How to Prioritise Potential Transfers: Everything Goes Better With RICE (or MICE?)
Introducing a decision-making framework adapted for transfer decisions called the MICE framework.
When you’re running a club and have to choose between potential players you can sign, how do you make that decision?
Most clubs have strong budgetary limitations and can’t simply sign who they like, when they like. Transfer teams have to make choices between whether to get a new winger or get a backup right back.
One of the best-performing sides for squad building in the last few years has been Liverpool FC.
Klopp hasn’t had the budgets to work with that City, Chelsea, or United can typically provide (transfer bans notwithstanding). Yet he, Edwards, Graham, and co have slowly assembled a devastatingly good squad.
Fans have been occasionally annoyed as the club moves slowly to find the right addition. People begged for a new left-back for years. Fear was well placed when starting last season with only 3 centre-backs. And the front 3 had wanted support for a while before Jota came in.
But Liverpool chose to operate slowly.
Robertson then Tsimikas, Konate and Jota. The holes were filled and, while Konate is too green to judge, all the others are rightly seen as bargains.
So how do you choose who to get, what position to fill, and how to prioritise on a limited budget? After all, a successful transfer is always cheaper than a failed one, as now you don’t need to keep buying players for that position, paying them crazy wages, and your overall performance levels increase - boosting revenue. Successful players can pay for themselves.
It’s super important to get the right player rather than scattergun-buy a mixed bag.
Most transfer discussion concerns individual player stats: are they good or not?
I’m starting this scenario further down the journey. You have identified your needs and highlighted the pool of players you want to choose from. Now you look at your limited funds and try to work out which to buy…
Presenting: the RICE framework.
What is the RICE framework and how does it work?
RICE is used by software companies to help make decisions about which features they want to develop. The world of SaaS is where huge tech money is going and where organizational advantages are constantly being carved out to survive in an increasingly competitive but lucrative space.
RICE stands for: reach, impact, confidence, effort.
How many users will this feature change reach? How much value will this feature add? How confident are we that our assumptions are correct? How much effort will this require for us to do?
Here’s an example:
Feature A will reach all users on the Premium pricing plan, which is 40% of all users. It’s a very high-impact feature that you have heard mentioned as the reason why potential clients decided not to choose you; this will help you close deals and expand existing business. We’re very confident in those assumptions because we have a large sample size of customer and prospect preference data. This will require its own dedicated team for a month or so and a small amount of ongoing support; it’s a medium-effort project.
So our figures are reach:40 (scale 0-100), impact:5 (0-5), confidence:0.9 (0-1), effort:0.6 (0-1 - lower is harder).
((40 * 5)*0.9)*0.6 = 108
Our RICE score is 108. You can then do this with all the other features in contention and see what the score is for each.
Some features might be dependent on others existing, and some may not feel like they fit, but generally the score will be useful.
Once you’ve done the work to understand the RICE scores for different potential features, you can rank them descending to understand what a potential priority list can look like and what factors contributed to each assessment.
This doesn’t mean you have to follow that order to the letter - it merely gives you a framework through which you can work through all variables carefully and in a considered manner.
You’ll likely find that the order the RICE scores give you is broadly what you agree on, with a few key strategic or operationally-driven changes to it.
It’s useful and aids the decision-making process.
But how do we adapt it to football?
From RICE to MICE: A model around impact per minute played
My big tweak is to change Reach to Minutes. Feeding RICE to MICE.
Minutes is an effective swap because some players are signed as starters, some as rotation, some as backup, and some as youth - the variable being how much they play. On top of that, some have a tendency for suspensions or a higher propensity for injury issues.
You can’t contribute if you’re not on the pitch. The most valuable attribute is a player’s presence.
The thing you need to optimize is the value gained per minute played by the whole team.
I could go out now and try to sign Umtiti, Dembele, Keita, Phil Jones, Gomez, Lallana, Sturridge, and Gareth Bale and have one of the most experienced and talented teams in the whole league. But we’d barely get any minutes out of them, spend a fortune, and potentially damage a carefully balanced squad morale as most of the players who do actually play the minutes see themselves as the de facto backups.
