Why Data Strategies Stall After Approval
The strategy was approved months ago and little has moved. Before asking for more money or a rewrite, work out where it stopped. There are four places to look, and they should be checked in order.
A stalled data strategy does not announce itself. Nobody withdraws support. The steering group still meets, the document is still cited, and the status report still shows amber. What has stopped is the part that costs something: funds have not been released, people have not been reassigned, and the decisions the strategy called for have not been taken.
The temptation at that point is to rewrite the strategy or to ask for more money. Either may be right, and neither should come first. A stall has a cause, the cause can be found, and the remedy depends on which one it is. This article offers a diagnostic in four questions, to be asked in order, with what a bad answer looks like and the smallest change that restarts movement.
It is written as analysis of published evidence. It does not draw on client engagements, and the examples of what a reader might observe are illustrative.
What the evidence shows, and what it does not
There is a figure in wide circulation for the proportion of data strategies that fail. This article does not use it. No primary source for such a figure was found and read in the research for this piece, and a number that cannot be traced should not be repeated.
What can be cited is narrower and more useful. The UK National Audit Office examined data across central government in 2019 and found that “developing data strategies has not been a priority for departments”. Of the seven large departments it looked at, five had a data strategy, of varying maturity, and it reported that “in one department the strategy has been approved but no additional funding has been agreed to implement it”1. That sentence describes the condition this article is about, recorded by an auditor.
The government’s own later account is franker still. Its 2022 roadmap for digital and data says of earlier efforts that “previous strategies have lacked specificity, cross-government endorsement, clear lines of accountability and business ownership”, and that subsequently “former flagship programmes have slowly shut down and failed to deliver results”2.
These are findings about UK central government. They are used here because they are public, specific and independently audited, and because the four weaknesses the roadmap lists are not peculiar to the public sector. They do not establish how often strategies stall elsewhere.
Question one: does anyone with a budget own it?
Start here, because the other three questions cannot be fixed without it.
A strategy has an owner when one named person is answerable for its delivery and controls, or can obtain, the money and people it needs. A sponsor who endorsed it at the board, or a data leader who wrote it and holds none of the budgets it draws on, is something less.
What a weak answer looks like. The strategy is described as owned by a committee. Its named owner cannot move funds between lines. Requests go to several budget holders separately, each of whom treats the request as one bid among many. When money is short, data work is the first thing deferred. The National Audit Office saw the pattern across government: “Data projects have sometimes been set aside when funding is under pressure”1.
The 2023 audit of the UK roadmap explains why the structure matters. The central team that led the roadmap could not deliver it on behalf of departments, so its success “depends on the engagement of senior departmental officials who have the required authority”3. Authority over the resources sat somewhere other than responsibility for the plan. The auditors also recorded the risk that follows, that support from leaders outside the digital function would “not be sustained without continued and proactive involvement” from the most senior officials3.
The smallest move. Do not ask for a new governance structure. Ask for one decision: a named executive who is accountable for the strategy’s outcomes and has the delivery budget, or the standing to direct those who do. The UK roadmap’s answer was to place “ultimate responsibility for delivery” with the most senior official in each department, and to give each of its missions a sponsor at that level2. An organisation that cannot name such a person has learned that the strategy was endorsed and not adopted, which is worth knowing before anything else is attempted.
The same weakness appears in governance programmes, with roles in place of budgets. Why data governance programmes lose momentum after launch traces that version.
Question two: is it in the operating plan?
A strategy can have an owner and still sit beside the organisation’s real plan.
The operating plan is where the year’s commitments, budgets and objectives are fixed. A data strategy that is not reflected there competes with everything that is. The UK government’s maturity framework describes the weakest position as one where data strategy is linked to organisational strategy only for external reporting and “any connections are not acted on internally”4.
What a weak answer looks like. The data strategy has its own launch, its own governance and its own reporting, and is absent from the documents that business units are held to. Managers asked about it say that they support it and that it is not in their objectives. Initiatives are described by the capability they build and not by the outcome they serve, so no business owner is waiting for them. The framework’s lowest level puts it concisely: data policies, principles and strategy are built “in a disconnected way”4.
The smallest move. Take the two or three commitments in the strategy that matter most and get each one written into the plan of the business unit that benefits, with that unit’s executive as owner of the result. This is a narrower task than re-aligning the whole strategy, and it can be done within one planning cycle. It also tests the strategy. If no business unit will accept a commitment as its own, the outcome it serves was never agreed, and the method in how to align a data strategy to business outcomes needs to be applied before more is spent.
One pitfall should be avoided. Attaching a data initiative to a business plan in name only, with the data function still doing all the work and carrying all the risk, satisfies the paperwork and changes nothing.
Question three: who is paying for the foundations?
Strategies that pass the first two questions can still stop at the point where shared foundations are needed.
Foundations are the things several uses depend on and no single use can justify: common identifiers, agreed definitions, ownership of shared data, a platform that is operated as well as built. Each business sponsor will fund the use they want. None volunteers for the part that benefits everyone.
