By now most of us have heard about the catastrophic attempt by the Spanish-owned TSB to introduce a new IT platform for their UK customers.
As my first mortgage was with the TSB many years ago, and I was also in the U.K. when the story broke I took a little more interest.
TSB, (Trustee Savings Bank), merged with and was spun out of Lloyds Bank after the EU ruled that it was a monopoly, because of the state aid it had received at the time of the banking crisis. TSB used Lloyds IT at a cost of about £220 million a year, but later moved to the Proteo platform, also used by its new owners, Sabadell. The Proteo system design goes back to 2000 and was specifically for mergers, and was used for successful integration of the four Spanish banks.
Proteo is based on Accenture’s Cobol-based Alnova system, and is customized, installed and managed by TSBs staff and runs on Amazon Cloud.
At the launch of Proteo4UK, Paul Pester, CEO of TSB, boasted that they had “created a more digital, agile and flexible TSB”. Carlos Abarca, the CIO, agreed, “It’s the technology journey that we are on together with our customers!” Similar ‘digital transformation’ good news messages from cloud providers are all too familiar.
This was to be “customer-centric by design” platform to “enable the open banking revolution”.
Well there was a revolution alright – from the locking nearly two million banking customers out of their accounts for up to ten days.
This was over a month-end, when businesses rely more heavily on access to their accounts.
TSB turned to IBM, to help get the system under control and “to help identify and resolve performance issues in the platform”. This included customers : experiencing zero balances, incorrect currencies, massively inflated mortgage amounts, and e-mails saying that there are no records of recent direct debits. IBank customers puzzled over on-screen messages, such as: ‘BeanCreationNotAllowedException exception: Error creating bean with name ‘contextManagerPostController’: Singleton bean creation not allowed while the singletons of this factory are in destruction (Do not request a bean from a BeanFactory in a destroy method implementation!)’
Customers who tried to make transfers got errors like: ‘ArrayIndexOutofBounds’ and java.lang.NullPointer and some Branches reported the systems spewing out error messages in Spanish. When I travelled back form U.K. early May, problems with internet banking wee still being reported by customers.
Instead of saving TSB over £100 million a year, this has greatly reduced public confidence in the bank but also in other banks and other financial services on the cloud generally. TSB are likely to be fined by the Financial Conduct Authority (FCA) and the Information Commissioner’s Office (ICO), which is the last thing they need to get things sorted – the loss in reputation alone is huge.
Supporters of Peter Pester believe he is being scapegoated for tech disaster:
• Allies said he has been ‘betrayed’ by a ‘bunch of Spanish numpties’
• Software that caused the problems was installed by Sabadell’s technology offshoot, Sabis
• Regulators could put TSB under a Section 166 probe – a formal investigation by an independent expert
What are the lessons?
Well, the first is not to claim success until the job is done. ( A damning report on the Guardian website suggests there were plenty of warning signs, up to a year before this all happened. Quoting an anonymous insider, the report explains how a mixture of poor technical and business decisions led to the eventual crises TSB finds itself in today.)
Which leads to the second lesson- bearers of bad news may have appoint to consider and is a hint at least the challenge needs more attention.
It seems Sabadell, the company that bought TSB, was warned about the high risk of its migration plans, which were seen by some as having too short a deadline and not big enough a budget. But Sabadell was not to be discouraged, and it pressed ahead with its plans, confident that it could successfully transfer TSB customers to its own Proteo software, as it had done with other customers in the past.
If you are doing some thing big and complicated consider the worst case and what that means for: insurance, contingency plans, contractual and legal protection, (so far none of the original contractors on the TSB redesign and upgrade have acknowledged any culpability) and PR mitigation:
PR week called it the flop of the month …….. and recipe for reputational disaster. Pester is well respected in the industry, but took too long to accept responsibility, was too quick to assume the problem was over, and too slow to appease customers. Easy to say from an armchair in Dubai but why do corporate leaders fail to heed the lessons of the past and to recognise the potential for disaster and that that when disaster arrives the only way to avoid reputational damage is to offer maximum compensation and care and to call in reinforcements asap.
Sabis is understood to have given TSB a written assurance that the parts of the system for which they were responsible had been comprehensively tested- maybe TSB needed to be more involved in those tests.
