How football scouting works, from data screen to signing

The seat is high in the corner of a stand at a ground holding a few thousand people, chosen because nobody else wants it and because the whole pitch is visible from there. The man sitting in it is not watching the ball. He is watching a left-sided defender who is forty metres away from the action, and he keeps watching him during the eight seconds after the ball goes out for a throw, because that is the window in which players stop performing and start being themselves.

That image is what most people picture, and taken alone it gives a badly wrong impression of how football scouting works. The live viewing is real, and it matters, but by the time it happens the club has already spent several weeks removing thousands of other names from consideration without anyone leaving the building. Scouting at a professional club is a funnel, and the part with a stopwatch and a notebook sits near the narrow end of it.

The short version is this. A search is defined by a role rather than a position. A data screen turns the global pool of registered professionals into a few hundred candidates who fit that role, the age band and the money available. Video work cuts those to a shortlist. Live scouts then watch the survivors several times, in person, looking for the things a camera crops out. A recruitment meeting turns the shortlist into a ranked list with prices attached, and a sporting director turns that into a bid. Everything else is detail, and the detail is where the interesting failures live.

The four different jobs that all get called scouting

The word covers at least four distinct functions inside a club, and confusing them is the source of most public arguments about recruitment.

Talent identification is the search for players the club does not yet know about. It is the classic job, it is heavily weighted towards lower divisions and younger age groups, and its output is names.

Player assessment is the evaluation of players the club already has on a list. Its output is not names but judgements, delivered in a format the recruitment meeting can compare across candidates.

Recruitment analysis is the quantitative side: building the search filters, checking whether a player's output is the kind that repeats, modelling how a signing changes the squad's age profile and wage structure, and putting a defensible number on what he is worth.

Opposition analysis shares a skill set with all of the above and has nothing to do with transfers. The person who catalogues next Saturday's opponents is often in the same department, using the same video tools, and their work never touches a signing.

A club that has one person doing all four is not scouting badly, it is scouting cheaply, and the trade-off is visible in what gets missed. Most clubs of any size separate at least the transfer-facing work from the match-facing work, because the deadlines are different and the second will always eat the first.

• 211FIFA member associations a search can theoretically reach
• 2Registration periods a club normally gets in a season
• 18Minimum age for most international transfers
• 3Defined exceptions to that minimum age rule

Regulatory and structural facts rather than any club's figures. Registration periods and the age rules come from FIFA's Regulations on the Status and Transfer of Players.

Where the list starts: the data screen and what it can and cannot filter

The first filter is a database query, and understanding what it is doing removes most of the mystique from modern recruitment.

The input is a role specification rather than a position. "Centre-back" is not a search. "Left-footed centre-back who defends a high line, is comfortable stepping into midfield with the ball, and has played at least eighteen hundred minutes in a possession-dominant side in a top-fifteen European league, aged twenty-two to twenty-six" is a search. The second version can be written as filters. The first cannot.

On top of the role sit the constraints that have nothing to do with football: an age band that protects the squad's resale value, a wage ceiling, a fee ceiling, and a set of leagues the club believes it can buy from without paying a premium. Those constraints usually do more filtering than the football ones.

What comes out is a longlist, and its size is the first honest signal about a club's process. Too short and the filters were written to produce the names somebody had already decided on. Too long and nobody will watch them.

The screen has three limits worth stating, because they are the reason the rest of the pipeline exists.

The first is that data describes what a player did inside a particular team, not what he is capable of. A midfielder in a side that dominates the ball will produce passing numbers that a midfielder in a side defending a low block cannot produce, however good he is. Adjusting for that is possible up to a point and impossible past it.

The second is that the volume of data available falls off a cliff below the top divisions. Detailed event data exists for a decent number of leagues; tracking data, which records every player's position several times a second rather than just the on-ball actions, exists for far fewer. The further down the pyramid a club looks, the more it is relying on eyes rather than files, and the leagues where the best value hides are exactly the ones with the thinnest data.

The third is that everyone has the same screen. The metrics that identify an underpriced forward are not secret, the data vendors sell to everybody, and a name that a filter surfaces has probably been surfaced at fifteen other clubs the same week. Competitive advantage in recruitment does not come from the query. It comes from what the club does with the two hundred names that survive it, and from being able to move on one of them faster than the rest.

What a live viewing is actually for

Given the video available, the obvious question is why anybody still buys a plane ticket. The answer is that a broadcast feed is an editorial product and a scout needs an unedited one.

