While some pundits argue that wearables represent little more than a short-term fad or fashion statement, market forecasts suggest the global wearables market has enormous growth potential. The recent ‘Smart Wearables Market Forecast’ from Technalysis Research suggests the market could triple in size within the next five years, rising to 175 million wearable devices and revenues of $31.5 billion.
Fitbit, a leading fitness tracking startup, filed for an IPO last week. The SEC filing revealed Fitbit sold 11 million devices last year, generating $745 million in revenue and $132 million in profit. However, as Avital Andrews pointed out in the science magazine Pacific Standard, some researchers have suggested that FitBit and other general fitness trackers are “essentially glorified pedometers” that limit users to “extremely low-level data views”.
In contrast, wearable devices designed and developed for team sports often offer users access to vast quantities of valuable data, and can include sophisticated data analytics tools to help users understand the data and improve their performance. But these advanced wearables, which can track and analyse thousands of datasets, have generally been prohibitively expensive and remained the exclusive preserve of elite athletes and professional sports teams…until now.
Dundalk-based wearables startup PLAYERTEK, which recently raised €1 million from Danu Investment Partners and entrepreneur Brendan Gilmore, brings the power of Big Data and the insights of professional-grade Big Data Analytics to amateur sports. Far from being a fashion accessory, PLAYERTEK’s innovative wearable device is custom-designed to help players and coaches track, analyse and improve performance.
Irish Tech News was recently given an exclusive interview with one of PLAYERTEK’s Co-Founders, Ronan Mac Ruairi, who discussed the startup’s advanced Big Data Analytics technology and the actionable insights it offers sportspeople. This is what Ronan had to say.
PLAYERTEK pods collect lots of performance data, but data alone won’t improve a player’s performance. How does PLAYERTEK process and analyse raw performance data to create valuable feedback and insights for players and coaches?
We set out to create a product that gives real actionable data – data that will improve your football. The wearables market is crowded with products that are too general, and they don’t solve any specific problem other than making a fashion statement.
Our focus is 100% on football (and other outdoor team sports), it’s very specifically about this application, rather than making a technology or fashion statement. Along the way we may have hit a few fashion goals and we’ve certainly had to break some new ground in terms of consumer wearable technology. In fact, we collect over 2000 numbers every second so our job is to simplify this data and to demystify the analytics process without sacrificing rigour and accuracy.
To do this we have structured our dashboard around the key tasks that players (and coaches) want. For example:
- Firstly, it’s useful to have quick feedback immediately after training or a game. For this, a few scores (simple numbers) is really the absolute most someone can use. You want to be able to compare your latest training session directly with the previous session and answer questions like…was the total volume of work similar? Was this session more intense than the last one?
You need answers that are quick to digest…for example: in this game I ran 8.5km compared to last week’s 7.6km.
- Secondly, there’s what we would call a five minute session analysis task during which a player examines a specific session in greater detail. An example might be quick review of a competitive match: by looking at the 5-minute breakdown chart a player can see if he/she maintained their performance during the last 15 minutes of the game.
It’s a great indication of fitness levels and whether a player can maintain their intensity right into the business end of game. In this task a player might also look at pitch coverage and compare first half with the second half, again looking for visual differences. They might also compare themselves with a pro player from a recent EPL match, apart from the fun side the serious aspect is that there is an opportunity to learn from role models.
- Thirdly, there’s a planning-analysis task in which a player (or coach) looks in depth at their workload across a period of time. This is a more reflective task and the question the player examines are around the amount and type of training they are doing. You might look at patterns in training volume and intensity for weeks in which game performance was excellent (maybe for games won vs games lost).
The answers should help to optimize training –there are only so many training hours in a week so a result from this analysis might be to decrease volume and to increase intensity (make sure that a small-sided game reaches peak match intensity). If your analysis shows that your power plays (during Saturday games) are lower than all your team mates you might consider extra gym or plyometric work to increase leg strength.
So, using PLAYERTEK can be as simple as measuring your distance and motivating yourself to play or train harder, or you can apply techniques that have previously would have needed a team of sports scientists .
Both – there are two levels of sharing, players can friend each other (or a coach), or a coach may have an account with several player profiles.
The first option, ‘friends’ is available to all players and is activated by request agreement between players. This allows players to see a daily and monthly graph of the key PLAYERTEK metrics. It can be used to facilitate players who play in the same team or friends you play in different clubs.
