Blog

A treasure trove of facts that matter: why I built Reports that Matter

"Doubt is our product since it is the best means of competing with the 'body of fact' that exists in the mind of the general public. It is also the means of establishing a controversy." A Brown & Williamson strategy paper, quoted on p. 1537 of United States v. Philip Morris (2006), the judgment that found the American cigarette companies had deceived the public for at least 50 years. The same paragraph has the Tobacco Institute's procedural memo of 1968: "The most important type of story is that which casts doubt on the cause and effect theory of disease and smoking. . . . [T]he headline should strongly call out the point – Controversy! Contradiction! Other factors! Unknowns!" This is why the case matters: it is the inside record of how an industry handled science it privately accepted. The answer was not denial, which could be tested, but doubt, which could be manufactured, and here is the method being planned on paper decades before anyone outside saw it.

The 9/11 Commission (2004) was given more than 2.5 million pages of documents and more than 1,000 hours of audiotape, and from them it rebuilt the morning of 11 September 2001 to the minute. Its account of the response to the first hijacked flight by NEADS, NORAD's Northeast Air Defense Sector, the military post responsible for that airspace, ends on p. 21: "In summary, NEADS received notice of the hijacking nine minutes before it struck the North Tower. That nine minutes' notice before impact was the most the military would receive of any of the four hijackings." The question asked for years afterwards was whether the planes could have been stopped. The record answers it: the air defences had been built for an attack arriving from outside the country.

The Senate's financial-crisis report (2011) came out of a two-year investigation with subpoenas and tens of millions of pages of documents. On p. 297 it quotes a Standard & Poor's analyst in a December 2006 email about a security the agency had been asked to rate: "Let's hope we are all wealthy and retired by the time this house of cards falters." And an instant message from April 2007: "[W]e rate every deal[.] [I]t could be structured by cows and we would rate it." The crisis is usually told as a failure of complexity, of instruments nobody understood. The emails are the record of people who understood them well enough, saying so to each other at the time.

None of these is a conclusion. They are the evidence, reprinted, and it is the evidence rather than the verdict that tells you what actually happened. That is what the news has no room for. Reports that Matter puts it back: 24 inquiry reports so far, Challenger to Thirlwall, as web pages you can read, search and link to the sentence.

What these reports are

After a disaster or a scandal, a country sometimes appoints people with the power to compel evidence and asks them to find out what happened. The big inquiries take years and read hundreds of thousands of documents (the Philip Morris opinion alone runs to 1,682 pages, much of it the companies' own files), and then publish what they found with the evidence attached. What comes out is the most detailed factual account of that moment anyone has assembled. The Columbia Accident Investigation Board (2003) reprints the emails that NASA managers and engineers sent one another on Wednesday 22 January 2003, 10 days before Columbia broke apart on re-entry, with their times, as a request to have the shuttle's wing photographed in orbit was raised and then cancelled. On p. 157 it prints one that an engineer, Rodney Rocha, wrote that afternoon and did not send: "In my humble technical opinion, this is the wrong (and bordering on irresponsible) answer from the SSP and Orbiter not to request additional imaging help from any outside source." What the record shows is less a single mistake than a process: the engineers asked for pictures, the managers decided the question had been answered, and the engineer with the sharpest doubt kept it to a printout because, he told the Board, he did not want to jump the chain of command. You get that from the emails, in order, with their times, not from a summary.

Then the record disappears from view. A report gets one news day, and the news and the response to it are built from the executive summary, which is the analysis. The evidence, which in Leveson's case is most of 983,000 words, goes into a PDF, and the PDF sits on a website that is often wound up when the inquiry closes. If it survives, it survives in a web archive. The Iraq Inquiry's web edition of Chilcot is in the UK Government Web Archive. The CIA's web pages for the Duelfer report on Iraq's weapons no longer exist except in the Wayback Machine. And a PDF cannot be linked to by the paragraph, so even the people who do read these reports cannot easily point anyone else at what they found.

I find these reports absorbing, and I want other people to be able to read them as I do: whole, as the inquiry published them.

