On a single weekend in April 2018, TSB moved about five million customers onto a new banking platform. The program had been years in the making and was expected to save the bank £160 million a year. By Monday, many customers had lost access to their accounts. Some saw money leave unexpectedly, and some could see other people’s accounts. Fraud attempts rose to 70 times their normal level.
The bank commissioned the law firm Slaughter and May to find out why. Its report ran to 262 pages. Several findings concern the decision itself and the way it reached the board. The platform was not ready for the bank’s full customer base. The “big bang” approach was chosen without adequately weighing its risks. Live testing was too small to reveal the problems that appeared at full launch. The supplier’s ability to deliver had not been thoroughly assessed. The board was not shown a full range of migration options and did not obtain independent advice on the program as a whole.
Every item on that list could have been raised before the weekend. Each one is a question a board can ask.
Why Large Technology Programs Surprise Their Sponsors
Bent Flyvbjerg and Alexander Budzier studied 1,471 IT projects. The average cost overrun was 27 percent, which looks manageable. The average conceals the distribution. About one in six projects was what the authors call a black swan, with a cost overrun averaging 200 percent and a schedule overrun of almost 70 percent.
Overruns in 1,471 IT projects, percent
All projects
27%
One in six (black swans)
200%
~70%
A leadership team that plans around the 27 percent average has no plan for the one in six. The right question for a major program asks what happens if it lands in the worst tail, and whether the organization can survive that outcome.
What Independent Judgment Adds
An independent adviser stands outside the delivery team, the supplier and the budget owner, and brings a single contribution: a clear view of the decision before resources are committed. In practice that means four things.
Business context. The decision is tied to the outcome leadership wants, the constraints it operates under and the people who must live with the result.
Evidence. Claims about readiness, cost and timeline are checked against what can be shown, such as test results, reference sites and supplier track records.
Dependencies. The critical links are named: data conversion, supplier capacity, regulatory timing, internal skills, the calendar.
Alternatives. Leadership sees more than one path, with the implications of each laid out before anyone commits. A phased migration might cost more in the plan and less in the worst case.
Five Questions Before Approving a Major Technology Program
- What are the other two or three ways to reach this goal, and why were they set aside?
- What is our plan if the program lands in the worst tenth of outcomes?
- What are the go-live criteria, who signs them, and what happens if they are not met?
- How did we assess the supplier’s capacity to deliver at our scale?
- Who outside the delivery team has reviewed the whole program, and what did they say?
These questions take an hour to ask. They cost far less than the report that follows a failure.
Good Decisions Leave a Record
The TSB review exists because the bank had to explain itself to customers, regulators and shareholders. Leaders can build that record in advance. Set down the options considered, the evidence reviewed, the risks accepted and the people who challenged the plan. A board that can show that record makes better decisions and stands on firmer ground if one goes wrong.