 18 metres in height. Mr Evans told us that the complaint had been taken seriously at Celotex,^15481548 Evans {Day73/166:22-25}. but the marketing literature remained unchanged. He was unable to explain why Celotex had not been concerned that it was not clear enough to avoid a problem with a major developer and why there had not been discussions within Celotex about improving it.^15491549 Evans {Day73/165:3-9}; {Day73/167:8-15}.

152

LABC approval

¶24.133 Celotex also sought to obtain an LABC certificate for RS5000. It was seen as an important form of approval,^15501550 Roper {Day72/26:9-16}. in part because it would protect the use of RS5000 from the risk of challenge.^15511551 Roper {Day71/35:11-16}; {Day72/44:15-22}. It was also seen as a route by which Celotex could market RS5000 as being suitable for use in systems other than that tested.^15521552 {CEL00000735/4}; Roper {Day72/29:24}-{Day72/30:7}. Before the test in February 2014 Mr Roper had been in contact with David Ewing of the LABC, who was then responsible for managing the Registered Details scheme,^15531553 Ewing {Day217/95:18-22}. and had been told by him that as Kingspan K15 was described as Class 0, it could be described as a material of limited combustibility.^15541554 {CEL00000939/5}. Mr Roper realised that that was wrong but he did not correct the LABC's mistake because he thought it was advantageous to Celotex not to do so.^15551555 Roper {Day72/34:14-23}.

¶24.134 The first LABC certificate for RS5000 was issued on 21 August 2014.^15561556 {CEL00000010}. The accompanying drawing and document list stated that it was "for use in rainscreen wall construction including above 18m in height".^15571557 {CEL00000009}. The only test information provided was contained in the Advice Notes, which said that RS5000:

"had successfully tested to BS 8414:2 2005, meets the criteria in BR 135 and is therefore acceptable for use in buildings with storeys above 18m in height (subject to the board being fixed to a non-combustible substrate) alternative compliance to ADB".^15581558 {CEL00000009}.

¶That wording had been taken directly from an email Mr Roper had sent to Tim Bartlett of

¶West Suffolk building control, the person responsible for preparing Celotex's Registered

¶Details certificate, even to the point of including the transposition of a typographical error.^15591559 {CEL00001995}. However, the wording and the message it was intended to convey was untrue and misleading.^15601560 Roper {Day72/41:14-20}. The certificate, together with the document and drawing list, were sent by

¶Jonathan Roome to Daniel Anketell-Jones of Harley on 27 August 2014.^15611561 {CEL00000006}.

¶24.135 In early November 2014, Samantha Li of the LABC sent an email to Jonathan Roper explaining that the LABC was intending to change the form of its certificates and that the certificate for RS5000 would be reissued in the new style.^15621562 {CEL00002021/2}. Mr Ewing explained that he had decided to change the form of Registered Detail certificates because he thought that the previous version did not give enough information to local authorities considering whether to approve the use of materials.^15631563 Ewing {Day217/122:13-25}-{Day217/123:1-3}. He had been prompted to take that course by concerns raised in relation to the certificate relating to Kingspan's K15.^15641564 Ewing {Day217/123:16-20}.

153

¶24.136 The new certificate was attached to Ms Li's email. It stated that it had first been issued on 21 August 2014 and was valid until 21 August 2015.^15651565 {LABC0000312}. It now stated on the first page that RS5000 complied with BR 135 for use in rainscreen applications on buildings over 18 metres in height, but referred the reader to the conditions of the certificate for more information. They made clear that RS5000 was acceptable for use on such buildings "subject to matching the specification criteria of the BRE fire test report 295255 carried out".

¶24.137 Mr Roper, whose role as product manager of RS5000 had by this time passed to Debbie Berger, sent the email to Ms Berger on 3 November 2014, asking her to reply to the LABC with a request that it remove references to the fact RS5000 was the same product as FR5000.^15661566 {CEL00002021}. Ms Berger replied to Samantha Li on 6 November 2014, attaching some suggested text which removed the reference to the test report and repeated the claim that RS5000 complied with BR 135 "subject to the board being fixed to a non-combustible substrate". She also included a reference to a BBA certificate.^15671567 {CEL00002022}. Ms Berger accepted that that her proposed text had been misleading,^15681568 Berger {Day78/75:2-8}. as it suggested that the product could be used in any system with a non-combustible substrate and not merely the system that had been tested. Furthermore, it was not true that there was a BBA certificate relating to RS5000.^15691569 Berger {Day78/77:5-16}.

