Big Data: Don’t Let Analytics Box You In
The sports world is using statistical analysis more and more. Dominic Pollard takes some lessons back to business
At times it can be useful, or at very least mildly entertaining, to gain inspiration for business practices from other, seemingly unrelated, areas of life. One such area which could provide some important insight into the world of data analytics is the use of ‘sabermetrics’ in modern sports.
Sabermetrics, to put it simply, is the statistical analysis of baseball records when scouting and signing new players. It’s a recruitment technique that has become increasingly popular in stat-driven American sports ever since the book Moneyball (since made into a film of the same name, starring Brad Pitt) was published back in 2003. The book examines how Billy Beane, general manager of the Oakland Athletics baseball team, used sabermetrics to assemble a very competitive side despite the club being one of the most financial restrained in the league.
Baseball metrics score a home run
Inspired by the Athletics, John Henry, now the owner of Liverpool Football Club, hired Theo Epstein, a 28-year-old Yale graduate, as his general manager at the Boston Red Sox in 2002. Epstein utilised sabermetrics and led the Red Sox to two World Series titles in 2004 and 2007.
Henry has taken this approach with him to Liverpool and sabermetrics, although a term applied to baseball, is now a technique also deployed by many leading football clubs as a means of identifying both polished gems as well as rough diamonds in the highly competitive transfer market.
This isn’t a merely a glance into sabermetrics, though. The point is that there are lessons to be learnt, good and bad, from this practice. Sport, like the vast majority of the business world, is extremely tightly contested and it is often only fine margins which separate the winners from the losers.
Finding value in unused data
Looking at the positives, sabermetrics is based on using stats that may otherwise have been ignored to give you a more accurate view of an asset, in this case a sportsperson. The same can be applied when it comes to big data. It is about extracting the value from hitherto unused data sets, whether that is because those data sets are new or have been too complex to analyse.
In football, scouts using sabermetrics – supposedly common at clubs like Newcastle, Arsenal and Liverpool in the Premier League – will amass files on players looking beyond goals scored and instead examine some of the following:
- Number of touches; indicates a player’s ability to get into the right position to receive the ball, his fitness and how willing their team-mates are to pass to him.
- Shot creation; shows how many chances they create, not simply just assists which can be misleading.
- Ball retention; helps to show the use of players who may not be as involved in attacking play.
- Balls won and percentage of successful challenges; highlights a player’s defensive ability.
- Ground covered; illustrates a player’s fitness and their willingness to work for the team.
As with the crux of big data, the objective is to uncover hidden things for competitive advantage. It is ultimately a matter of turning data into value – whether that is finding undervalued players or creating a team capable of winning trophies – and this is a fundamental approach that can be adopted by companies worldwide.
If using data analytics can show you something which other people who are not performing such analytics are unaware of then there is value in it. It may well be a case of unearthing the next Claude Makélelé, an under-rated older player who proves to be a rock in the heart of your team. Alternatively it could be finding the next Antonio Valencia, plucking a player from obscurity with all the attributes to become a great player. Either way, using all the data available for greater insight and improved decision making is vitally important.
Scoring in a marketing campaign
In the business world it might be improving the success rate of a marketing campaign by three per cent through better understanding of how customers use your website, or it may be real time visualisations of the efficiency of the machines running in your warehouses. Both can result in enhanced profits for the company. Therefore the value of using big data analytics for greater insight cannot be doubted, not that there are many who are.
However, there are warnings to be heeded from sabermetrics. Toby Moore, CTO of London-based start-up Mind Candy, was keen to stress to the Big Data Insight Group when we met him earlier this year that data cannot be the deciding factor in an organisation’s decision making process. If that becomes the case then you fast remove the human element and eventually any fun from what you do.
In sport and business alike, there is still plenty of room for gut instinct. Creating a data dictatorship can lead to decisions being made based upon anomalies, misleading data sets or in ignorance of some much-needed human evaluation.
If you’re unwilling to get your hands dirty and rummage through stats, instead taking them at face value, then it is easy to be deceived. Look at David Nugent; the Leicester City striker averages one goal a match in International football. but comes courtesy of a tap-in in his sole appearance as a substitute for an injury-plagued England team.
Luton Town’s spy-in-the-boot
At the turn of the year an interesting example emerged of how stats are driving transfers in football; Chelsea FC signed Jay Dasilva, 13, and his brothers, 12- year-old twins Cole and Rio from Luton Town’s youth squad for what could be a total of £1m.
Luton Town’s youth squad are fitted with sensors which monitor their fitness levels and movement around the pitch, data which supposedly inspired Chelsea’s triple sweep. The danger here is that data is clouding decisions and leading people to act without asking further questions. You cannot account for mental weaknesses or tactical deficiencies. Big data technology is not a silver bullet; the human brain still has an integral role to play. In this case, only time will tell if such an investment was the right call.
You cannot turn your back on the instinctive, humanistic approach. Let deeper analysis of data (which big data in essence now offers in the business world) provide otherwise unavailable insight, allowing you to improve decisions and gain knowledge that your competitors may not have or may have overlooked. But let the data work for you, don’t work for the data, lest you end up shelling out £50m on Jordan Henderson and Andy Carroll.
Dominic Pollard is editor at Big Data Insight Group .