This is a slightly edited/expanded version of a post that was published recently on the Credera website.

If you’re a CDO (in either name or responsibility), chances are you’ve had to write a data strategy – heck, perhaps you’re even writing one right now. If you haven’t, you may feel that everything would go much more smoothly if you were able to pull it out of your bag and wave it in the face of every naysaying executive stakeholder who dares to question your work, with a righteous cry of, “it’s in the data strategy!” Sadly, naysayers are not so easily swayed. But there are some things you can do to at least raise the chances of your data strategy causing the C-suite to fall gratefully in line.

Why bother with a data strategy?

The CDO role is notoriously hard to pin down, and its success criteria even more so, as I’ve previously discussed. Some sort of document that puts flesh on the bones of an organisation’s aspiration to “be more data-driven” therefore seems like a good idea — it can help define the areas where the organisation needs to focus and provide a framework for determining whether progress is being made.

It’s not, of course, a magic bullet. A data strategy, however well-written and blessed by executive leadership, only represents the start of the CDO’s job of cajoling stakeholders and execution partners to pull together. But it can establish a set of principles that the organisation can use to guide data investments, and help to energise disparate teams who must work together and compromise in the name of progress.

Keys to a successful data strategy

Not all Data Strategies are created equal. Some drive real change and transformation; others gather dust on a shelf, or – worse – cause confusion and conflict. But what can you do to ensure yours falls into the first category, and not the second? Here are ten tips.

1. Connect it to your organisation’s goals

I’ve lost count of the number of times I’ve talked to CDOs who, when asked what their Data strategy is, have said something like, “We want to build a high-quality enterprise data platform to allow us to transform into a data-driven organisation”. When pressed on exactly what this means, they’ll mumble something about data quality, or data cataloguing, or perhaps a single customer view.

I call this the “Field of Dreams” approach to data strategy — If you build it, they will come. It’s a strategy based on the hope that if good quality, easily accessible data is available, people will magically know what to do with it and how to leverage it to achieve their goals.

Instead, the explicit goal of a Data strategy should be to enable the organisation’s overall strategic goals. As one of the attendees at our recent CDO community dinner said, “Business Strategy is Data strategy” (might have been the other way round; I had had some wine by that point in the evening).

Of course, you might object that your organisation’s overall strategic goals are too high-level to just copy/paste into the data strategy. You can’t just say “Our data strategy is to grow revenue by 20% and increase customer satisfaction and retention”; you have to be able to sketch out how it is going to enable these things to happen. But that’s where your friends in the business come in.

2. Establish some principles

The life of the CDO is one of careful compromise, cat-herding, and looking over your shoulder. In this environment it’s easy to conclude that expediency is the safest route through the corporate quagmire; and indeed, it’s important to pick your battles – no CDO will last long if they go to war with the entire organisation.

However, in order for it to have durability and integrity, your data strategy needs to lay down some principles that will guide all the other decisions and plans that you make. What these principles are is up to you; they can cover all the aspects of your strategy from technology to operating model to skills and empowerment. The important thing is that even as the details of your strategy change and evolve (as they are bound to), your core principles are much more persistent. This means that it’s important to have strong agreement on the principles with your stakeholders and set the expectation that you will all abide by these principles, and not abandon them when they become inconvenient.

3. Build it together with your stakeholders

If the purpose of a data strategy is to explain how data will be used to achieve the organisation’s objectives, then it is essential to develop that argument in concert with the business stakeholders who can actually achieve these objectives.

The biggest challenge to doing this is the chicken-and-egg problem of identifying how data can be used specifically to support business objectives: stakeholders can’t identify opportunities until they know what data is available, and data teams can’t determine what data is needed unless they know the objectives.

This is where the CDO can add value, by coaching their key stakeholders in how to leverage data. One very simple way to do this is to use the framework proposed by Florian Zettelmeyer of the Kellogg School of Management:

• How can data be used to decide what to do?
• How can data be used to enable what needs to be done?
• How can data be used to measure the success of what has been done?

For example, if a business stakeholder says, “My goal is to increase customer retention by 10%”, then the CDO’s team might suggest one or more of the following:

• Decide: Do product analytics to identify product frustrations that are causing churn
• Enable: Use customer data to drive targeted retention messaging
• Measure: Define a common measure of retention and include it in executive dashboards

Importantly, these are end-point use cases or applications of data; they are not merely data assets or capabilities (however important those are to enable the use cases).

The other reason to build the strategy with the stakeholders is that it raises the likelihood that they’ll be bought into the strategy from day one. A strategy that can be linked to strategic organisation goals, and which was built collaboratively with the business, has a better chance of surviving to implementation.

Finally, the other set of stakeholders who must have a hand in crafting the data strategy are those who will be needed to help implement it – particularly IT/engineering teams. They too will be motivated to prioritise work that has a clear link to top-line business value, and will definitely appreciate being brought along for the ride.

4. Don’t make it all about technology

An unfortunate side-effect of the fact that many CDOs are drawn from the technology function in their organisation is that data strategies are too often much too focused on questions of technology and architecture. Technology has an alluringly concrete feel to it (in contrast to the more nebulous topics of data use cases or operational models), and is often where the money is, which exacerbates the problem.