What a weak answer looks like. Visible projects have gone ahead and each has built its own workaround. The same data is cleaned repeatedly by different teams. A platform was delivered and has no operating budget. Foundation work appears in every version of the plan and is always in next year’s column.
The audit evidence identifies one cause precisely. Departments told the National Audit Office that “it is easier to bid for capital funding for new developments than resource funding to maintain existing services and keep them up to date”3. New things attract money, and the continuing cost of making them work together does not. The same report notes of one cross-government service problem that “there is often no end-to-end ownership of services”, with different people responsible for different parts3, which is the organisational form of the same gap.
The smallest move. Stop presenting foundations as a programme in their own right. Tie each one to the first funded use that needs it, scope it to what that use requires, and put its running cost into the sponsor’s case. A foundation attached to a wanted result has an advocate. From data strategy to roadmap sets out that sequencing in detail, and running a data platform as a product covers the standing team and funding a platform needs once it is live.
Foundations built this way arrive piecemeal and may sacrifice some of the efficiency of designing the whole foundation at once. The trade-off is that each increment is attached to a funded result rather than waiting for a standalone foundation programme to be approved. The justification is that the single programme, in the conditions described here, does not get funded at all.
Question four: what do the measures reward?
The last question is about what gets reported, because a strategy can appear to be moving while nothing changes.
Activity is easy to count: frameworks published, standards agreed, people trained, datasets catalogued. Outcomes are harder: a decision made faster, a cost avoided, a report that can be trusted. A strategy measured on the first kind can report steady progress and deliver none of what it was approved for.
What a weak answer looks like. Status reports list completed deliverables and no change in any business measure. Milestones are met and the benefits case has not been revisited since approval. The people who were meant to gain from the work cannot say what is different.
The 2023 audit contains a careful example. On the roadmap’s data mission, it reported that “most work has been on developing frameworks and definitions and other data initiatives”, intended to be brought into use from the following year, while on the roadmap as a whole early milestones had been met and “existing risks have become more acute”3. Both statements were true together. That is not a criticism of the work, which had to come first. It shows how a programme can meet every milestone in a period and still be further from its outcome than the report suggests.
The smallest move. For each major commitment, add one measure that belongs to the business and would move if the work succeeded, and report it beside the milestones. Keep the activity measures; they show whether the work is being done. The outcome measure shows whether it was worth doing. The Federal Chief Data Officers Council’s playbook asks data leaders to “continuously communicate the value and measured progress” of the strategy to stakeholders5, and warns what a data function becomes without that focus on value: the chief data officer “faces the risk of becoming a compliance driven position”5.
Reading the four answers together
The questions are ordered because the remedies depend on each other. A strategy with no accountable owner cannot be written into an operating plan, since nobody has the standing to negotiate it. A strategy outside the operating plan cannot get foundations funded, since no sponsor is attached to a result. And outcome measures mean little when no business owner is waiting for the outcome.
Where a data strategy has stalled: a quick reading
| If this is true | Look first at | The first thing to ask for |
|---|---|---|
| Funds are approved in principle and not released | Ownership | One accountable executive with the delivery budget |
| Business units support the strategy and have no objectives for it | The operating plan | Two or three commitments written into business plans |
| Projects deliver and the same data problems recur | Foundations | Each foundation tied to a funded use, with its running cost |
| Milestones are met and nobody can name a result | Measures | One business-owned outcome measure per commitment |
The table is a starting point. More than one condition may hold, and where several do, begin with the earliest in the order.
When a rewrite is the right answer
Sometimes the diagnosis is that the strategy itself is the problem. Three findings point that way: the outcomes it serves are no longer the organisation’s priorities; it made no choices, so there is nothing to hold anyone to; or it committed to a scale of change the organisation was never going to fund.
In those cases, write a shorter strategy. Fewer outcomes, fewer commitments, each with an owner and a cost. What a strategy has to decide in order to be usable is set out in what a data strategy is and the decisions it has to make. It may also help to re-establish the starting point with the Data Strategy Maturity Assessment, so that the next version is sized to what the organisation can do.
Where the strategy is sound and one of the four questions has a weak answer, a rewrite is an expensive way of avoiding a conversation. The plan can be repaired where it stands, and the Data Strategy Roadmap gives a structure for re-planning the committed work without restarting the whole.
Sources
- National Audit Office (UK). Challenges in using data across government (HC 2220). 21 June 2019.
- Central Digital and Data Office (UK Government). Transforming for a digital future: 2022 to 2025 roadmap for digital and data (updated September 2023). 9 June 2022.
- National Audit Office (UK). Digital transformation in government: addressing the barriers to efficiency (HC 1171). 10 March 2023.
- Central Digital and Data Office (UK Government). Data Maturity Assessment for Government: Framework. 27 March 2023.
- Federal Chief Data Officers Council (United States Government). CDO Playbook: Advancing the Federal Data Strategy. 2021.
This article is general professional information. It is not legal or regulatory advice.