3 ways NHL teams can get more data driven

The NHL isn’t the only business experiencing seismic ripples and paradigmatic shifts in the information age. The time of “Big Data” is upon us, driven by mass adoption of powerful communication technologies. There’s deep wells of information everywhere for those who care to look, waiting to be plumbed for valuable insights and new efficiencies. 

Of course, access to information is necessary but not sufficient when it comes to operationalizing data in an organization or business. NHL teams have begun to dip their toe in the “advanced stats” pool, but are no doubt encountering challenges in effectively leveraging the various analyst bloggers who have been scooped up in the past year or two. It’s one thing to bolt on a “stats consultant” to existing organizational structures and quite another to integrate empirical processes from top to bottom. 

I recently read the book Winning With Data by Tomasz Tunguz and Frank Bien. In it, Tunguz and Bien lay out a roadmap for successfully operationalizing data, as well as the many roadblocks businesses often encounter in trying to become more empirical. 

Here are three key insights I discovered in Winning With Data that could help NHL clubs become truly data driven.

Perhaps the biggest obstacle is teaching people (especially Very Important People) to treat their ideas as theories to be tested and not axiomatic. 

When I say “ideas” here, I mean the conventional wisdom and legacy conceptions about how to win in the NHL. Most executives have been in the game for a long time and are a part of the NHL’s insular monoculture. This no doubt grants them a level of knowledge and insight that is not available to outsiders and consultants. 

But it also makes them susceptible to insidious processes that can skew or short circuit effective decision making: the suppression of “outsider” or non-conforming ideas, the seductive uniformity of groupthink, the coronation of seniority (i.e.; experience) over innovation.  

Decision makers must be willing to accept that their current view of the game is not necessarily correct in order to find new insights. Operationalizing data means constantly asking new questions, thoroughly testing them, and then re-orienting assumptions based on outcomes (and then starting all over again). 

The best, most useful insights are often counter-intuitive. Exploiting a market inefficiency often means finding and acting on something everyone else is ignoring (or thinks is completely wrong). This can’t be accomplished if those at the top of the pile treat their assumptions as immutable principles and not hypotheses subject to inquiry and proof. 

Which brings us to…

HiPPO = highest paid person’s opinion.

Deference to the HiPPO in the room can lead to decision making that is biased, inaccurate or in direct opposition to evidence, because a HiPPO is free to ignore or undermine data-based insights on a whim. 

This is a significant hurdle in NHL front offices that have borrowed their structure from classic military organizations: rigid and top-down hierarchies where the information and authority only flows from top to bottom. This is why, in response to the kerfuffle surrounding Matt Pfeffer leaving Montreal after the P.K. Subban/Shea Weber trade, he was told to shut and up and “know his role” by former NHL players like Corey Hirsch and Patrick O’Sullivan on Twitter. The sense that guys down the totem pole are to be seen and not heard pervades dressing rooms and upper offices alike.

Operationalizing data requires a flattening of org structure so that info flow freely across departments and upwards towards management. Brutal intellectual honesty means, simply, humility before the facts. Companies that are effective at defusing groupthink and killing the HiPPO often employ a Devil’s Advocate in meetings for projects – a person who consistently tests assumptions and claims for their accuracy and never stops asking “what if we’re wrong?”

Proper empiricism in an org should often render the HiPPO’s personal opinion moot. By asking questions, carrying out tests, and surfacing data-backed insights, it brings a clarity that denudes bald opinion and gut feel of their apparent legitimacy. 

This is likely where most teams are starting: with their own databases and data collection teams. 

There’s more to it than just gathering information, however. Clubs must find ways to make data distribution and communication friction free and clear so that it can be employed and understood by everyone throughout organization. They have to create a “data dictionary” and “data fabric”, where everyone involved understands the terms, objectives and value of the information. If everyone knows where the data is, what it means and how to communicate its implications, leveraging it for decision making becomes far easier.

The ideal end state for NHL teams would be for everyone in the organization to be able to ask their own questions or conduct their own experiments with the available data without having to constantly make requests of an analyst. Making data universally accessible, agreed upon and clearly communicated can break down barriers, reduce arguments and increase understanding at all levels of a company or business. 

Test your Assumptions, Kill the HiPPO and Clear and Accessible Data: if NHL teams can integrate these three primary insights, they will take very strong strides towards becoming more empirical and data driven.

I have only skimmed the surface of what Winning With Data offers to NHL clubs, as well business owners and companies in general. I highly recommend it to anyone interested in how to create and operate a more data oriented culture.