top of page
  • Writer's pictureChelsea Wilkinson

Strategy first; data second

At DataDiligence we have a mantra:

Start with the business problem, not the data.

Sounds obvious, right?

But, if there’s one thing that I’ve learnt in the last year running a data science consultancy, it’s that people nearly always try to jump to the end! They start conversations with: we have this data, what can we do with it? Or: what are the best use cases for my sector?

And while these are interesting musings, they shouldn’t be the opening question.

Rather, we start all our client engagements with a session about the business strategy – examining the successes, challenges, and risks to the business. We actively steer clear of talking about data. Instead, we talk KPIs, 5-year business plans, etc.

As data science experts, it is then our job to translate the overarching business issue(s) – say, the need to increase annual recurring revenue in a SaaS business – into an analytical problem. And only once we have the analytical problem defined do we ask: what data is needed to solve that?

By starting with the business problem, you avoid four common pitfalls (and yes, there are more!), to set your data project up for success by:

  1. MAKING IT RELEVANT – if you are working towards solving one of the top KPIs, your project matters to the business. And, crucially, is aligned with the business strategy.

  2. GARNERING SENIOR CHAMPIONS – to the first point, if you are solving a business-critical KPI, it can be assured that your project is on a C-suite executives’ agenda. And you need this support to drive budget and resource allocations, corporate ‘airtime’, momentum, adoption, ...

  3. MEASURING THE OUTCOMES (NOT OUTPUTS) – it’s easy to be distracted by the lists of exciting (though often short-term) outputs that data projects produce, like a great POC or line of code. But what you really need to be measuring is the direct impact your activities have had on achieving of the KPI - like how much did ARR increase due to your cross-sell, upsell and churn models?

  4. AVOIDING ORPHANED ANALYTICS – many data projects start life as pet hobbies. While this is fun, it doesn’t shift the corporate needle. Projects aligned with the business strategy have longevity, replicability, and relevance.



bottom of page