What we are doing

Reports that Matter puts these reports on the web as pages you can read, search and link to. Every paragraph has its own address, with its page number or the inquiry's own paragraph number. Select any sentence and you get a link to exactly those words, so when you tell someone what a report says, you can show them. The official PDF is one click from every page, and where our text and the PDF differ, the PDF is right; we say so, and we list what we have fixed. Every report's full text is also on GitHub as plain Markdown.

Getting a million words off a PDF with every footnote marker and heading where it belongs is its own story; the About page tells it, and the code is public.

A decade, on and off

I have wanted something like this since I was a teenager. South Africa's Truth and Reconciliation Commission published its report online, and anyone could read it. It showed me that putting the record where everyone can read it is part of how a country reckons with what happened to it.

In 2014 and 2015 I started building it, under the name Official Inquiries. It came from the same conviction that had got me to start the Open Knowledge Foundation: information a society has already paid for should be usable, not merely available. In 2016 and 2017 we built a prototype at Datopian. Then the funding ran out and the project stopped. My note from 2017 says "Official Inquiries => does not mean anything to most people". A later one just says "Needs funding and focus." For 7 years it had neither.

What changed is that the work a funded team was needed for can now be done by one person with very little time. One rule has not changed: no language model wrote, summarised or rewrote the report text. AI coding agents built much of the software and the site, under my direction, and the code is public. The short introduction at the top of each report is ours: drafted with AI help, checked by me, and labelled. Each report is still weeks of work: getting up to a million words off the page intact, then checking the text against the PDF and fixing what went wrong, a footnote marker read as a number, a heading gone missing, by hand and with the agents, until the report is fit to read. Nothing went up in an afternoon. In December 2024 I renamed the project Reports that Matter, and through 2025 and 2026 the reports went up one by one.

What is there

The reports are from the UK and the US, published between 1986 and 2026. The full list is at the end, and all of them are here.

Next are the later volumes of the reports that are only partly here, and whichever reports people keep asking for. If there is one you need, write to contact@reportsthatmatter.org. The Bluesky account posts one passage a day, verbatim, with its page and its link, and once a month the newsletter gathers a few passages from the archive on one subject.

https://reportsthatmatter.org

— Rufus Pollock

Appendix: the reports

In date order.

