Boston Consulting Group tracked 1,700 global transformation programmes. Only around a third accelerated growth in a way that created genuine value, 70% fell short. That figure has barely shifted in years, even as organisations have invested considerably more.
The usual explanation is that transformation is hard, which is true, but it does not get you very far. A more useful observation is that most programmes are designed at the wrong level.
Where things tend to go wrong
The strategy is usually sound. The business case is reasonable. The technology choices are defensible. The board is supportive and then the programme meets the organisation and begins to slow.
Some of this is legacy technology. The UK government’s own State of Digital Government Review, produced by Bain and Company with input from over 500 leaders across 120 organisations found that 28% of central government IT systems are now classified as legacy, with maintenance costs in some cases tripling or quadrupling those of modern alternatives. A global survey of over 1,400 senior technology decision-makers by NTT Data found a consistent picture across the private sector, with 80% of organisations reporting that outdated technology is holding back their progress.
The harder problem is the one that cannot be put on a roadmap with a go-live date. Legacy structures and legacy thinking slow transformation just as much as legacy technology does and they are considerably more difficult to address.
What cybersecurity tells us
The KPMG Global Tech Report 2026 found that more than half of UK organisations plan to increase their cybersecurity budgets by over 10% in the next twelve months, a higher proportion than any other country surveyed. Only 13% say their capability is fully scaled today and 45% report hitting blocks even with funded strategies in place.
Analysis from Osborne Clarke found that two thirds of large businesses that experienced a significant cyber incident went on to experience another. The investment went in and the incidents continued. That points to gaps in governance, behaviour and accountability that tooling alone will not close. Cybersecurity is a useful lens here because the failures tend to be visible. The same pattern, though quieter, runs through AI adoption, data modernisation and most other programmes that depend on people doing things differently rather than simply having better technology.
The layer that tends to get missed
How roles change. How decisions get made differently. How the organisation actually operates once the new technology is in place. This is the part that gets deferred, then compressed, then treated as a communications exercise rather than the design problem it actually is.
The organisations making the most consistent progress tend to treat transformation as an organisational challenge as much as a technology one. They bring the same rigour to the operating model as they bring to the technical architecture and they look for people who can work across both, which turns out to be a harder search than most programme plans account for.