Learning from Sports Analytics

With all due respect to my dear American clients and friends, let’s be honest, watching baseball is a very boring way to pass time. I could be showing my Western Canadian bias here, but compared to the sports that I follow, hockey and football, the pace of play in baseball is just too slow to keep my attention. But the one aspect of baseball I find really interesting is the advanced statistics. Popularized by the book and subsequent movie ‘Moneyball’, the arrival of advanced baseball analytics changed the way the whole industry evaluated their assets and processes – and not just in the sport of baseball either.

In hockey, up until fairly recently, the only statistics people cared about were goals, assists, and plus/minus. These are measures of directly observable offensive actions, but even for forwards, they aren’t the greatest metrics in terms of gauging a player’s overall contribution. The worth of defensemen is even less-so encompassed by these old school metrics. Yet hockey teams have long understood that a roster of quality defensemen has immeasurable benefits to the rest of the team, and good defensemen are always in high demand. Marketers have faced a similar dilemma with tactics like SEO, content creation, and branding; while we understand these are foundational elements for marketing success, we lack metrics to truly appreciate how impactful they are. Instead we tend to place an over-emphasis on the things that are directly measurable.

This over-reliance on direct measurement has changed in recent years. Hockey fans and pundits have become well-versed in advanced statistics that aim to measure the players’ more subtle contributions to the game. The most well-known is ‘Corsi’, which measures a team’s shot differential when a player is on the ice, and is meant to be seen as a measure of a player’s contribution to controlling the play. It puts offensive and defensive players on even ground. A team’s overall Corsi rating also happens to be the most highly correlated statistics to overall team success.

In football, the concept of controlling the play is also seen when the team who’s ahead elects to run the ball instead of the comparatively risky strategy of throwing. Taking a run-first strategy results in fewer points scored, but also fewer turnovers, and it helps bring the game to a prompter conclusion. If you were to do a statistical analysis, there’d be a correlation between teams winning and teams running the ball more frequently. But this isn’t a prescriptive insight from a strategy perspective. Teams run the ball more when they are already winning late in the game to run out the clock and reduce the risk of a counter-attack. But they usually throw the ball a lot to actually get into a leading position. Running the ball frequently from the beginning of the game is a poor strategy. This wasn’t so well understood in the past. It’s no co-incidence that modern NFL quarterbacks are achieving record setting passing totals, while running backs and full backs are seeing a diminishing role.

There is much debate in the hockey community regarding how much stock should be put into advanced stats like Corsi. On one end of the debate, you have last year’s Stanley Cup Champions, The Washington Capitals. They’ve made a point of de-prioritizing Corsi, electing to optimize for the quality of their shots rather than the quantity. On the other hand, you have the Carolina Hurricanes, a lower budget team akin to a Moneyball situation. Carolina’s radical new owner, Thomas Dundon, has been on the record about how the team won’t be prioritizing the selection of defensemen in the draft. They have decided to go all-in with advanced stats as a driver of strategy, and have taken the quantity over quality of shots approach. Indeed, over the last two regular seasons, Carolina has had the highest Corsi rating in the league, despite being just a middling team in the standings. While Washington is in the lower half of the league in Corsi, but actually won the championship. Co-incidentally, at the time of writing the Capitals and Hurricanes are dueling in the 1^st round of the NHL Playoffs, so we’ll see which strategy prevails. My money would be on Washington, as it seems Carolina is making the same mistake NFL teams did for so long in taking statistics out of context, believing correlation equals causation, and incorporating those factors into their underlying strategy.

The NHL’s youngest ever General Manager is the Arizona Coyotes’ John Chayka, who first got the job at 26 years old. Chayka is seen by many as a leader in hockey analytics. On the 31 Thoughts Podcast, Chayka was asked about how his team utilizes analytics when drafting players, and he highlighted the importance of taking data in context:

“Everyone just wants a magic number or an end result. That’s a very difficult thing to do, to judge players in different leagues, so context is key. When we’re using quantitative analysis in our draft process, there are certain things we can do to understand a player playing in different leagues… We have a lot of techniques and different ways that guys that are very advanced in math approach it with. I’m not a big believer that any of them are the true answer and give us an exact response. But I do believe it adds a lot to our process. And then when we get input from our scouts and we go through that entire draft day process, it’s very helpful and very meaningful. But we’re not taking a player just based off his data.”

In marketing, it’s my belief that our ability to differentiate correlation and causation is going to be further hampered as time passes. There’s an on-going battle between MarTech companies vs. privacy seeking consumers; with near mainstream adoption of privacy tools and with legislation like GDPR coming into effect, from my perspective it seems the consumers are winning. We marketers should expect less available data, especially about what’s traditionally been directly measurable like the ability to attribute sales to particular marketing tactics. If your marketing strategy for 2019 and beyond is optimizing only for what is directly measurable, you’re not going to be placing enough emphasis on more important tactics like SEO, content creation, and branding, whose contributions are greater but more subtle.

Of baseball, football, hockey, and marketing, baseball is the easiest to measure by far. America’s greatest past time is a turn based game with an awfully generous amount of time in between events to allow for the recording of data. Hockey, football, and the modern marketing landscape are much more dynamic and harder to measure and to understand in context. You don’t want to be running the football excessively, shooting the hockey puck at will, or spamming end-of-funnel marketing tactics over the fundamentals, only because that’s what you’re best able to measure. Sticking to what makes sense over what’s measurable and taking a quality over quantity approach is the best strategy.

Want to talk about your marketing metrics and the implications on your business’ strategy? Contact us today.