What SEO Marketers Can Learn From the Analytical Revolution in Basketball

More and more sports teams count on in-depth analytics to improve their odds on and off the field. How do they do it and what can SEO marketers learn from this approach?

The analytical revolution in sports started with the 2002 Oakland baseball team. Despite a limited budget, this team managed to win the championship against all odds, and did set a record for most consecutive wins.

So how did this team maximise their wins when given a limited budget? They used mathematics and advanced metrics to estimate player values. They noticed a big impact of on-base percentage on the amount of wins. The story of the Oakland baseball team can be seen in the 2011 movie “Moneyball”, starring Brad Pitt and Jonah Hill.

A dedicated data division

Nowadays, every professional American team in baseball, football and basketball has its own data division. There are cameras everywhere on the court registering every detail, PHD statisticians are influencing how the game is played and which players are selected to form a team. Every year, the MIT University holds its own conference “MIT Sloan Sports Analytics” attracting sports analysts from all over America.

The concept of using advanced metrics can be translated from the game of sports to the game of SEO. As I’ve never seen a game of baseball in my life, and thus being unable to interpret these metrics, I use the game of basketball, a sport which I’m more familiar with. Now, I will explain to you how you can strengthen your SEO strategy with advanced metrics.

The rise of advanced metrics in basketball

The first book “basketball on paper” was written by Dean Oliver in 2003. Dean Oliver, together with John Hollinger, was discussing basketball analytics way back on a Yahoo group and were then hired by professional NBA teams which, in turn, lead to an analytical revolution in the game of basketball.

A few years later, PHD mathematician Stephen Sea and software developer Christopher Baker wrote the book ‘Basketball Analytics’, which contains some more elaborate advanced metrics to measure the value of players.

We can use the thinking behind these advanced metrics used in basketball to analyse SEO results.

Rebound rate and SEO acquisition rate

Is the player with 10 rebounds in a game always better than the player grabbing 9 rebounds? The answer is no. It depends on how fast the game is played, how many shots were taken during the time the player was on court and how many of them were missed. You can’t get rebounds if no one misses a shot.

The rebound rate is calculated as the percentage of rebound.

Rebound rate = rebound taken by a player / amount of missed shots during time on court

The same way as you can’t rebound if there are no shots, you can’t get any SEO traffic if there are no searches for the relevant keywords you optimise for. Therefore, to have a clearer view of how your SEO is working, you have to take into account the seasonality and customer trends to more clearly assess the effect.

So an accurate advanced metric would be SEO acquisition rate.

SEO acquisition rate = SEO clicks on relevant keywords / keyword traffic on relevant keywords

Adjusted +/- and SEO component analysis

The +/- score in basketball is the amount of points your team scores, e.g. L. Hunter played 11 minutes 23 seconds, during the time he played the team scored 22 points and allowed only 21 points, so he has a positive +/-.

The adjusted +/- is a more advanced metric that takes into account the level of the players you are playing with. While this metric is not a holy grail, it can indicate problems if a player has a big negative +/- and detect those players who are good team players without the stats.

Also in SEO you can check which components do drive the traffic. Which pages are important for SEO, which of the pages do attract traffic and leads and which pages don’t get any traffic and do not contribute to the end result?

(source image: winninghoops.com)

Three-pointers and SEO conversion value

Popular offensive metrics are effective field goal percentage (eFG%) and points per possession.

eFG% is calculated as FGM+0,5*3PM / FGA, with:

  • FGM: field goals made
  • 3PM: three pointers made
  • FGA: field goals attempted.

Statistically speaking, teams shooting a lot of three pointers had a higher eFG% and thus changed their game plan to benefit more from the added value of the three pointer. This fundamentally changed the game of basketball in the previous years.

As you can see in the graph below, the increase in three point shots taken and scored was huge and this had a positive effect on the eFG%.

Also in SEO you don’t have to look at just traffic, some of the traffic is more valuable than other traffic. You should weigh your metrics for conversion value rather than number of conversions to be sure you score three pointers.

Entropy in basketball and SEO

A team consists of more than 5 players. You can select the best 5 players by value, but not have the best team. The key to winning games is having a balanced team with no real weaknesses. This can be summarised in an “entropy” metric. The more diverse the skills of the players on the court are, the better your team will play.

In SEO as well, you should consider entropy and differentiate your SEO efforts to avoid weaknesses within your strategy.

  • Do you have a lot of distribution in SEO traffic or does one page get 80% of the traffic?
  • Can you make a competitor analysis of your keywords to see which areas you are lacking on — compared against your competitors?

