Continuing our theme of roles within Analytics or Data Science teams, let’s consider the loneliest position.

We are all human beings, no matter how technically gifted. We all make mistakes. Have you had the experience of realising you messed up? Perhaps you risked the reputation of your team, through a mistake in the data or your analysis. How did you handle that?

Based on his experience of watching Milwall play (a braver man than me), guest blogger Tony Boobier returns to discuss what this role means. What does he mean about the loneliest role…

The loneliest position when things go wrong

It’s a moment that will be indelibly be etched in the minds of not only Millwall supporters, but also especially in that of the Millwall goalkeeper who, in the dying moments of the game which victory was within their grasp, turned the ball into his own net.

Being the goalkeeper is the loneliest position on the field (maybe apart from the referee). He (or increasingly ‘she’) is the final line of defence, and if they get it wrong, then the consequences can be significant. It’s a burden which they carry, and sometimes it extends into their non-football professional life as well.

History is littered with goalkeepers who have made their mark: Albert Camus, the renowned author who famously wrote about ‘L’Etranger’ – ‘ The Foreigner’, or ‘Stranger’; Pope John Paul II; Luciano Pavoritti, the great Italian tenor; Julio Inglesias, the crooner; and even Sir Arthur Conan Doyle of Sherlock Holme fame, who spent some time between the sticks at Portsmouth Association Football Club.

So what’s this got to do with the roles needed in a data science team?

Being one Data Science team

The first point is that the whole of the Millwall team had a single vision, which was to win the game and play the next at the National Stadium, Wembley, despite being the underdog.

That vision was not only inspired by the manager Neil Harris, but embraced by the team. You might have thought that Harris was the ‘leader’, but the reality is that the ‘leadership’ – with all that means nowadays – had passed from one individual to the entire team who all knew their place, and what they had to do.

In a data science context, team members may know the technical necessities, but equally importantly don’t they need to have, and own, the vision? For all of them to understand what ‘success’ looks like within a project. Let’s call that idea the ‘democratisation of leadership’, one which has moved away from the flag-waving individual, to one of a group function.

Your Data Science football team

Secondly, shouldn’t we think about the out-field players on the team, all knowing their roles and responsibilities? Knowing how they fit into the storyboard of (hopeful) success.

  • The attackers applying innovative ideas;
  • The midfield dynamos who kept distributing the ball, and providing the link between the forwards and the defenders;
  • The defence itself, providing what is in effect the ‘hygiene’ elements of any project such as timely governance and compliance.

But even with these apparent demarcations, the most effective players were mobile and agile, with defenders attacking when appropriate, and attackers defending when needed.

The loneliest position = taking responsibility

And finally, what’s the analogy to a goalkeeper in the data science team? Well, I guess it’s about taking personal responsibility for our actions, both as an individual and as part of the team. Isn’t it how we feel about our contribution to the success of the team, and how we might feel that we’ve let down the team if we fail personally?

For David Martin, the Millwall Goalie, the best reference came from his manager who said afterwards: “He’s made a mistake and he walked in and put his hands up, and said he had let the boys down. He hasn’t let the boys down, he’s made an error of judgement. He’s a great character and very popular in the dressing-room, so he’ll recover.”

None of us are error-proof, and sometimes we make mistakes, but when that happens it’s critical that we are open and transparent. We don’t try to conceal them.

A Data Science team needs bad news reporting

Many projects fail simply because we are ashamed to give bad news, and issues such as delay or extra cost simply don’t surface until the situation reaches a critical stage. (Take for instance the Crossrail debacle, with a delay of over a year and an additional cost which could be as much as an extra £3.0 Billion). Plus, there are no shortage of IT implementation failures which could be equally referred to.

So there it is. Transferable and empowered leadership; agility and flexibility; technical expertise; and honesty. Don’t these sound like good attributes for a successful team?

How are your team performing?

But here’s the rub. Some might say that, even with these skills and attributes, it still hasn’t prevented the Millwall team from struggling against their competition in the Championship, which is a fair comment.

Maybe we should all be fortunate that there isn’t, as yet, a league table for successful data science teams. If there was, what position in the table do you think your own team might hold?

Answers on a postcard” to Tony, no doubt, or feel free to comment below.