In this series we have spoken about the current data tsunami, the value that can be gained from systematically analysing data, and the critical importance of individualising data to optimise the outcome. In this final article we discuss the thorny aspect of athlete data collection, data privacy, and ownership rights of this data, and present a framework to address the variety of challenges in this space.
The value and use cases of these large athlete data sets are perceived differently within the various stakeholders. What’s valuable to a team or league may differ from what sponsors and fans are interested in, or what’s useful to an athlete for training and performance analysis. When looked at in isolation, some metrics may seem to offer minimal value. However, this same data, if treated correctly or combined with other data, can offer beneficial insights and thereby huge value to the stakeholders. Today ownership is unclear, and the rules and regulations differ between sports and countries. In some cases, the club owns everything collected during regular training hours, but are not allowed to ask athletes to collect data off site, whereas other situations give the athletes the right to collect and access their data. In the case of athletes who prefer not to share their data or allow the use of their name, the situation is further complicated by the fact that in some countries, contracts are signed at the club level and not at the player level. A current controversy is that the international union for football players, FIFPRO, is licensing the names and “likeness” of the players, including for example game stats, to EA Sports for use in the game FIFA21. In both cases, the individual players may not have explicitly agreed to sharing their data or using their name. Fundamentally, a cohesive common infrastructure is lacking to practically own, collect and share athlete data.
At Svexa, we believe that athletes should own their data. They should have the option to share a part of or all the information with coaches, support staff, and external stakeholders, or decline to do so. A crucial point though, is educating the athletes on the potential benefits of sharing data with respective stakeholders, especially when it comes to optimisation and personalisation of training. If done correctly, some benefits include staying healthy and performing better, i.e. potentially longer careers and sustained performance. As we discussed in the previous article in this series Winning with Data – Athlete Profiling, data is king in athlete training optimisation and peaking performance, and none of the successful examples would have been possible without the athletes sharing their data.
Bring Your Own Data (BYOD)
There is an overwhelming quantity and complexity of data arising from athletes and sports teams. Elite athletes are among the most quantified, and data is more often collected in controlled environments. This data is an asset that can provide insights, not just for the individual and the team, but also for a broader audience. The integration of data from training sessions, performance metrics, subjective ratings, wearable sensors, molecular profiles, etc provide a gold mine for AI models.
In some sports, teams train together year round, but individual players may have short training stints with other clubs or in other countries. In other cases, individual athletes train on their own but come together prior to a competition as part of a national training camp. Yet another frequent scenario is athletes training with multiple coaches, eg. a sport specific coach, and a strength & conditioning coach outside the club. In either case, there is data that is lost or left behind, when athletes transfer from one training environment to another. In these situations, the loss of data can severely decrease the accuracy of all analytics efforts. From a coach’s perspective, the gap in data could lead to misunderstanding total load the athlete can endure, or longer learning curves for new athletes, or sometimes even starting from scratch with data collection and analytics.
Clubs and athletes may use different systems and devices to collect data. For example, a state hockey club could use GPS tracking from one manufacturer, while the national team from another, or the athlete could change their device for sleep tracking. Another hurdle is the differences in data formats, thresholds, and reported values. For example, if the number of accelerations is an important metric to track performance, one vendor could be using an acceleration threshold of 2.5m/s2 in 0.5 seconds, while another reports it for every 0.3 seconds. Comparing this data set at face value would be like comparing apples to oranges. In order to perform comprehensive analysis, the data must first be harmonised between providers. Data standards are key, and we envision that this will be an important field in the near future.
Ethics and Regulations
Historically, the team has owned all the data and the individual has had few rights or little control on their data. Are the historical employment contracts, consents and NDAs enough to own an athlete’s data? “Health” data is one of the major data streams collected on athletes. Almost across the board health data is in a special category under most laws with significantly more restrictions and control given to the individual. Control over health data by an employer, especially for actions that are not in the interest of an individual is ethically and legally a significant issue that will impact teams increasingly. There are substantial issues around the jurisdictional (what state has control) and sovereignty (who owns the data) with teams composed of international players with multiple national and professional affiliations.
There are legal and policy requirements arising across the globe that are focused on the individuals’ right to own and control the data collected on them. It is likely that individual athletes will increasingly push the envelope on these rights. Legal requirements are substantially increasing the risk/exposure/liability around breaches or non-compliance for appropriate data use. The risk isn’t just regulatory but being outmaneuvered on the competitive landscape. GDPR in the European Union has become a blueprint for data regulation across the western world. These new laws define who has access to data and for what purpose. This has dramatic implications for sports teams and the associated commercial organisations such as TV, gambling, gaming who invest huge money into sports. Teams and sports organisations that quickly and offensively act on individual data policy proactively will have advantage over those who slowly and defensively respond to individual data issues reactively.
A framework that addresses the considerations of data handling in sports, including respect of the individual athlete balanced against contextual commercial interests, offers mutual benefit. In summary:
- The individuals should own and control their own data.
- Providing value to athletes, teams and stakeholders through the collection and harmonisation of data within an open and portable passport.
- Being responsive to the legal requirements among nations to maximise the exchange of data internationally.
- Being harmonised across different situations (e.g. team, manufacturers) following the individual.
- Emphasising data standards to ensure the framework is interoperable with and open to other stakeholders.
- Extensible to multiple sports, team vs individual, professional vs recreational, and the general population.
By Shikha Tandon, Darren Montgomery, Filip Larsen, Daryl Waggott, Euan Ashley & C. Mikael Mattsson. This is the final article in the series, titled ‘Winning with data’, on sports science and analytics. The previous articles discussed optimisation in training, the differences between sports analytics and exercise analytics, insight into what data is important, and the benefit of athlete profiling. The authors are founding members of SVEXA, an exercise intelligence company, based in Silicon Valley in the United States, and can be contacted at [email protected]