What Marketers Can Learn From Football Analytics

Digital marketing is going through its own analytics-led revolution. The death of the cookie is forcing marketers to rethink how they use data to assess the contribution of marketing to business outcomes. Forward thinking digital advertisers should draw inspiration from Brentford’s approach in terms of how they use data to find unrecognised value.

By: Aidan Mark – Global Director, Performance Strategy at CvE

4 min read

In some ways, both football and marketing are very simple. A simple view might say that marketing is about driving sales, whereas football is about scoring goals to deliver a win.

For anyone not into football, you may not realise that Brentford will be competing with England’s most decorated and wealthy teams in the Premier League next season. This is a huge achievement for a team like Brentford, who were competing in England’s lowest professional division as recently as 2007. But what does any of this have to do with marketing? The parallels become clearer when you understand what Brentford consider their competitive advantage to be. It is not a star striker or decorated manager. Instead, Brentford’s advantage comes from data analytics.

Applying data analytics to a sporting environment is not new. The Oakland Athletics baseball team famously coined the ‘Moneyball’ concept and were so successful that Hollywood cast Brad Pitt in a film that tells their story. Achieving a level of success that goes over and above budget expectations is a consistent theme when sport and data meet, and this is where the parallels with marketing become even more apparent.

Unfortunately, it is the simplistic views of marketing and football that have often been detrimental to success. Reductive thinking has meant the drivers behind why sales are generated and why football teams are able to score goals are not fully recognised or quantified.

How a brand drives sales and how a team scores goals is of huge importance. Context matters because sometimes a player adds very little when scoring a goal. Think of a goal hanging striker scoring an easy tap in. When a player scores that type of goal, it does not really add much to the collective team. In this metaphor, the ball was heading for the net long before the striker got a final touch.

Digital media often acts in a similar way. In the media example, ads takes credit for sales and revenue that were likely to happen. Performance marketers regularly attribute success to the media channels that are scoring metaphorical tap-ins. Think retargeted banner ads claiming via last click attribution that they are 100% responsible for driving a sale from you as the consumer, when you were simply waiting for payday to make your purchase. Perhaps you as the consumer never noticed the banner ad that is getting credited for driving your sale. Similarly, when you look for a discount code from your favourite brand, an affiliate may claim they are 100% responsible for driving the sale. But perhaps you would have paid full price if the discount were not available.

And the biggest culprit of them all, branded search. It is a fallacy that people searching for your brand by name should be assessed with a blanket performance metric that is applied to all forms of performance marketing. But that is what goes on when brands use context free performance metrics like CPA or ROI, deeming everything that achieves the target as being effective marketing, regardless of context. In footballing terms, this is like comparing a goal hanging striker to a star creative midfield player based on goals scored alone. Of course, the goal hanger is probably going to score more goals. But that does not mean the goal hanger is the superior player or is contributing most to the team.

Context matters in digital marketing and its metrics, just as it does in football.

Brentford have been wrestling with this issue and have found a data-led solution that helps them win. One of Brentford’s enabling metrics is expected goals, often referred to in the football world as xG. This metric allows for quantitative assessment of both players and teams, based on the number of goals they might expect to score, rooted in the data-led context of actual on-field play. When football matches result in low scoring games, it can be hard to assess how each player is performing. Using xG helps inform how many chances the striker had to score, and how difficult those chances were to convert. xG therefore shows the specific contribution of each player, by comparing the number of expected goals by actual goals. The best players regularly outperform their xG metric because they score goals that the average player would not. With xG, the contribution of the player to the team becomes more accurately recognised.

Brentford applies this same thought to all aspects of the game. xG analysis is used to identify valuable players in both transfers and team selection, which gives Brentford a competitive advantage over their wealthier rivals. xG and similar creative statistics have been heavily used by Brentford in all aspects of the game. This is where the opportunity for digital and performance marketers becomes clearer. In marketing we need to adopt our own version of the xG metric to recognise the incremental contribution of each media channel, which allows marketers to identify the difference between a media ‘tap in’ versus a wonder goal scored out of nothing. Both types of goal can be scored within a marketing mix, but most brands still cannot recognise the difference. This is a similar market-wide knowledge gap that Brentford are exploiting in football.

Instead of xG, in marketing we might use expected revenue or expected sales which can be achieved with basic propensity modeling. When propensity can be estimated accuracy, the uplift created in real business outcomes can also be observed and attributed back to the appropriate marketing driver.

Indeed, there are many ways for marketers to better quantify the relationship between marketing inputs and outcomes. At Control vs Exposed we run lift experiments, rooted in scientific method, to isolate the exact contribution of any element in the marketing mix. In addition, progressive digital advertisers are increasingly using big-data powered econometric models to quantify the incremental contribution of media channels, including digital. Multi-touch attribution models also help marketers, but only when viewed through the lens of incrementality.

Each measure gives a new perspective to aid data-led decision making, equipping them with a similar toolkit to those used at Brentford. Brands like Adidas, P&G and AirBnB have discovered that the basic metrics available in digital do not represent value back to their businesses. These brands also employ creative use of data to get to that level of understanding.

Football, like analytics, has often relied on the power of simple data and ‘gut feeling’ to establish a winning strategy. Footballers have nuanced contributions to the team that are hard to identify with the naked eye, even from highly experienced experts. This is no different to marketing, where the drivers of sales are part of a complex and joined up ecosystem. Football goals are usually the result of team play involving multiple players, and similarly in marketing, sales and revenue are usually most effectively driven by a combination of media channels and marketing drivers.

Brentford have proven that intelligent use of data can enable a lesser fancied team to compete at the top table. If you as a marketer are competing with rivals that have deep pockets, you would be mad to not seek your own data-enabled advantage.


News & Blog