Providing Effective And Organized Feedback Amongst The Chaos:
Football is chaos – And your job as an analyst is to package/collect that chaos into something that is discernible and palatable. Fun, right? While that can be tough at times, having clarity in your organization juxtaposed with the game itself (not-organized/fluid, mostly) is a huge factor in supporting one’s ability to do just that. This manifests itself in a number of ways:
- Use of technology/applications to database information.
- Clear processes in your workflow so you can “automate” (without going on autopilot) the day-to-day admin.
- Knowing how and when to feedback necessary information.
In part two of the series, I’ll give my thoughts and tips on how to do exactly that. This can be a bit of a dry subject: The operations of performance analysis are much less fun than the actual analysis process (i.e. watching games/training, looking at data, interacting with other people, etc.) but your ability to do those things effectively are dramatically affected by said “administration.” Hopefully by giving examples of it in practice this will provide concrete ideas of how to improve your own efforts!
Whether you’re someone more oriented towards data or video, having said information put together in a way that is easy to pull and utilize day-to-day cannot be overstated. My first couple weeks working professionally in a performance analysis role, I learned this the hard way: I had videos strewn across my desktop, data was separated into random silos, and because of this I was late with providing reports. Simply put: Poor organizational procedures meant I spent less time doing the things I was hired for and required to do, and more time searching for the information I wanted to use. There’s no “right” or “wrong” way to database your work – Everyone’s workflow will slightly differ depending on the demands of your week and what your specific role is within a team/club. However, there are some constants to organizational skills in analysis that are constant.
There are a lot of tools, thankfully, that support these efforts of clarity – Be it video analysis software such as Sportscode/NacSport/Fulcrum Angles/SBG Matchtracker, more telestration based platforms like Coach Paint and Metrica Play, or any multitude of data platforms: Be them as basic as Excel, or complex and learning intensive as R and Python. However, in a football analysis context, there’s some slight things to consider on top of these:
- How does this information fit in with the game model (i.e. what’s happening on the pitch)?
- How do we utilise this information across multiple game weeks and the demands of a season, and the changes in personnel/results (positive & negative) that will happen?
- What are MY strengths as an analyst, and how can I best convey what I am good at?
For me personally, I always use video as the starting point: It’s what I am most comfortable in picking apart, and it’s always what is the easiest to relate to most coaching staffs. Every game is broken down in accordance with how we want to play as a team (the game model) and reviewed post match to include clips of every single individual player in the game. Concepts which are highlighted, at a basic level, are:
- Offensive Organization
- Defensive Organization
- Set Pieces
- Smaller more granular concepts of play (ex. Breaking lines, pressing jumps from wide players, etc.)
Breaking these principles of play down the same way, week by week, and staying consistent with this affects your work positively in both the short term and longer term. In the short term, your post-match analysis process has a lot of structure and can be done in a matter of moments. In the long term, it allows you to easily look at clips/video and judge how and why performances are improving or not: Essentially, databasing properly can help objectify the “subjective” content that is video review.
Data follows a similar format to video: How can I avoid separating all my metrics/numbers into very distinct silos which have no interaction with the other information I utilize. To do this, as you can guess, they need to be synthesized in such a way that encompasses the immediate/short term goals of analysis and the longer term trends.
There is a study that was conducted by the International Journal Of Performance Analysis In Sport that found coaches had very low recollection/accuracy rates when remembering key events in a match – Around 42%. The day after a match, being able to provide metrics across a spectrum of angles (How did we do in regards to chance creation? How did we defend transitions? How did Player A do with his set piece delivery?) This is like gold dust, especially when it’s done in a way which is contextualised in football (hello video!). In the long run, having all your data in one place can help determine trends: If you were only to look at at each match in distinction, you and the staff could get reactive with your decisions.
Take this Inter Milan trendline from my employers, StatsBomb, as an example. It’s a well known fact that Conte and his players made a cohesive effort to drop off their pressing habits (PPDA the metric shown) as the season went on. This change in tactical approach on the pitch, was no doubt informed by the analysis team in conjunction with a long term view of data – xG against was “poor” (in relative terms) and pressing was high. This was slowly adjusted, and allowed Inter to secure the title for the first time in over a decade. Conte is obviously a world-class coach, but data and it’s importance cannot be understated – This wouldn’t have been possible in the decision making process without an all encompassing approach to utilizing data. Lukaku helped too, I suppose.
