Conceived in 1996 as one of the first private finance initiative contracts, between the Post Office and the Benefits Agency on the one hand and computer company ICL on the other, the Horizon IT system had an unpromising start. It had been set up to create a swipe card system for payment of pensions and benefits from Post Office branch counters.
When, in May 1999, the plug was finally pulled on what the Commons public accounts committee called 'one of the biggest IT failures in the public sector', taxpayers had lost around £700m. Something had to be salvaged, however.
So, against the better judgement of its IT specialists, the Post Office decided to use the system to transform its paper-based branch accounting into an electronic system covering the full range of Post Office services. The new Horizon project became the largest non-military IT contract in Europe.
And, though I can't find the reference today, I have read that considerable pressure was put on the Post Office. Its executives were told they could whistle for more government investment if they didn't buy Horizon.
All of which reminds me of Roger James, who wrote a book seeking to apply the insights of Karl Popper's philosophy to public affairs. That 1980 book, Return to Reason, can be found online.
One of the useful concepts James introduces is 'solutioneering', which he characterises as:
Jumping to a solution before clearly formulating what the problem is (or indeed if there is one at all) or how success or failure are to be judged. Achievement of the solution then becomes the goal; and, when opposition develops, the problem becomes how to get the solution accepted, while the question of how best to solve the original problem, if there was one, never gets discussed at all. I call this mistake solutioneering.
Horizon wasn't a solution to one, clearly defined problem, but the Post Office was nevertheless determined to defend its reputation at all costs. Anyone who raised doubts about whether it could do all that was being asked of it was treated as a threat to the organisation.
Thank you for the link to Return to Reason, I'm going to read the book.
ReplyDeleteI can't find direct evidence either, but the reason given for adopting Horizon was due to pressure from the Japanese government, they threatened to pull out of japanese investments in the UK. The government of the day wanted to be seen as creating more jobs, some of these jobs were being created by japanese investments.
The scandal does indeed go all the way to government level, but also being arm-twisted by a foreign government. Global inter-dependence not working out for the common man.
To explain - ICL was 80% owned by Fujitsu at the time. It didn't rebrand until 2000 or so.
ReplyDeleteI don't know how Return to Reason will stand up after all these years. For some reason there was a copy in the university library at York when I studied there.
ReplyDeleteThere are even those who trace the Horizon scandal back to the establishment of ICL by Tony Benn, in his days as the arch technocrat Anthony Wedgwood Benn. He hoped it would be a worldwide competitor for IBM, but in the event it became something of an in-house supplier for the British government.
Going back to the 1960s and ICL as a world-wide competitor to IBM, it always was a mish-mash. In the late 1960s after leaving uni, my first job was with ICL at Kidsgrove (the old English Electric Leo Marconi) and their System 4 computers were essentially UK hardware equivalents of the IBM 360 series. The old ICT 1900 series were very different in hardware design. As far as the customer is concerned all that matters is that they can run their programs written in high-level languages on the new system with the minimum of modification and an expectation of few bugs (no computer system is bug-free and I had experience of bugs in the Fortran compiler that manifested themselves on the 4-70s at Risley, but which we could not reproduce on our 4-70 at Kidsgrove).
ReplyDeleteThe management of ICL made the decision to end development of System 4 and the 1900 series and concentrate on developing the new 2900 series which was deliberately designed not to be compatible with either of the previous two designs. Whilst this may have made it "better" in a computer science sense it hardly encouraged IBM users to switch.
See Wikipedia for more details https://en.wikipedia.org/wiki/ICL_2900_Series