The second approach is more applicable when a club purchases a set of PLAYERTEK pods for use with a team (or several of their teams). In this case the coach has a single logon with several players associated with the account and the coach sees all the data. The coach can choose to share or show the data to individual players.
Both methods can also be mixed and matched so it’s very easy to build up a really functional team system with PLAYERTEK bit by bit as clubs and individuals purchase bits of kit.
In both cases, as well, the charts or comparisons can be shared over Twitter or Facebook so a coach or team player can feedback quickly to their team mates. It’s a really powerful, and objective way to improve the feedback and communication between coaches and players.
How much historical performance data does the PLAYERTEK dashboard store?
We store everything you measure with PLAYERTEK and the dashboard gives you views that range data inside a second, to 5 minute breakdowns, to daily, weekly, monthly and season long views. This generates quite a lot of data storage, it would not have been feasible before the advent of cloud storage & processing.
For team sports, is it possible to use multiple PLAYERTEK devices to understand how players are interacting and working together?
Yes, in the performance aspect mentioned above you can quickly see of some players are working more than others and to use that data to match with expectations of the different roles and positions allocated to each player. PLAYERTEK also provides a tactical feedback in that the pitch coverage can be examined and compared directly across players. This is a first for a consumer system and it’s a really powerful way of graphically explaining to a player where they’ve been (and where they should be) on a pitch.
Is PLAYERTEK able to suggest that a coach might want to consider playing a specific player in a different position, based on the player’s particular strengths and weaknesses?
In theory yes, but in practice we would urge caution, a lot depends on how users tag and categorise data. We can automate some of this, but the use of data always depends on the story around the data.
I think this is a lesson for big data in general, if you gather data without good context and metadata then you have the potential to read data incorrectly.
We do have a background in deep data analysis and AI techniques but for now we want to provide a platform that is accurate and reliable and sticks to presenting real information.
In our application, only the players and coaches know the full story, what role was assigned to a player, what injuries they might be carrying… and so forth. So, what we can do is to help coaches and players ask the right questions. A player might see that their heatmap is more central, like Wayne Rooney, or that he/she tends to cover the wing, leading to a more informed conversation between a coach and a player.
Given the volume of data PLAYERTEK devices capture, is PLAYERTEK intended for professional sportspeople more than amateurs?
PLAYERTEK is intended for amateurs and we think it’s time amateurs had the same facilities and tools as professionals. We used some fairly neat electronics, algorithms and cloud storage to make it possible for have amateurs to have these facilities, but more importantly we have put a lot of work into how information is presented to the
Has PLAYERTEK developed any partnerships with professional sports teams to help improve their team’s performance?
We have worked with a lot of professional teams prior to PLAYERTEK, across Soccer, Rugby, NFL and Basketball and we understand how performance data is used. As we are a startup we need to be focused and manage our resources, professional teams require a different service model so we are concentrating on bringing the power of professional analysis to amateur teams and consumers by harnessing a cloud based model.
Could professional sports scouts use PLAYERTEK data to assess sportspeople before signing them?
Absolutely, the potential is there and it is something that we will explore as we collect more data. It’s an area where big data has a key role to play in sports as physical performance is one of the pillars of
team sports but tends not to get measured for aspiring footballers.
Is the PLAYERTEK pod still useful without a subscription to go.PLAYERTEK.com?
The system needs to work as a whole, the power is in the software. Academics & scientists could license our system for raw data from the pods and chart it using Math Lab, but it’s a lot of work and requires
specialist skills – not for most people. We’ve included the first year’s subscription to go.PLAYERTEK.com and we intend to keep the on-going cost to a very reasonable sum.
Are you planning to launch an analytics dashboard for Mac and/or for mobile devices?
Our dashboard is web based and works anywhere you have internet, so Macs, Mobiles & PCs are all supported -the one caveat is that our uploading app is only on Windows right now, but the Mac version should be ready within a few weeks,
We are also working on apps for mobile devices that will give a better experience on smaller screens, that’s a few months away…iPhone first then Android.
The great benefit of the apps is that they will provide almost instant feedback on training without needing an internet connection.
PLAYERTEK’s wearable device costs €249 / £199, and includes a 12 month subscription to go.playertek.com