Building on our month focussed on controversial topics, let’s turn to what will set your team up for success.

Different contexts can require different types of analytics team. A lot of the advice that I offer within the Opinion section of this blog is based on a lifetime leading teams in large corporates. So, I’m pleased to partner with guest bloggers from other settings.

On that basis, I am delighted to welcome a new guest blogger. Alan Murray. After a career in the worlds of consultancy & large corporate, Alan has spent plenty of years in both medium-sized firms & lately startups.

So, over to Alan to explain why getting “fuzzy” is the way for an analytics team to see success in the world of startups…

Get fuzzy! Why it is needed

My co-founders and I have recently had to face up to this challenge of creating a new data analytics team having set up our new firm Vistalworks, earlier in 2019.  Thinking about this challenge, reflecting on what we know, and getting the right answer (for us) has been an enlightening process.

With 70-odd years of experience between us, we have plenty of examples of what not to do in data analytics teams, but the really valuable question has been what should we do, and what conditions we should set up in order to give our new team the best chance to be successful.

As we talked through this issue my main personal observation was that successful data analytics teams, of whatever size, have very fuzzy boundaries between the different roles (or functions).  In fact, the fuzziness of the boundaries feels much more important than the precise specification of the roles.  

That is a relief in many ways, because roles are really not standard (my friend Dr Natacha Lord recently wrote a great LinkedIn post on that very issue), and, especially in a small company, roles are going to evolve pretty quickly. Reflecting on that made me feel that the notion of fuzziness was even more important than I had initially thought.

What do we mean by fuzziness?

Which is great, but fuzziness is almost by definition a bit vague, so to make this observation useful, it was important to be a bit more specific.  Therefore I started to ask myself: What is fuzzy?  How can one be fuzzy?  How does a group get fuzzy???

That lead me to thinking about the behaviours that lead to the desired state of fuzzy boundaries, in order to get down a level to something that we could actually ask for in a team. The next realisation was that that successful (and fuzzy) data teams need to have people who can move out of their own role (or function) and meaningfully discuss the challenges and choices that their teammates in other roles face.  And that has to apply to everyone, not just a central product manager-type.  Fuzziness is a two-way street.

Rephrasing this again, we wanted an environment where there is no need to spend time breaking down walls, because the team are happy jumping over the walls.  Breaking down walls proves to be time consuming, draining, and actually quite a negative thing.  Personally, I’d always rather focus on something positive: making ladders, or trampolines, or whatever apparatus is needed to get from one side of the wall to another.

Being fuzzy in practice

Ultimately, this has manifested itself in 3 ways:

  • Having a really clear set of values, which are put front and centre in all recruitment
  • Commitment from the co-founders to provide coaching, especially on softer skills 
  • Willingness to learn and redesign sprint over sprint, improving the approach each time

As for the ‘why’ of all this, to me it strikes at the heart of another truth that has come up a number of times over the years: it is easier to build new things than it is to integrate two things that have already been built.

Therefore, by setting up the team that will integrate things well, you are taking on the hard part, and that is probably why these teams have been more successful than others over the years.  Integration inherently comes later in the process, when contingency has been used up by other people, and deadlines loom closer.  Having people who can work together effectively at that stage is a real boon.  When the going gets tough, the fuzzy get going…

Fuzziness for your Analytics team, is it for you?

Thanks to Alan for sharing his experience and the analytics team that Vistalworks needed.

What about you? Are you convinced by the need for fuzziness or does your team require more well defined roles?

Please share your experience in the comments boxes below or on social media, so we can continue to develop best practice in analytics team leadership. Let’s see if it is relative to context.