The Real Data Transformation Challenge is Changing How People Work

For years, organisations have treated data transformation as a technology project. Invest in a new platform, consolidate data, build dashboards, and give employees access to better information.

While the technology certainly matters, increasingly it is not the hardest part. The real challenge is getting people across an organisation to define and use data differently in the decisions they make every day.

Much of that work happens far away from the data platform, in business processes and individual judgement. That distinction matters because a data platform can be delivered successfully while data transformation still fails to deliver the expected business value.

The evidence is increasingly difficult to ignore. The 2026 AI & Data Leadership Executive Benchmark Survey found that 93.2% of senior data and AI executives from nearly 110 Fortune 1000 companies and leading global organisations identified cultural challenges and change management as the principal barriers to data and AI adoption, compared with just 6.8% citing technology limitations.

That does not mean technology is irrelevant. Rather, it suggests many organisations have moved beyond the point where another platform alone will solve their data problems.

Technology is only the beginning

An organisation can spend heavily on data infrastructure while adoption remains patchy and different teams continue to use competing definitions and measures. Dashboards multiply, but important decisions are still made according to habit or instinct. 

The problem is that data transformation is fundamentally different from implementing a conventional business application. With an application, there is usually a clearly defined thing to adopt, whereas, data transformation asks people to change behaviours that may have developed over many years. 

Employees need to treat the data they create as something other people depend on, rather than as a local by-product. They need to manage quality as part of their work instead of expecting problems to be fixed downstream. They need to take ownership of definitions and data lifecycles, apply governance during delivery, and challenge numbers that do not appear credible.

This is why measuring transformation through logins, dashboard views, or software licences can be misleading. Those measures demonstrate technology usage, however they don't necessarily demonstrate changed business behaviour.

The OECD makes a similar distinction in its work on data-driven organisations and government. Having strategies, mandates, standards and governance arrangements does not necessarily translate into systematic data use, effective quality management or routine monitoring.

Designing behaviour into the transformation

If the objective is changed behaviour, organisations need to design the environment in which that behaviour occurs.

The first requirement is simplicity. The governed way of working must be easier than the workaround it replaces. That could mean consistent definitions, role-specific views, or putting relevant insight directly into the tools employees already use. The aim should not be to hand people more data, but to make the right action easier to take.

The second is reinforcement, as training is rarely enough to change established habits. People need prompts at the point where behaviour matters: agreed metrics in a meeting pack or an alert when a threshold is reached. The design focus should be the decision event and not the training session.

Trust is equally important as employees will not consistently act on information they do not trust. So declaring data to be 'trusted' is not enough. People need to understand its definition, ownership, lineage, limitations and how it is used to create insights.

The fourth ingredient is making data-led behaviour normal, and this needs to start at the top. When executives routinely ask what evidence supports a decision, employees receive a much stronger signal than they do from transformation communications.

Gartner's research reinforces the point, identifying organisational and human issues alongside technology as major challenges for data and analytics leaders. Its 2024 research also found that 61% of organisations were at that time changing their data and analytics operating models in response to artificial intelligence, with governance central to that evolution.

From change management to behaviour change

None of this makes conventional change management obsolete, but instead moves where change management begins. Rather than beginning with a technology rollout, organisations should begin with a business decision they want to improve. 

They should look carefully at who needs to behave differently, what evidence would make the information credible, and how will the organisation know that behaviour has actually changed. The answers can then feed directly into technology requirements, process design, leadership actions, and measures of success. Behaviour design should not become another isolated workstream but rather should inform the design of the transformation itself. 

For business leaders, the test should involve asking these key set of questions:

• Can an organisation identify the specific behaviours that must change? 
• Does it know which decisions require those behaviours? 
• Is the governed approach genuinely easier than the workaround? 
• Can employees see why they should trust the information? 
• Are leaders visibly demonstrating the same behaviours?

A data platform is delivered when the technology works. A data transformation is realised when people repeatedly work and make decisions differently because of it.