Investing in analytics

We hear a lot of talk of the growth of analytics in football. 

It seems everyone has read “Moneyball”, has a strong opinion on the usefulness of xG, and has scouted French Ligue 2 for the next Benrahma or Mahrez. 

In the last few years many people (like us) who started messing around with data and posting bits on blogs and twitter have gone on to start working within the professional game.

I think we’ve reached the point where the evidence in favour of analytical thinking in football is undeniable. 

This is subject to the rule of the ridiculous reverse though. Nobody is suggesting that it would be better not to think analytically.

What we actually mean is that teams make better decisions when they have more, and more accurate, information to base that decision upon.

Where the analytical thinking element comes in is that this information has to be valued too. If you aren’t using it in the decision making you aren’t using analytics, just paying for it.

Assuming the club will actually take that information into account then the next question, how much will it cost?

You can scroll to the bottom if you just want that answer, for more detail read on.

First define what you want included.

Analytics tends to focus on three key areas:

• Scouting and recruitment

• Analysis of games

• Project work – unique research and macro issues

To estimate a cost we need to look at the 2 main areas where costs can arrive:

• People
• Goods and services

Your first choice is whether you have existing in-house skills and capacity. And, if you don’t, if you should hire in specific skills on a consultancy or permanent basis. 

People

Do you need a big in-house team?

Some clubs who have been very successful in the cutting edge of analytics are staffed by large teams of exceptionally clever people. 

The question for the club is, how much of their time is going to be spent on things that require the skills they have? 

The process in most clubs we have worked with is there are people who gather the information and a small group of decision makers who make the final choice.

The amount you invest in analytics has to be linked to the amount of value you place on it when making a decision.

You can have an entire department of Ivy League particle physicists researching the optimal athletic workload for a player in the club’s preferred playing system but if the coach takes no notice it has not been worth the cost.

We also need to consider just how much of the research undertaken by clubs needs to be proprietary. Just how secret does ‘secret sauce’ have to be? If you can get almost all the value from commercial products, at a fraction of the cost of inhouse development, then is the extra expense worthwhile?

Is the process more important than the output?

Most Premier League teams will have some sort of analysis department. Some of the bigger clubs in the world have huge research laboratories. But even when speaking to the staff there they put their contribution at 0.01% of the team’s performance.

So is investment in analysis worthwhile?

There is fairly decent evidence at Premier League level that goals are worth a lot of money. If £1m is spent on all sorts of “might help a tiny bit” projects such as specialist coaches, analysts, and researchers and together they add just one goal a season then it should be cost-neutral. Of course, you will never separate out that goal and be able to claim that the technical staff contributed to it.

Obviously as a company we believe marginal gains exist and should be exploited.

We also believe that you can get most of the gains for relatively little money and that can bring it into scope for teams outside the financial elite.

We use a shared cost model. If you only require the skills of a skilled data scientist or specialist goalkeeper scout for a few hours every week then why pay for one full-time? We separate out our specialists so that nobody works with two clients in the same competition, so there are no conflicts of interest when recommending players or managers.

Decisions, decisions, decisions

The choices on people are therefore a combination of:

• Existing in-house staff.
• Consultancy to support.
• External hires.

And on products:

• Free public resources.
• Commercial products.
• Bespoke external development.
• Bespoke internal development.

Putting these options together into packages allows us to give rough cost estimates that probably hold up in £, $ or Euro.

Silver (No additional cost option)

With the assumption that your club has access to Wyscout you can run a reasonably effective “technical scouting” department based on a good understanding of the search functionality.

You can use the search facilities to narrow target lists, and monitor the metrics you want your team achieving in the playing style. Combine this with bookmakers odds and public xG models and you have the outputs from advanced data models available to you at no cost. 

Information can be stored on existing club shared drives.

Gold (<50k a year)

Having now worked with clubs the idea that there are staff with the spare time to run their in-house analytics is quite unlikely. It is therefore more likely that a small annual investment (from low thousands to 50k) is enough to get you decent player evaluation and team evaluation software and, at the higher range, the additional capacity (in-house and/or consultancy) to turn this data into actionable information.

At this price you are likely to concentrate the resources on player trading, the area with the most obvious impact. With annual wage budgets running into the millions in League Two, and tens of millions in the Championship then this level of investment in due diligence and ensuring all options have been considered should be a simple decision. Within this range of budget you should have dedicated scouting databases, and a lot of video scouting time, available to you.

Elements can be added to the gold package to increase the quality of data, or depth of detail. 

Platinum (>100k-150k a year)

Once around 100k a year is committed you can add club wide elements such as dedicated document, video, and workflow platforms. One of the craziest issues in football is the lack of centralised, institutional knowledge. Far too many clubs allow knowledge to be stored in heads.

Increasing the recording of information comes with increased costs, both in staffing and systems so we should consider a budget of up to 300k annually to become a club that retains information centrally and adds to it daily.

Diamond (>300k, could easily reach £1m)

We should now have a dedicated department with access to the best products on the market heavily involved in the decision making of the club.

As the department grows it should be supplying data science skills to the whole organisation, from scouting and recruitment, to performance, sports science, medicine, commercial, media and more.

You can begin to gather your own data for analysis and run research projects.

A truly smart club who uses the power of data to aid their decisions.