The Great Post Office Scandal

The candidates were whittled down to a shortlist of three. In May 1996, the winner was announced. Pathway, a consortium owned almost entirely by a company called International Computers Ltd – better known as ICL – got the gig. ICL once saw itself as Britain’s answer to the American computing giant IBM, but its glory days were long gone. In 1990, the Japanese tech corp Fujitsu had taken an 80% stake in ICL with a view to outright ownership. Whilst Pathway was being tendered, Fujitsu was in the process of swallowing ICL whole.

Pathway managed to come bottom in eight of the 11 scoring criteria drawn up to quantify the strengths of each bid, but it was the cheapest in terms of its cost to the government. Fujitsu’s business model for Pathway saw it shouldering almost all the project’s commercial risk by paying for its development and rollout.

In return Pathway would get a transaction fee every time a shiny new Benefits Agency card was swiped through a shiny new Post Office card-swiper, for a guaranteed period of eight years. If all went according to plan, Project Pathway would deliver automation of the Post Office and vast sums of cash to Fujitsu, without any risk to the public purse. As an incentive to get things operational as soon as possible, the eight year guaranteed transaction fee period would start the moment the PFI deal was signed.

Almost as soon as the ink on the contract was dry, Pathway ran into difficulties. Serious, big-picture difficulties. It seemed everyone had badly under estimated the complexity of automating 20,000 geographically isolated and disparate Post Offices carrying out multiple (and sometimes region-specific²) transactions which were many times more complicated than the average bank.

The logistics of building, testing and installing a bespoke IT system of this size was uncharted territory for the government, and Fujitsu. 20,000 Post Offices would need 40,000 terminals. 67,000 people with variable technical skills would have to be trained to successfully keep handing out £56 billion in benefits alone to their 28 million annual customers.

The hardware would have to be robust. The software would have to process millions of transactions a year. Training and support would have to be first rate. Fujitsu had bid for IT’s Operation Moonshot by undercutting the competition. Having won the contract, they were realising how complex and expensive moonshots can be.

Conceptually, Horizon was relatively straightforward. The front-end terminals would be custom-specced PCs running Microsoft’s Windows NT operating system. These would sit in a box under each Post Office counter window. Each PC would be connected to a custom key-board, a barcode scanner, a 3.5 inch receipts printer and a touchscreen, which would sit on top of the counter. Running bespoke Horizon software, each counter PC would write both manually-and barcode-inputted transactions into the branch accounts. Every night all the information collected from the branch would be uploaded via ISDN or (in some cases) satellite to a centrally-located Post Office mainframe.

That was the theory. Writing the code to make this system work reliably appeared to be completely beyond Fujitsu. Despite burning through £10m a month, by the middle of 1997, Fujitsu was not able to deliver an (already renegotiated) contractual obligation to demonstrate Horizon’s ‘satisfactory, sustained’ operation. Even in a controlled environment, they simply could not get it to work.

Fujitsu was not helped by its clients, two state-owned entities with different and sometimes competing priorities.

Goalposts for the project were constantly being moved. Fujitsu would later tell parliament that over three years, Pathway’s 366 original contractual requirements had been subject to 323 formal change requests by the Post Office and the Benefits Agency. This was exacerbated, it said, by numerous informal changes and ‘clarifications’ which necessitated ‘significant system re-design.’

Frustrated with the delays, in December 1997, the Benefits Agency and the Post Office served Fujitsu with a formal notice of breach of contract. Fujitsu, fed up with the way it was being treated, refused to accept the notice. It then upped the ante by threatening to stop work on the Horizon project altogether unless it received a guaranteed hike in the fee it was going to get from each swipe of the DSS benefits card.

The stand-off turned critical when the Benefits Agency responded by announcing it no longer wished to use swipe-cards to dispense benefits payments. Instead, it was going to encourage claimants to have their benefits delivered direct to their bank accounts. In one swoop, Fujitsu’s potential income from the project had been reduced to zero. On top of that, the Post Office was looking at losing a third of its customer base overnight.

In January 1998, serious consideration was given to abandoning Horizon altogether, but the project limped on. Despite a shifting brief and its lack of operability, there was a feeling in Whitehall that Horizon had become too expensive to kill. Someone was going to have to find a way through, because the fallout caused by cancelling the project would be politically disastrous.

Richard [Roll, a well-trained and experienced IT engineer in the RAF and an ex-Fujitsu UK employee who was a whistleblower that aided the case] found it a tough – at times punishing – environment [at Fujitsu UK, Horizon creator and owner]. The scale and complexity of Horizon and its inherent shakiness meant there was a lot to do. Spotting and fixing problems before they triggered a fine under the service level agreement Fujitsu had with the Post Office¹ could save thousands, sometimes tens of thousands of pounds. The system of financial penalties for service breakdowns meant Fuitsu did not have an incentive to keep the Post Office fully informed about every problem as it manifested itself.

