How analysts can do a better job selling their product

Back to regularly scheduled programming...

Photo by James Wheeler from Pexels

Sometimes in writing about this stuff you get some cool, Zeitgeisty moments, and one of those came this past week when Nikos Overheul tweeted out a link to this story on Hockey-Graphs titled “It’s time to stop talking about analytics”, which I hardly could have put better myself.

The author wisely argues that analytics is a tool like a hammer, but unfortunately, when all you have is a hammer, everything starts to look like a nail. Part of the problem is how all of us—from the media to analytics bloggers themselves—have come to use ‘analytics’ as an umbrella term that captures ideas that are far more foundational, such as ‘basic competency.’

Any team can add an analytics department; what matters is how you integrate that work into your decision-making process. Getting this right, says the author, “is where the real market inefficiency is”:

Absent that decision context, stats, no matter how fancy, are just numbers. It’s fine to have lots of data and statistics, but that data has to be in service of the overarching needs and goals of the organization. And the way to do that is to ensure the data and analytics capabilities are deployed in support of making better decisions. Because ultimately, it is the decisions made every day at all levels within an organization that determine whether it meets its long term goals.

This does not only apply to how you use analytics, of course, but also to a club’s hiring process. For example, rather than hiring a person with a great reputation with the expectation they’ll magically make everything better, it’s probably better to ask how will hiring X person in Y role assist us in making better decisions. That way, you don’t abdicate responsibility for outcomes because you hired a big name and expected big results.

Again, I think analytics writers and companies advertising their wares need to explore these questions a little more deeply. To take a good example, the uber-talented James Yorke published a fascinating piece a few weeks ago on Statsbomb’s ability to identify ‘pass footedness’ across teams or players.

It’s an interesting snapshot on how players behave in the Premier League, but its real utility is implied almost up until the end of the piece, where we find this tantalizing hint at how this information might be effectively used by an interested club:

Each of Arsenal’s three main centre backs are almost entirely right footed, but have different passing profiles. Koscielny looks to switch between his fellow defenders, from side to side, while Sokratis will pass forwards with range, left or right. Should an opponent run a press against Arsenal, and consider blocking passing lanes and preventing easy outs, there is information here to work with.

This paragraph could almost be spun out into a standalone piece, and it probably should! This piece might explore how could a team practically work with this information, exactly? What are the possible outcomes? How might having this data tip the scale over the course of a few games, or even a relatively tight contest?

Because while the utility of understanding pass footedness in this context might be obvious to a manager or first team coach, it’s not at all going to be obvious to the layperson, who, unfortunately for the analyst, is more often than not the person who will make the decision about hiring an analytics company in the first place.

Having this information will also make it easier on whoever is considering paying for this information to explain its potential to those in the club overseeing the football side of things—including their own in-house analyst (should they have one).

It will also prompt the club to ask better questions when hiring the analytics company before purchasing their product—how up-to-date is this information? Is there variability in pass footedness from match-to-match or tactical context to tactical context, etc.?

I think making the potential gains a particular analytics tool might offer a club as explicit and clear as possible without giving it away for free will help convince doubters, because it’s clear that clubs are desperate for a comparatively affordable edge.

This past few weeks saw not one but two pieces focusing on the incredible income disparity not only in the Premier League but the Championship as well.

Of the former, on NYT Rory Smith points to a new status quo, one where the required points tally for a title win is increasingly absurd, and where historically big clubs can slump their way to a reasonably comfortable finish:

It is just how the Big Six executive wanted it as he sat in that hotel conference room, the Premier League’s greatest miracle shimmering on the horizon. An elite so strong that they break records as a matter of course, that they win almost every game, that they can never be caught. Three years later, his vision has been realized. Perhaps, in hindsight, this season will not seem quite so exceptional. Perhaps this is just the new normal, the Premier League as the Big Six want it to be, whether anyone else likes it or not.

And the Championship, whose inequality is fueled by grossly inflated parachute payments, is faring little better in encouraging more open competition. As Rotherham United manager Paul Warne said, quoted here on traininground.guru:

“If the smaller teams come up, the task is getting greater and greater and it’s turning into a Premier League 2 and the rest in my opinion. I think it was the Brentford manager after we played them who said there are 19 teams in the division who are ex Premier League.

"The parachute payments are crucial, they make a massive difference, but the money from being in the Premier League allows you to improve your facilities and infrastructure. I think now the divide is getting bigger and bigger.”

This all sounds hopelessly dire, and it is certainly not ideal for those who prefer something approximating fair competition. But it is also an environment where less well-off clubs might be theoretically sympathetic to someone approaching them with the promise of a proprietary edge.

But this only works if the highly specific nature of that edge is spelled out in the clearest and simplest terms. Analytics might be a tool, and a tool might only be as good as the person using it, it helps to know exactly what the tool can really do.