Automating The Process
There are bits and pieces of this subsection that I’ve highlighted before, but with a good database you have everything you need at hand – The issue now is upkeep those standards of organization. Keeping your data clean, your videos/clips constantly updated and easily pulled, etc. can be cumbersome: Unless you automate your processes! Let me be clear with this: Automating your daily workflow does *not* mean you go on autopilot, and generate the same tired reports week and week. What I mean is mechanizing HOW your work is produced and put together, not the work itself. Outside tools, as I mentioned prior, do a great job of helping you clean up your processes and producing your reports in an instant. Taking my old personal working experience, this was done a number of ways.
For video, I’ve always personally used HUDL and Sportscode (other providers are available, etc.): HUDL’s service has online server hosting capabilities which mean you can upload a match, tags and all, as soon as the match is over. You can also add personal “sorters” to your video database to separate out matches, training, scouting reports, individual clips, etc. and get even more granular with the video you capture. Built in to Sportscode as well are database features where you can look at different phases of play and more that are personalized to the code windows you make/use. The outputs you can generate within them mean your scouting templates are the same, and all your videos are in one place: Saving hours of time.
Post-match data were all shared to the coaching staff in Tableau through interactive dashboards. Since getting my first job as a performance analyst, I’ve been lucky enough to have access to a data provider: To varying degrees of accuracy. Through direct API feedsbor simple CSVs our provider provided us following the match, I was able to take this data, replace it in Tableau and instantly the information from that game (as well as across past games/recent seasons) was put into the visualizations I created. Using this application (which is free to download), I could get the data on our performance into the hands of the coaching staff as soon as it was available – If I didn’t set up post match analysis this way, I would be spending hours pulling data and creating bespoke reports every single matchday +1. Tableau is a great entry level tool for this, but these similar types of mechanisms can be utilized with Excel, R, Python – you name it!
How Do I Communicate Key Concepts?
Finally, the important part: This is where analysts really separate themselves from those who are good technically – being able to understand the game and analyze it, and those who can *really* affect change within their role. In other words, how do you communicate said information. To do this, you need to:
- Know what is important for your staff, and what they want to know.
- Understand how they want that information shown to them.
- Know how and when to step in and affect change, provide feedback.
I’ll touch on all three of those points below.
The first one is rather self explanatory: You must understand how the team plays! There’s no use blindly spitting out information that is relevant: If you’re a team which likes to drop-off and sit in a deep block, don’t show them numbers about how anemic the pressing has been. This obviously takes time: You can’t jump into a club and be familiar with all the quirks/style of play in an instant. Discussions with the coach on this front are imperative, as well as asking questions of them – Something I wish I had done when I first started in football.
Now that you’ve got the type of information your coaches want, you must understand how they want it presented to them. While the concept of various different styles of learning types has long since been debunked, there’s no doubt that some people like looking at their data/video in different ways than others. I’ve made mistakes in the past where I produced reports that were incredibly detailed, but the visualizations were those which the head coach couldn’t understand. As with all presentations to the “public sphere” you must think of your audience, chiefly. If bar charts are what they like, bar charts it is. However, I’ve found from my experience that one’s ability to contextualize data in a real world/actionable way is imperative. It allows coaches to use your findings and train it on the pitch, and takes much of the translation time away.
Finally an analyst, be it video/data/or other is a member of the support staff – without being too dismissive or cruel, you must know your position in the hierarchy. It’s a tough line to balance, but you must be able to step in when needed and affect change without overstepping your boundaries. As with all areas of this subsection, it requires time in the working relationship to know what is allowed, and adjust accordingly. Touching on my own experience again, doing your job well and showing your knowledge is the best route to more “freedom” in how you interact with others. Hopefully this series will put in some building blocks for you to get better at that very thing!
The third part of this series will be a sort of “day in the life” installment: Showcasing what a match day and training day is like, and how to use the database information (as well as those touched on in part one) to become the best analyst you can be.
Alongside this, I will be updating some of the Tableau dashboards in the coming days – One of them being a sample template for an automated match report (such as the one I talked about above). I hope you’re enjoying this series, and find it useful and somewhat as an inspiration for anyone who wants to make this their career!