What are advanced metrics?

Basic metrics in basketball are: points, rebounds and assists. Basic metrics in SEO are visits and keyword rankings.

While these basic metrics can tell you something and are very easy to communicate and understand, they won’t tell you the whole story. Advanced metrics do a better job showing you nuances within the metrics. You can score 30 points in a basketball game, but if you took 100 shots, you didn’t have an efficient game.

By using those advanced metrics, you have a better sense of the value of a player. The disadvantage of advanced metrics in basketball is they are not very easy to understand by casual fans, so the media uses basic metrics.

In the same way your SEO clients may just be interested in total traffic and rankings for 5 keywords they think are important. While you may report the basic metrics, the advanced metrics can give you more insights about the results of your SEO actions. Advanced metrics may be “black box” (not comprehensible) for the client but “white box” (totally comprehensible) for the expert.

Metrics can also be super advanced. An example of this is the probability to win the game of GO, a metric AlphaGo used to defeat the best human player Lee Sedol. This metric is a super advanced one that only machines can come up with with months of training using black box neural networks and reinforcement learning. The result is a totally incomprehensible black box metric. Even the best GO players of the world were surprised by the famous “move 37” of AlphaGo

Systems thinking

Basketball and SEO are different types of sports. So even though you can make some analogies to the advanced metrics used in basketball to find advanced SEO metrics, you are going to miss many more. A possible approach to find advanced metrics is to think in systems.

So what is a system?

According to Wikipedia, a system is a group of interacting or interrelated entities that form a unified whole.

You can visualise a system by drawing rectangles (entities) and arrows (relations)

The end goal of our SEO system is to get new leads through organic search. The entities are the company (with website / GMB) profile), the people and the search engines. These entities are related. For each entity you can list some numbers which only rely on this entity, which are considered basic metrics.

You can link 2 directly related entities within the system and then you can find a percentage, ratio or correlation, e.g. bounce rate which combines the entity “web page visit” with the entity “search result” clicks.

When you link two directly related entities within the system, but adjust your metrics for other entities in the system, you can have a weighted advanced metric, e.g. SEO acquisition rate, which combines the entity ‘search result’ click with ‘keyword’ searches.

When you combine metrics or link non-direct related entities, you can find combined advanced metrics such as number of visitors coming from SEO who visit a page with a loading speed over 3 seconds, which combines the entity ‘visitor’ with the entity ‘webpage’.

Practical examples of advanced metrics

You may report keyword positions in a table. This is nice, but if you optimised 3 keywords, you would want to know the effect of this optimization on the amount of people clicking on your website. Assume these 3 keywords have comparable business value and conversion rate, which of these optimisations have the most value?

At first glance, you don’t know. Keyword 1 has the lowest increase in ranking.

But if you account for CTR% and search volumes, you see Keyword 2 has a low CTR in its new position and keyword 3 has a low search volume.

Therefore , you can use the advanced metric “Effective traffic increase”, which is calculated by ‘(CTR estimate (new position) — CTR estimate (old position)) * Search volume’ to easily compare the effect of these optimisations.

Probably the SEO conversion rate is higher on your homepage than on other pages, because of branded search queries and people who already know your brand.

To calculate non-branded conversion value rate, you will want to avoid putting the pages with mainly branded clicks in the equation. Another option is to weigh conversion rate for amount of branded queries. This way, you can see which non-branded pages and search terms generate the most conversion value.

If you also engage in search advertising campaigns, you can also use this data to see how effective a page is on getting value out of non-branded queries.

As you can see a centralised data system with the combination of different data sources (e.g. Google Ads, Google Analytics and Google Search console) may sometimes be necessary to calculate advanced metrics and gain better insights.

Conclusion: don’t be biased, use metrics

First of all: don’t be biased, use metrics. Use both basic and advanced metrics, do not only rely on simple metrics like SEO visits and keyword rankings, but weigh and adjust your metrics. Do think in systems. By thinking in systems you can see your strong points, your problems and your potential.

System-thinking allows you to find the causes for a drop and increase in SEO performance and evaluate values of pages and keywords using advanced metrics more quickly and easily. To do this, you may need to combine data from different sources in a central data warehouse.

But the most important advice: use the advanced metrics to change your SEO game plan and win the game.

SEO Lead @ Intracto Belgium — Python, Marketing, Data , Economics, Climate https://www.twitter.com/vdrweb https://www.linkedin.com/in/michaelvdr