Apart from compounding the issues discussed above of failing to engage with stakeholders and connecting to business strategy, making a data strategy all about technology ensures that it will be a contentious and divisive document – after all, what topic do business leaders like to argue about more than technology? Unless the CDO has complete control over the entire data technology stack (which is effectively impossible in any but the smallest start-up), this will put them in conflict with the technology team, and slow progress.

5. Make it about people

The yin to the yang of “Don’t make it all about technology” is “Make it about people”. Your strategy should contain real goals and plans about how you plan to enable people throughout your organisation to feel more comfortable and empowered to use data to support their work. This is about a lot more than just putting data in their hands and telling them about it. Rather, it’s about championing cultural change within the organisation so that people feel more confident about forming hypotheses, trying things, and then seeing whether those things worked out, all supported by data, confident in the knowledge that they’re not going to lose their jobs if they make a mistake.

A good data strategy will include concrete plans and goals on how to improve data confidence and data-supporting thinking. It’s also not, by the way, just about educating the “grunts”. At Microsoft we spent a lot of time educating our senior leadership on how they could integrate data into their strategic thinking and ideation, which resulted both in the better use of data and better outcomes for key projects.

6. Don’t make it all about saving money

Saving money is incredibly important. It’s also incredibly boring. Focusing your data strategy only on saving money will ensure it reads like a cross between a Calvinist tract and a credit card statement.

Even if your organisation is not focused on making money (for example a public sector or non-profit), think hard about how your data strategy supports the creation of value and not just the avoidance or minimisation of cost. Cost-cutting is by definition a game of diminishing returns; every bit of progress you make makes it harder to find the next efficiency to squeeze out. Endlessly cutting costs is a soul-crushing way to think about the value of data, and is almost guaranteed to make it hard for you to find and retain good people in your organisation.

7. Make sure it reflects where the organisation is now

When many people hear the word “strategy”, they tend to think of top-of-the-mountain, aspirational goals to making fundamental changes to the way the organisation functions through data. It’s fine to have these aspirational goals, but if people cannot relate them to where the organisation is today, they will not be taken seriously. I know, I know, shoot for the stars and you’ll reach the moon, etc.; but on the other hand, a journey of a thousand miles starts with a single step.

If, today, your organisation is struggling to manage its core data effectively, and there are limited data skills across your business teams, then a data strategy that waxes lyrical about Data Science and AI is unlikely to resonate or be particularly useful; far better to focus on landing a solid enterprise data platform and get people’s data skills to a good basic level.

It’s actually fine for your data strategy to address the issues that are in front of you now; people in your organisation who encourage you to focus on loftier goals may be trying to ensure that the current state of affairs (which may benefit them) does not change quickly.

8. Treat it as a living document

I’ve had conversations with CDOs who say that they want their data strategy to last for the next five years. This is… optimistic, to say the least.

You should expect your data strategy to guide the next 12 – 18 months of work. That doesn’t by any means mean that your data strategy should only focus on the next 12 months; just that you should expect to update it approximately this often, to respond to changes in your business and technology circumstances. Not least among these is the fact that most CDOs only last around two years – creating a five-year plan seems fanciful in this context.

Your data strategy shouldn’t be constantly changing, because then people will just expect to be able to ignore it in the expectation they can bend it to their will, but you should not expect to get it right first time. Let your stakeholders know that the strategy will be regularly reviewed and adjusted based on what you’ve learned along the way. but control this process yourself. Signalling to your stakeholders and leadership that the strategy will evolve over time shows the right kind of willingness to adapt and adjust that will mark you out as a pragmatic, outcomes-oriented CDO.

9. Make it actionable

As we already discussed, a good data strategy will lay out a compelling top-of-the-mountain vision for how data can support the organisation’s overall goals and aspirations, supported by a set of principles that can guide investments and provide a way of measuring progress. But it’s also important that it lays out a concrete set of actions that can help you start to climb the mountain, and which are aligned with the principles. Not every step of the journey has to be mapped out in order to identify these first steps; given the inevitable evolution and updates to the strategy, it would be impossible to do so.

Naysayers will point to the lack of a perfect roadmap, or incredibly detailed end-state description, as evidence that the data strategy is half-baked. These people should be ignored, or asked to provide evidence of the same level of rigour in their own projects.

10. Sell it

My final piece of advice doesn’t concern the data strategy itself, but what you do with it once it’s written. Here’s the ugly truth: Most people in your organisation don’t care that you put together a data strategy (and the ones that do care probably wish you hadn’t, because your strategy says they should stop running some tactical data solution that they love). So you gotta sell the thing – to leadership, of course, but also (again) to the stakeholders who helped you put it together, and crucially to the people that you expect to benefit from it.

If your reaction to this piece of advice is, “Yes, but end users won’t really understand what we’re trying to do; telling them about the data strategy will merely confuse them”, then I’m afraid that’s a problem with the data strategy, not the end users. Sure, not everyone in your Finance team needs to understand the finer points of data virtualisation, but they probably will get excited if you tell them that you’re aiming to make it much easier to bring together all the data they need to do regulatory reporting each month. Plus, if you get these folks excited about what’s to come, they’ll champion the strategy, and keep you honest about getting it done.

In a nutshell

There’s no such thing as the perfect data strategy, and what yours specifically needs to focus on will depend on the particular challenges you’re facing. But if you keep the above points in mind, you’ll maximise the chances that the strategy will help you achieve your goals and help your organisation make better use of data.