Richard’s team was under pressure. When he worked there, Fujitsu UK was not the all-conquering tech behemoth it is today. Without the Horizon contract, Richard felt that Fujitsu UK might be in deep trouble. The relationship with the Post Office therefore had to be protected at all costs.

And this is what rankled. The Subpostmasters who were calling in the problems and whose livelihoods depended on Horizon functioning properly, did not appear to feature much in Fujitsu’s corporate thinking. The Post Office was the client, not the Subpostmasters, and Horizon was the golden goose. So long as Horizon gave the appearance of functioning as it should, the client was going to be happy. What they didn’t know, couldn’t hurt them.

The other side of the service level agreement was that whilst the Post Office were entitled to (sometimes large) payments for service interruptions, if they wanted information from Fujitsu, it would cost them money. They had a certain quota of requests which came with the contract but any requests above that number triggered a fee. Subpostmasters with a discrepancy were not entitled to any investigation. The Post Office could ask Fujitsu to investigate a branch discrepancy if it wanted to, but a finite number of ‘free’ requests made them reluctant unless it was in their interests to do so.

Even if the Post Office did ask Fujitsu to investigate a problem, Fujitsu had no imperative or incentive to find one. As the courts would later find, this meant bugs and errors could be misdiagnosed, ignored or quietly fixed without the Post Office or Subpostmasters being any wiser.
The contractual environment created a dangerous situation. It was cheaper and easier for all concerned to blame a Sub-postmaster (and then go after them for a discrepancy with threats), rather than spend time and money digging into allegedly dodgy code.

Richard eventually got out, fed up dealing with angry and upset Subpostmasters, and fed up with Fujitsu’s work culture. He retrained as a chiropractor and set up a practice in Reading. On 7 February 2011, as Richard sat eating his dinner, he watched Seema, Jo and David’s story being told, and he resolved to do something.

The other downside to an insular culture, over-promotion and a lack of fresh thinking was the Post Office employees’ emotional and financial dependency on ‘the business.’ There were plenty of people within the organisation whose identity was bound up in the Post Office’s own self-image. They saw the Post Office as virtuous, so they were virtuous for working at the Post Office. Criticism of the Post Office was therefore a direct attack on their own personal integrity.

Colin [not his real name] said it was strange to be around. ‘The main thing everyone said at the Post Office, was that, “We are Britain’s most trusted organisation.”’ At senior management level, Colin found this obsession both weird and dangerous.

He described it to me as ‘a sort of lifebuoy which became a cage. It stops you from making sensible decisions. It leads to cultural blindness.’

Colin had seen media reports about
Horizon and the problems Subpostmasters had been having, but once he was on the inside, he realised that bringing it up was not the done thing.

‘It was a really strange situation. It was one of those things where everyone knew what was going on, but very few people spoke about it.’

[...a chapter or so later...]

Colin was working with the team organising the public response to the report [about Post Office processes]. ‘At the time there was a complete belief that there could be no findings other than what was expected, which was that there were no issues with Horizon.’

By this stage Colin was feeling really uncomfortable with the ‘official version of the truth’ and the whole approach around Horizon. He began to ask questions. Shortly afterwards another red flag went up. Colin was having a casual conversation with a senior member of the Post Office operations team. As they were chatting, Colin asked him if he thought it was a coincidence so many of their Subpostmasters were criminals. The ops team member told Colin that Subpostmasters were a specific class of people: ‘failed coppers and retired publicans’ who were ‘all on the take!’

Colin felt this glib reply was the ‘mentality’ which gave people at the Post Office permission to believe what they needed to believe to carry on in the way they did.

[...and then a few more chapters later...]

He [Colin] remembers the incredulity amongst his colleagues when one of the first Case Review Reports into an applicant’s case came through. ‘It was clear Second Sight believed there was an issue and there could have been an unsafe conviction. No one on our side could believe it. Everyone was like ... “What?!” Then they produced a couple more. And they were coming up the same!’

The internal reaction was one of deep concern. Not for the Subpostmasters, naturally, but for the Post Office’s reputation.

Colin watched this with interest. ‘What was crazy about the Post Office was that there was almost this mentality of refusing to admit that there is a problem. It’s almost like an alcoholic being in a denial that they are an alcoholic. It really was that sort of mentality.’

Things took a darker turn. Instead of confronting the issue, the Post Office retreated further into corporate denial, and got nasty.

Colin says, ‘It became clear we needed to “switch” and start discrediting Second Sight. And that’s what began to happen. Post Office began to discredit the people we had hired by saying things like ... they’re “too old”, or it’s been too long since they did a job like this, or they’re not up to the job. And that was purely because the Post Office did not like the conclusions that Second Sight were coming to.’