An opening data strategy

What can chess teach us about data strategy? Is there a data strategy or do we need a playbook of data strategies?

I still remember the feeling of disappointment and sheer confusion the first time I saw a data strategy.

I’d been working in a data consulting team for just over a year, earned my first promotion and loved everything about the world of data so far. So, when a volunteer was needed to give feedback on a client’s data strategy, I jumped at the opportunity.

I’m not sure what I was expecting. I’d worked on a lot of tactical data solutions and ‘rapid BI’. So I think I was probably expecting a scaled up version of this. Perhaps proposals for what we might now call data products and how they could deliver value.

What I encountered was a set of policies and technical specifications centred around what I soon understood to be a DAMA wheel.

DAMA (Data Management Association) Wheel

I spent an evening googling new terms, just to make sure I understood the document.

And the more I read it, the more I recognised it for what it was: utter techno-babble. Techno-babble for me is something quite specific. It is the the phenomena of a technical document that makes complete sense on close inspection and yet is devoid of all meaning and purpose when viewed as a whole.

Simon Sinek would have needed sedating. Not only did it not start with a why, there wasn’t a single why anywhere in fifty-plus pages. No business outcomes or success measures. No prioritisation mechanism… It was like a broken pencil, pointless.

Sadly, at that point in time, I lacked the experience and confidence to be this blunt and probably fed back that “it seemed logical and thorough but lack purpose”.

Sadder still, I later discovered that this data strategy was not an anomaly and encountered many more like this.

What should a strategy be?

Let’s leave the worlds of data and business for a minute.

If you search for an image of strategy on Unsplash, 3 of the top 5 images are of chess. Chess also provides a great way to explain what strategy is and is not.

A chess strategy is not a list of allowable moves or a book of moves from historic games. Whilst these are both useful tools for the chess student, they are not strategies.

A chess strategy is a series of high-level principles that aim to increase your chances of winning. There are many strategies. One I learned was to focus on three principles:

1. Gaining control of the centre

2. Piece development - move to positions where your pieces can participate and support each other

3. Protecting your king

A common opening strategy here is the queen's gambit, where you quickly sacrifice a pawn to gain control of the centre, bringing the queen and other vital pieces into play and putting pressure on your opponent.

There are, of course, alternative strategies. Flank openings focus on controlling the sides to gain indirect control of the centre. But you cannot choose to simultaneously pursue a flank opening and a queens gambit.

So what does that teach us about strategy? An effective strategy simplifies things (e.g. gain control of the centre - not the board) and has a clear purpose (activating pieces and improving the odds of winning).

It also teaches what a strategy isn’t. It isn’t perfect or singular. It is OK to discuss multiple viable strategies, but you then need to make a choice; you cannot pursue them all.

Chess vs data strategies

Make no mistake about it. Data strategies are infinitely more complex.

In game theory terms, chess is a finite game with a fixed number of possible moves and a clear endpoint. Data strategies are more like an infinite game where the rules and objectives change over time and there is no definitive endpoint.

You can win a chess game but you cannot win a data strategy.

Picture a game of chess where the board mutates and grows, new players enter and new pieces arrive with new rules. This is more like the level of complexity of data strategy or any business or organisational strategy.

In environments like these strategies themselves are not fixed. Instead the best strategy is to expect your strategy to emerge, adapt and evolve over time.

One of my goals for this year is to develop a playbook of data strategies. A series of different principles and approaches that can be applied to data leadership along with guidance on when best to apply them.

Today though I want to start by introducing one data strategy. It is the data strategy that I have applied most and that I suggest as a starting point for many organisations.

An initial purpose for data strategy

I don’t put much faith in the advice that a data strategy’s purpose should be to support and deliver a business or organisations strategy.

It makes too many assumptions.

1. that there is a clear enough strategy to base the data strategy off;

2. that the organisation’s strategy will last long enough to deliver a data strategy; and

3. that those producing the organisations strategy have considered the role data can play - note if they haven’t then the data definitely strategy needs to influence and lead the organisation’s strategy (think of this as the ability write back to source systems)

A starting purpose that I advocate for a data strategy that aims maximise the benefits that an organisation will get from its data.

This strategic purpose works because creates a feedback loop. By measuring the benefits of data you gain the ability to stop low value work and put most of your effort behind initiatives that deliver material benefits. It is an evolution of sorts - survival of the fittest data products.

An initial set of data strategy principles

Now we have a purpose, we need our equivalent guidance to gaining control of the centre, piece development and protecting the king.

Gain control of the centre

Our purpose (maximising the benefits of data) requires our strategy to touch all areas of the organisation. But we need to start somewhere. Start by gaining control of the centre i.e. your team, wherever that might be.

Lead by example and put in place processes to maximise the benefits that come from your work, by:

1. Establishing a process to agree the value that comes from data work;

2. Only approving new data work where benefits are likely;

3. Measuring the likely value of data work throughout product development to stop work if benefits are no longer likely;

4. Improving your development and product management processes in order to maximise the throughput of your team.

That is it. Together I think of these steps as value focused data product management.

Piece development

Now to maximise the benefits of data across your organisation we also need to work towards building an organisation that is better at delivering benefits from data a.k.a. develop data culture.

This is very much like piece development in chess. It involves making system changes that create a culture where everyone is expected to measure outcomes, where there is support for data careers and skills development and where intricacies like security models support collaborative data work (not silos).

It is easy to get overwhelmed by the size of the to do list and therefore we need to take a product management mindset to identify priority features and prioritise them based on (you guessed) anticipated benefits.

In short you need to start managing data culture as a product.

Protect the king

Finally you need to ensure you do both of the above in a way that maintains and builds stakeholders confidence. Fail to do this and you and your strategy are at risk of check mate.

Honestly, this comes down to nothing more than the basics of programme and change management.

The final principle I advocate is to run a data transformation programme in much the same way as you’d run any other major formal enterprise change programme.

Summary

Strategies are not a list of allowable rules, they are a set of guiding principles that, combined with a clear purpose, simplify decision making.

There are many viable data strategies but the one I advocate most is to focus on maximising the benefits of data through:

1. Data product management

2. Managing data culture as a product

3. Running a data transformation programme

Next

In the next few weeks there will be a deeper, more technical dives, into these three areas.

But I’m also keen for feedback, especially as I look ahead to developing a playbook of data strategies.

Is this helpful?

What strategies have you followed?

What books on strategy/ business strategy/ data strategy would you recommend?

Comment below.