The world’s leading publication for data science, AI, and ML professionals.

When data speaks, and no one listens

Intuition vs information.

We’ve all been there; you take the spec for the information required from management, you go back to your (data/BI) team, discuss it, they do their magic and create a report that holds key information of the direction that the company should take.

You then present the report back to management, and they ignore it.

It so happens that ‘someone’ has a ‘hunch’ different to what the report says and manages to sway the decision resulting in the company taking a different direction.

As a Data Lead, you have to go back to your team and somehow explain why that report you asked them for, the one they probably had to drop everything to finish, has been put aside and why management has opted to go with Joe Bloggs’s hunch rather than factual data.

Motivation in data teams is a hard one to maintain.

To work in data today, one needs a specific mindset – a resilience to withstand those who disregard authentic information and instead go with their gut feelings.

Data Leads are often stuck in between a rock and a hard place between their team and the rest of the company. They need to convince their teams that their work is valued, and at the same time, minimize the damage from those who insist on feeding their own egos and overturning data-driven decisions with random intuition decisions.

The solution is not simple but can start with three steps.

Forever striving for more data-centric decisions, I’m always looking for ways to streamline the understanding of data so that everyone understands it and can relate what they’re doing on a day-to-day basis to whether or not it’s impacting the growth of the company.

Is the confusion and discount of data due to the fact we have such a vast array of metrics that we find it hard to make sense of it all?

As a result of so many metrics, we often fail to pin down the one main metric, also known as the north star. If we do manage – there seems to be a common problem in that we lack in communicating it effectively as well as relating it to everyone’s activities, efforts and intentions.

There are many pieces to this puzzle, and it’s not an easy one to solve. I have, however, tried and tested and identified that there are (amongst others) initially three steps that can help to start to embed data into decision making.

  1. There should be ONE main metric (North Star Metric) that everyone knows about and whether they agree upon it or not whoever has the right authority should be able to declare it and ward off any pursuits to change it- not saying it should never be changed or revised. Still, it shouldn’t be a regular practise.
  2. The main metric should be broken down into a minimal amount of KPIs (I suggest around 6) that relate directly to customer/prospect/user journey.
  3. In addition to the above (point 2) break down the 6 KPIs to the team level and again down to the individual level, this results in a correlation of what they are doing. That way, both teams and individuals within the teams work with a specific intention.

Put it all together, measure and optimise.

The solution could be the consolidation of what the team does (as in activities, i.e. feature development, sales and marketing campaigns etc.), with what the customer is doing (user/customer journey and behaviour). With that in place, you can easily measure if what is being done is, in fact, making an impact on the KPIs, which in turn is tied to the main metric (North Star).

Conclusion.

Once this fusion takes place, you can then measure what is working and what isn’t, which enables you to tweak and try again and once you’ve found what it is that is impacting growth, you can optimize and build on that.

This then becomes part of a framework that speaks a language that the whole team can participate in, as wells as understand to achieve the same goal.

The clear outcomes should be able to convince even the harshest sceptics and so-called hunch believers, letting the data speak for itself.


Related Articles