Everyone reading knows how that story ends.
So now the adapted MICE model reads:
Percentage of minutes they’re likely to play if signed?
How much impact do we think they’re likely to make in-game?
How confident are we in these assumptions?
How long have we scouted them for? How big are the sample sizes for the data we’re assessing? How can we work out how well they’ll settle?
What financial effort is needed to bring them to the club?
So, if we (low-Prem/high-Champ team) wanted to sign a Welsh winger it might look like
Gareth Bale
M = 10 (/100)
I = 5 (/5)
C = 0.8 (/1)
E = 0.2 (/1)
Dan James
M = 90
I = 3
C = 0.7
E = 0.4
Bale = ((10 * 5)*0.8)*0.2 = 8
James = ((90 * 3)*0.7)*0.4 = 75.6
Of these two, we can see that Dan James becomes the preferred option. He’ll play all the time and has youth enough to do it for a good number of seasons, and that has been the crucial winner for him.
Though, I’ve also recorded him as being cheaper than Bale - which, in all honesty, is probably not true in real life. So let’s say we run the numbers again and Dan James would cost us 25m + 100k a week, and Bale can come on a base of 75k with an incentive structure likely to average out around 125k gross. Bale’s a two-year contribution window. James might do three and we’d hope we can get the fee back with inflation when we sell him, but it still limits our budget today. Maybe Bale shifts some merch. Let’s see the scenarios where they’re equal and where James is a lot more expensive.
That financial equality pumps Bale’s number to 16 if James’ stays the same. All other numbers being equal, Bale would have to play ~47.5% of available minutes for his score to match James.
But if James is a lot more expensive and his 0.4 is reduced to a 0.1, then he comes in lower than Bale by a fraction of a point. But this depends on really how much your budget is.
An effective way to do this would be to know your budget for transfer spend and new wage expenditure, assess the player cost, and work it out as a percentage of that budget. Then invert it by judging how much budget remaining it leaves you with. i.e. if a player is 70% of budget then it leaves 0.3. Or, 90% leaves 0.1, etc.
Now, this framework isn’t really there for direct 1 v 1 player comparison.
The goal is to run it across each potential transfer you might make, and it will help direct your thoughts and assessments. Scrutinise each number and determine how sure you can be of every item.
How and when to implement MICE internally
Lots of clubs seem to go into transfer windows with a surprising lack of preparation.
It’s common to see clubs have bad windows or windows where gaps in the squad aren’t filled and deadline day rushes occur.
To fix this, we need careful prior planning and a team of people who have been preparing for the window in advance.
This team will ideally meet before the window opens and prepare a prioritised shortlist of targets, and scenarios that account for certain deals happening or not happening.
Scouting and data collection can occur continually over the long term, with the MICE model being the framework that structures how collective decision-making can happen - and identify where further study is required.
To show how this might be executed internally, using the DACI decision-making framework which determines who does what in your investigation:
Driver = Director of Football - leads the investigation
Contributors = Cheif Scout, Scouting Lead, Head of Data, Data Lead, Accountant, Head of Talent - provide the insight and scrutiny for each metric
Approver = Manager - has final say over the decisions
Informed = Board - is sent the eventual datasheet, recording of meeting, and manager choices for preliminary budgetary approval. (Manage up as well as down…)
That’s the MICE framework in a nutshell and how it could be executed. You could use this in a football club, in a fantasy league, in football manager, or even in its original version in another line of work.
The framework doesn’t have systemic flaws provided you agree with the impact per minute approach. The only question is how you measure these items and balance them player by player. The question of Impact may be different from position to position, and getting that right is likely the key to making this model work at its absolute best.
MICE isn’t here to pick players for you - it exists to facilitate more intentional, collaborative, and scrutinised decision-making processes; something many clubs are very clearly lacking.
If you enjoyed this piece, subscribe below for more infrequent newsletter updates:
Or share this with other people who you think might enjoy it: