Accurately assigning credit is incredibly difficult. In football we may look at a team of eleven players who have just won a game by one goal to nil and try to break down what their individual contributions were. The headlines will always be about the goal scorer. He scored the winning goal so he is most responsible for the win? But what if the goal was the result of the winger beating five players and passing it to the scorer with the net empty? Surely that was a bigger contribution? What if the goalkeeper made some world class saves? Doesn’t that deserve more credit? What if the defence or midfield were so good the goalkeeper didn’t have a shot to save? Which midfielder or defender was most responsible?  I think most reasonable people would agree football is a team sport and in teams individuals do all contribute at different rates in different circumstances. Some games it may be the goalkeeper who is the hero, some games the left back. The important thing is that the collective works. Yet some players are clearly better than their teammates. These players rise through the levels as that talent is identified. Football is quite meritocratic in that way. There is no hierarchy to navigate, no boxes to tick, no interview panel, you simply get independently assessed on your observed performance. Data and video help massively. Even at this fairly nascent stage of football analytics with only event data you can identify talented individuals within teams pretty accurately after only a few games. Jump onto Instat or Wyscout and verify with your eyes and soon you can pull together a list of players who are capable of playing at a higher level. The difficult bit is matching players to roles and finding the best move for them for their current ability level. A good concept to keep in mind is the zone of proximal development. Players need to play at the level they are just good enough to get regular minutes at to maximise their development. This is a huge, and underused, advantage available to multiclub networks.  But we don’t have data and video for “thinking” jobs If I am coaching track cyclists I can measure their output to ridiculously precise levels.   This is also true of traditional factory work. Piece-rate work where workers are assessed on how many units they process is incredibly common. Problems occur when you start adding externalities into the performance measures. For example you may try to rank surgeons by their performance. Would you rather be operated on by Mr Smith with a 95% survival rate or Mr Jones with a 60% survival rate?  The higher the better surely? But what if I told you the expected, pre-operation survival rate of the patients was 100% for Smith’s and 20% for Jones’? It gets even trickier when we move past measurement of individuals pursuing solitary enterprise into people working within teams. Imagine a sales team, another place where data can be, and is, used. You have 100 sales leads and 10 people in the team. How do you assign the leads? Do the best sales people always get the best leads? If so, how do you know they are the best, not just the ones given the best chances?  Sales is still very much output driven. You have a single figure you can, rightly or wrongly, attribute to an individual. What about long term projects? For example look at youth development at a football club. This is a multi-step project including: Initial talent identification: Someone has spotted the player. Signing the talent: Someone has persuaded the parent/player to sign up with the club. Coaching in multiple age groups: Someone has worked with the player and helped improve them. Player retention: Someone has persuaded that player to remain with the club (good players always have options). First-team debut: Someone has put that player on the pitch for the first-team. Often football clubs look enviously at the youth system of another club and poach one of the staff from that club. This may be someone from the talent identification system. But does this work?  The Matthew Effect is something we should always consider. If we only give opportunity to people with “track records” we risk missing out on people with better ideas. How do we know the person with time in the Barcelona academy is better than the one with Bradford? Much like the surgeon example earlier, it may be a bigger achievement to have got 3 players all the way through the system into regular League Two football than it was to get Barcelona’s star players in.  And anyway is taking an individual person from a successful system and putting them into another system going to always work? Clearly not. You need someone to have correctly diagnosed the problem with the system. A systemic approach? Going back to our youth system process we first have to ask where the system is failing. This is where I feel some clubs lack good quality user-centered information. Saying “we need to develop our own players” is simple. But diagnosing why that happens is often missed out. Take two clubs, Club A and Club B Club A has a thriving youth system with 5 current first team players and millions in the bank from player sales. Club B has only a few youth players but none are regulars, the academy runs at a huge loss. Simple isn’t it, Club A is good, Club B is bad. Club B should take some of the talent spotters and coaches from Club A? What if Club B finishes higher in the league than Club A almost every year? Maybe it is just easier to get into the Club A first team? Maybe Club A have a structure in place where they have a stated aim to get players into the first-team whilst Club B prioritise winning above all. To correctly diagnose the problem we would need to look … Continue reading Game Changing Investments – Coaches as surgeons