What does a good data strategy look like?

Chatham House rule session, so no direct reference to organisations or people included.

A data strategy should be in support of a goal.

What makes a good strategy useful?

What are you going to do?
How are you going to do it?
How does it fit into the context of what you are doing?

Sometimes you need to look beyond the immediate goals for open data. But it’s possible to look too high – pay attention to quick wins. What are the most effective things you can do? Those quick wins? Or doing the foundations for the next step?

Ultimately data doesn’t matter – the results do. Data is a step on that way.

A strategy is a live document – it needs to be adaptable to situations. It should change. It will change. That’s fine — it captures a moment, and guides towards the next moment. Recognising success while you iterate your strategy is really important.

Stakeholder engagement

Engaging people in the strategy process really helps with getting people to adopt it. They feel some ownership of it. Conversation works really well for some people, but different organisations won’t take things seriously until they’re on a slide or a Word document. Internal chat tools can be a good way of getting involvement early in the project – and there may not be a single way to engage everybody in the organisation.

Every data strategy should have principles, but they need to resonate with the organisation. You can’t tell people what their principles are.

When thinking about the strategy, think about how it impacts people’s work. You can help people see themselves — and their work — in the strategy.

Relate it to organisation vision

Any strategy must support the organisation’s vision or fundamental goal – and that includes a data strategy. For example, it might support single, accessible versions of the latest dataset for a particular dataset; authoritative, sharable data. Is you data fit to be seen?

Open Data supports this, but it has to be foregrounded, or it tends to get lost or forgotten about. Can you use standard data formats? Or clearly map between formats? Can you make helping other organisations achieve their goals part of your strategy (that might make a lot of sense in government).

Can you make sure you focus your resources where they are truly needed, and ovoid replicating work that might be available elsewhere? A cross-governmental data stagey will probably never work, because different organisations are so diverse, but you could feasibly have a set of uniting principles, that guide towards a unifying infrastructure.


It should not be a project plan. Action plans matter and they can be based on the strategy – but they are separate things. They clarify strategy by making it clear by how we deliver something.

Equally, you need to think about things like KPIs and other building blocks that will allow implementation to happen. And KPIs are data in their own right.

[Session Notes]