  • House Science Committee's Challenger report (1986). The US House of Representatives' own investigation of the Challenger accident, written by the committee that authorised and funded NASA. It agrees with the Rogers Commission on what failed but not on why, and names Congress among those who pushed NASA to fly too often.
  • Columbia Accident Investigation Board (2003). The independent board's account of why the Space Shuttle Columbia broke apart on re-entry in February 2003. It traces the loss to a piece of foam and then to a NASA that let foam strikes become routine.
  • 9/11 Commission (2004). The bipartisan Commission's account of the attacks of 11 September 2001 and of the failures across government that let them succeed, based on interviews with more than 1,200 people. Its central recommendations created the National Counterterrorism Center and the Director of National Intelligence.
  • Duelfer report (2004). The final report of the Iraq Survey Group, the team of some 1,750 people sent to find Iraq's weapons of mass destruction after the 2003 invasion. It asks what became of the weapons and why Saddam Hussein let the world believe he still had them.
  • United States v. Philip Morris (2006). Judge Gladys Kessler's ruling that the largest US cigarette companies had deceived the public about smoking for some fifty years, in violation of the federal racketeering law. It is a court's judgment, built from a nine-month trial and the companies' own internal documents.
  • Stern Review (2006). Sir Nicholas Stern's review for the UK Treasury, which set the costs of climate change against the costs of acting on it. It concluded that the benefits of strong, early action "considerably outweigh the costs".
  • Lehman Brothers examiner's report (2010). The court-appointed examiner's report on the collapse of Lehman Brothers in September 2008, the largest bankruptcy in US history. It sets out in full Repo 105, the accounting device Lehman used to move as much as $50 billion of assets off its balance sheet at quarter-end.
  • Saville Inquiry (2010). Lord Saville's report on Bloody Sunday in Derry/Londonderry on 30 January 1972, after a twelve-year inquiry into events an earlier tribunal had examined in eleven weeks. It found that none of those shot was posing a threat, and rejected the soldiers' claims that they had fired at gunmen or bombers.
  • Deepwater Horizon commission (2011). The presidential commission's account of the blowout of BP's Macondo well on 20 April 2010 and the largest accidental marine oil spill in US history. It traces the disaster to failures of management at BP, Halliburton and Transocean and to a regulator without the expertise to oversee them.
  • Senate financial-crisis report (2011). The Senate's two-year investigation into the causes of the 2008 financial crisis, built on the banks' own emails and documents. It shows how a bank, its regulator, the rating agencies and the investment banks each fed the collapse.
  • Hillsborough Independent Panel (2012). The Panel's account of the 1989 FA Cup semi-final disaster, in which 97 people died, and of what happened to the truth about it over the next 23 years. It examined over 450,000 pages of previously undisclosed documents and found that police statements were altered.
  • Leveson Inquiry (2012). Lord Justice Leveson's inquiry into the culture, practices and ethics of the British press, set up in 2011 after phone hacking at the News of the World. It concluded that decades of self-regulation had failed, and recommended a new independent regulator recognised in law.
  • Senate torture report (2014). The Senate Intelligence Committee's study of the CIA's detention and interrogation programme after 11 September 2001, drawn from more than six million pages of CIA records. It sets out what was done to 119 detainees and what the CIA told others about it.
  • Litvinenko Inquiry (2016). Sir Robert Owen's account of the poisoning in London in November 2006 of Alexander Litvinenko, a former Russian security officer, with a rare radioactive isotope. It concluded that the operation was probably approved by Russia's president.
  • Chilcot Report (2016). The Iraq Inquiry's own summary of why the United Kingdom joined the 2003 invasion of Iraq and what followed, after seven years of work. It concluded that the UK joined before the peaceful options had been exhausted and that military action was not a last resort.
  • 2017 Intelligence Community Assessment (2017). The declassified assessment by the CIA, the FBI and the NSA of Russian activities aimed at the 2016 US presidential election. It is the source of the judgments still argued over, with how confident each agency said it was.
  • Mueller Report (2019). Special Counsel Robert Mueller's report on Russia's interference in the 2016 US presidential election. Volume I covers the interference and the Trump Campaign's contacts with Russia; Volume II covers the President's conduct towards the investigation.
  • January 6th Committee (2022). The final report of the House select committee that investigated the attack on the US Capitol on 6 January 2021. It sets out its account of the efforts to overturn the 2020 presidential election, with its conclusions, criminal referrals and recommendations.
  • Grenfell Tower Inquiry, Phase 2 (2024). Volume 1 of 7 of the report on the fire at Grenfell Tower on 14 June 2017, in which 72 people died. Across the volumes the Panel sets out how the tower came to be clad in materials that let the fire spread, how the authorities responded, and its recommendations.
  • Jack Smith report (2025). Volume One of the Special Counsel's report on the case against Donald Trump over the 2020 election. The case was dropped after the 2024 election, so this is the evidence prosecutors would have put before a jury.
  • Post Office Horizon IT Inquiry (2025). Volume 1 of the inquiry into the Post Office's use of faulty accounting software to prosecute its own postmasters for losses that did not exist. It covers the human impact and redress; who was responsible is for later volumes.
  • Covid Inquiry, Module 2 (2025). The UK Covid-19 Inquiry's report on the decisions of the four UK governments from January 2020 to May 2022. Volume I gives its account of the lockdowns month by month; Volume II its analysis, lessons and 19 recommendations.
  • Southport Inquiry, Phase 1 (2026). Volume 1 of the first report of the Southport Inquiry, on how the agencies handled the risk the perpetrator posed before the attack of 29 July 2024. It sets out five fundamental problems and a chronology of the agencies' contacts with him from 2019.
  • Thirlwall Inquiry (2026). The report on events at the Countess of Chester Hospital from 2015 to 2018, in a summary and three volumes. It examines what the hospital's managers, the NHS and its regulators did when doctors raised concerns, and makes 17 recommendations.

Some are not whole yet, and each says so at the top. Chilcot is the executive summary; Lehman is Volumes 1 and 3; Saville, Columbia and Duelfer are Volume I; Jack Smith, Southport and the Post Office are Volume 1; Stern is the executive summary and selected chapters. The torture report is the declassified summary and findings; the full study of more than 6,700 pages is still classified.