What banks can learn about transformation from the defence and intelligence communities
Somewhere between 60 and 70 percent of bank transformation programmes fail. The figure has been remarkably stable for years, despite sustained investment in delivery, governance and reporting. Programmes still drift, still miss outcomes, still surprise their sponsors late.
The defence and intelligence communities have spent the last decade and a half wrestling with a similar problem: too many signals, not enough understanding of what they mean. Their answer is instructive for banks.
The intelligence community’s problem, in miniature
After 9/11, intelligence analysts found themselves with access to an enormous queue of signals. Phone activations, SIM registrations, flight bookings, money movements, border crossings. Millions of data points a day.
The challenge was not collecting data. It was finding meaning in it.
A single SIM activation tells you almost nothing. A flight booking tells you almost nothing. The signal lives in the relationships between these events, and in how those relationships evolve. A SIM activated in Hamburg at 3am is unremarkable. A SIM activated in Hamburg at 3am, belonging to someone with a flight booking to Boston the next morning, is a different proposition.
What intelligence communities learned was that data on its own is not enough. To act on it, you need context: a continuous view of how entities, events and behaviours connect and change over time.
The same problem, in banking
Banks have not been short of data for a long time. A typical Tier-1 transformation programme produces millions of signals across Jira, ServiceNow, GitHub, Workday, spreadsheets, change boards, risk forums and incident systems. Dashboards proliferate. KPIs, BI and MI multiply.
And yet senior leaders consistently report the same experience: even with an increase in reporting, confidence remains low. Escalations persist, interventions are made too late. Programme dashboards show everything is on track until, suddenly, it’s the opposite.
The reason is structural. Dashboards are snapshots of individual sources. They show what a team or a ticket looks like at a moment in time. They do not show how work, people and dependencies interact across the organisation over time.
Trying to understand a complex transformation from a portfolio dashboard is like trying to understand the plot of a film from a single frame.
The shift that matters
How did the intelligence community solve this? Not by collecting more data, or by building better dashboards on top of what they had. They solved it by recognising that data on its own was never going to be enough. They needed to change the lens through which they looked at it.
Instead of viewing signals through relational databases — rows, columns, individual records — they began looking at them as a network.
They mapped connections between data points — phones, people, passports, flights — and watched what happened between them over time. Snapshots became a live feed. Isolated signals became patterns. The question changed from “what does this record say?” to “what is this entity doing, and how is that changing?”
Banking transformation needs the same shift. Leaders need to see not just where individual projects stand, but how the delivery system itself is behaving: where dependencies are accumulating, where coordination is breaking down, where a delivery risk is forming. And they need to see these things early enough to act on — not days before two critical path dependencies collapse at once.
Without this, organisations remain in the position they have been in for years: well governed on paper, structurally fragile underneath, and surprised by failures that were forming for months in plain sight.
The intelligence community solved this in the decade after 2001. The question for banks isn’t whether to make the same shift. It’s how much longer they can afford not to.
If you want to see what this looks like applied to a Tier-1 transformation programme, get in touch.