¶24.138 Later the same day Ms Berger sent Ms Li another email, observing again that the new certificate was available on the LABC website and asking for access to it to be restricted until they had sorted out the wording.^15701570 {CEL00008691/2}. Ms Li replied, saying that she would check with her manager to see if they could prevent the page from being available to the public. The LABC appears to have treated that as a request for the registration to be withdrawn.^15711571 Phase 2 Module 2 – Written closing submission on behalf of LABC {LABC0019740/28} page 28, paragraph 73. Celotex does not appear to have been told that the registration was being withdrawn, but we accept that it was removed from the LABC's website as a result of its request.

¶24.139 Six months later Mr Ewing wrote to Celotex again, referring to concerns about the wording of certificates^15721572 {CEL00001286}. and attaching a further revised certificate in draft.^15731573 {CEL00003741}. The revised certificate was finally issued in September 2015,^15741574 {CEL00000389}. long after the initial orders of RS5000 for Grenfell Tower had been placed.

Handover to Debbie Berger

¶24.140 Ms Berger took over the role of product manager with responsibility for RS5000 from Mr Roper on 1 October 2014.^15751575 Berger {Day78/6:15-17}. On 26 September she had a handover meeting with Mr Roper during which nothing was said about the use of 6mm magnesium oxide boards during the test in May,^15761576 Berger {Day78/27:7-12}. although the concerns about the test raised by NHBC were mentioned.^15771577 {CEL00002555/2}. Ms Berger discovered that the 6mm magnesium oxide boards had been used only later when she spoke to Mr Hayes on 24 October 2014 with a view to understanding the nature of the test in greater technical detail. His demeanour on that occasion led her to think that he had concerns about the test.^15781578 Berger {Day78/39:17-21}. On 27 October 2014,

154

¶Ms Berger obtained the full 32 page test report from Mr Roper^15791579 {CEL00009742}. and met Mr Hayes again, either the same day or some days later, to discuss it. Mr Hayes told Ms Berger about the

¶6mm magnesium oxide boards and using a diagram showed her where on the rig they had been placed.^15801580 {CEL00002686}; Berger {Day78/41:23}-{Day78/42:4}. Ms Berger also made a note on her copy of the report, next to a diagram of the system before the test, with the expression "WTF? 6mm MgO + 8mm Marley Eternit panel".^15811581 {CEL00008507/13}. She had clearly been shocked by what she had been told.^15821582 Berger {Day78/52:14-19}. Despite her shock, however, Ms Berger took no steps to alert her colleagues or anybody else to the fact that the report did not accurately describe the system tested.^15831583 Berger {Day78/61:19-23}.

Contact with BRE

¶24.141 On 8 October 2014 Jonathan Roome attended a CWCT conference,^15841584 {CEL00001255}. at which Dr Sarah Colwell of BRE was one of the speakers. Mr Roome met her and together they arranged a meeting between her and one of her colleagues to discuss, as Mr Roome put it, how Celotex might improve its approach to marketing RS5000.^15851585 {CEL00001255}. The meeting was planned for 4 November 2014.

¶24.142 A week later, on 17 October 2014, Stephen Howard of BRE sent an email to Mr Roper and Ms Berger referring to an earlier conversation about the content of the classification reports and the amount of technical detail they contained.^15861586 {CEL00002131}. Ms Berger sent the message on to Mr Roome on 20 October 2014, saying that she would set up a meeting with Mr Howard to discuss how the report on RS5000 could be structured to provide relevant commercial information without disclosing anything sensitive.^15871587 {CEL00002131}.

¶24.143 In the event, Mr Roome and Ms Berger met Mr Howard on 4 November 2014. Dr Colwell did not attend. Notes of the meeting prepared by Ms Berger refer to growing uncertainty in the market, which had been "burnt by K15 approvals in the past".^15881588 {CEL00001260}. That was a reference to an unhappy experience which NHBC had had in the latter months of 2013 and thereafter with the BBA certificate for K15 and its widespread use above 18 metres.^15891589 See Chapter 26 for the detailed analysis. Ms Berger noted that accreditation bodies required more information in order to give approvals and that further testing with a variety of facades was required.

155

February 2015: market analysis slideshow

¶24.144 For the purpose of a meeting of the Service Product Innovation Group in February 2015 Mr Roome prepared a series of slides identifying three tiers of potential customers. (We have described the slides and what they depicted in Part 6, Chapter 56 dealing with the way in which Celotex RS5000 insulation came to be chosen for use in the refurbishment.) He denied that their implication was that the more ignorant a customer was of the restriction on the use of combustible insulation on buildings over 18 metres in height and the less well-equipped to ask difficult technical questions, the more likely it was that Celotex would be able to sell it RS5000,^15901590 Roome {Day 69/81:23}-{Day69/82:4}. but it is difficult to see how else they can be understood.

¶24.145 As we have said, the slides reflected a deeply cynical view that there were ignorant or reckless contractors in the market of whom advantage could be taken without any regard for the safety of occupants.

¶24.146 A very different approach to knowledgeable third parties can be seen in an email sent by Jamie Hayes to Paul Evans, Rob Warren and Louise Garlick on 25 March 2015 with advice on what to say where NHBC or LABC were involved and the client's proposed construction differed from the system tested.^15911591 {CEL00003597} In the case of the LABC, potential customers were to be provided with a copy of the Compliance Guide and advised that, if they were using a substantially different design from the system tested, they should discuss the position with building control. If NHBC was involved, clients were to be provided with the Compliance Guide and told that Celotex did not support the use of RS5000 unless the design of the proposed system exactly matched the description given in its literature (although that did not in fact reflect the system that had been tested). Following discussions with Mr Evans, Mr Hayes subsequently promulgated an internal instruction at Celotex that in all cases where it provided a U-value calculation, and where it was asked whether a particular design complied with the Building Regulations, a copy of the Compliance Guide should be provided and customers told that it was the responsibility of building control to judge compliance.^15921592 {CEL00003604}.

Desktop study (March 2015)

¶24.147 In due course a meeting took place on 20 February 2015 between Frans Paap and Andrew Evans of Exova and Debbie Berger and Rob Warren of Celotex.^15931593 Berger {CEL00010055/57} page 57, paragraph 213. The notes of the meeting record that Exova needed to see the full test report and the Class 0 reports and refer to the need for a "case by case basis assessment".^15941594 {CEL00002691}. Ms Berger said that she had not understood that to mean that a separate desktop assessment would have to be made for each project.^15951595 Berger {Day78/144:21-25}.

¶24.148 On 18 March 2015 Ms Berger sent Mr Paap the full 32 page RS5000 test report. Her covering email set out four systems which she said were commonly presented to Celotex at the design stage and which she wanted Exova to assess: brickwork, a ventilated cavity with terracotta cladding, a ventilated cavity with A1 cladding and a ventilated cavity with a Class 0 aluminium rainscreen.^15961596 {CEL00003589}. However, as she ought to have realised, any desktop study based on the test report was bound to be misleading and dangerous because the report did not correctly describe the system tested.^15971597 Evans {Day73/154:21}-{Day73/155:1}.

156

Exova's draft report

¶24.149 Exova produced a draft report dated 7 May 2015.^15981598 {CEL00002040}. It set out the test evidence relied on, which had been taken from the full 32-page BRE Report.^15991599 {CEL00002040/5}. It did not refer to the 6mm magnesium oxide boards or the 8mm Marley Eternit boards because they did not appear in the report and Mr Paap had not been told about them.^16001600 Berger {Day78/149:13-24}.

¶24.150 The report concluded that all the designs set out in Ms Berger's email would pass, save for the ventilated cavity system with aluminium Class 0 cladding. In respect of any such system, Exova concluded that it could not be judged with certainty to meet the performance requirements in BR 135.^16011601 {CEL00002040/9}. The report was formally issued on 26 May 2015.^16021602 {CEL00001116}.

NHBC's reaction

¶24.151 The draft report was presented to NHBC at a meeting on 19 May 2015 attended by Debbie Berger, Paul Evans and Rob Warren for Celotex and Graham Perrior, Steve Evans, David White and John Lewis for NHBC.^16031603 Berger {CEL00010055/61} page 61, paragraph 230; {CEL00003682}. It was not well received by NHBC, which considered that its generic nature indicated a failure by Celotex to understand the purpose of desktop studies.^16041604 Berger {Day78/156:22}-{Day78/157:6}. The monthly report to the management action group described the meeting as "quite heated".^16051605 {CEL00003710/8}.

¶24.152 NHBC's objections were set out in greater detail in an email sent by Mr Perrior to Mr Evans on 8 June 2015, which was sent on by Mr Evans to Ms Berger and Mr Warren. In it Mr Perrior identified a number of shortcomings and concluded that it appeared that the basis of Exova's analysis did not reflect a true model of a ventilated facade or take fully into account the fact that a large part of the Marley rainscreen had caught fire.^16061606 {CEL00001122/2}. NHBC did not, of course, know that the description of the system in the Exova report contained material omissions; its concern was whether a system such as that described would ever be built in practice. Mr Evans told us that he had been concerned by the fact NHBC thought Celotex's position was untenable and that nothing, in effect, had changed in the year since their meeting on 19 June 2014. However, he intended to work with NHBC to understand what needed to be done to gain its acceptance.^16071607 Evans {Day73/195:19-25}. The report to the management action group acknowledged that at some point it would be necessary to consider another test to broaden the scope of use of RS5000.^16081608 {CEL00003710/8}. However, Celotex kept selling RS5000 and no further testing was undertaken before the fire at Grenfell Tower.

157