Television follows the ball. It has to, because that is what viewers want, and the consequence is that roughly nine tenths of what a footballer does during a match is off camera. A scout in a stand is watching the player, not the game, and can therefore see the things that only exist in the empty seconds: the scan over the shoulder before the ball arrives, the two steps taken to close a passing lane that was never used, the position taken up when the team loses the ball and the player is nowhere near it.

Live viewing also picks up a set of signals that no camera angle captures reliably.

The condition of the player. How he moves in the warm-up, whether he is favouring anything, what his body actually looks like at close range rather than in a compressed broadcast image.

The relationship with the group. Who he talks to, whether anybody listens, what happens between him and a teammate after a misplaced pass, how he behaves when he is substituted.

The response to going behind. A scout will deliberately choose a fixture the player's team is likely to lose, because the useful information is in how a player behaves when the game is not going his way. A hatful of assists in a five-nil win tells you very little that video would not.

The environment. The state of the pitch, the size of the crowd, the level of the opposition, and the honest question of whether any of this resembles what he would be walking into.

The good scouts do one further thing that sounds trivial and is not. They watch the player for a defined stretch of the match without looking at the ball at all, usually a block of ten or fifteen minutes, and they write down what he did in that period. It is the only way to get an unbiased sample of behaviour, because the alternative, watching him whenever he is involved, guarantees a sample selected on involvement.

Inside a scout report: the sections, the grades and the recommendation

Reports vary between clubs, but the working format has converged, because a recruitment meeting has to compare a report written in Portugal against one written in Denmark and cannot do that if every scout writes an essay.

A typical report carries five parts.

The context block. Fixture, date, competition, weather, pitch, opposition standard, the player's position and role on the day, the minutes he played, and the state of the game while he was on. Without this the rest of the report cannot be weighted.

The graded attributes. A fixed list of qualities, each scored on a short scale, with the list varying by position. A defender's list will include aerial ability, defending in behind, one against one, positioning in the block, distribution under pressure and aggression on the front foot. Short scales are used deliberately: a ten-point scale invites false precision and clusters at six and seven, whereas a four or five point scale forces a scout to commit.

The written sections. Strengths, weaknesses, and the two that carry the weight, which are the description of the player's role in his current team and the assessment of how he would function in the signing club's structure. This is where a scout earns his money, because the transfer risk is almost always about transplantation rather than quality.

The projection. Where the scout thinks the player will be in two and three years, given his age, his physical profile and his rate of improvement. A report on a twenty-year-old is mostly this section. A report on a twenty-nine-year-old barely has one.

The recommendation. Almost always a ladder rather than a yes or no, with rungs along the lines of: not for us, keep on the list, watch again, shortlist, sign. The ladder matters because it separates a scout's confidence from his enthusiasm, and because it lets a department track its own decisions afterwards.

The most useful discipline any recruitment department imposes is that the recommendation must be time-stamped and never edited. A report changed after the player's transfer collapses or succeeds is worthless as a record, and without the record there is no way to find out which scouts are actually any good.

15%30%20%15%10%10%

• Warm-up and the player's physical condition15
• Uninterrupted blocks watching the player away from the ball30
• Following the player's on-ball actions in real time20
• Defensive phases and reaction to losing possession15
• Behaviour after mistakes, substitutions and stoppages10
• Writing while the match is still running10

An illustrative allocation of attention across a single live viewing, not a measurement and not any club's protocol. The point is that a scout is running a plan, not simply watching a match.

Show the numbers

Constructed example: how one scout allocates a ninety-minute viewing

Item	Value
Warm-up and the player's physical condition	15
Uninterrupted blocks watching the player away from the ball	30
Following the player's on-ball actions in real time	20
Defensive phases and reaction to losing possession	15
Behaviour after mistakes, substitutions and stoppages	10
Writing while the match is still running	10

The video stage, and why the honest number of matches is high

Between the data screen and the aeroplane sits video, and it does most of the work of reducing the list.

Clubs subscribe to platforms holding full match footage and tagged event data for a very large number of competitions. The tagging is the useful part: an analyst can pull every defensive action a specific player made in his last fifteen matches into a single sequence of clips, watch it in twenty minutes, and see patterns that would take fifteen full matches to notice.

That is also the trap. A tagged clip reel is a selected sample by construction, and selected samples flatter. Watching eighty clips of a winger beating a defender proves that he sometimes beats defenders. It says nothing about how often he tried and failed, what he did in the seventy minutes between the clips, or whether he tracked back.

Serious video work therefore alternates between two modes. Clip reels answer specific questions: does he take the ball on the back foot, which shoulder does he scan over, is his weak foot usable under pressure. Full matches answer the general question, which is what he is like when nothing is happening. Both are needed and the second is the one that gets skipped when the window is closing.

On volume, nobody publishes a policy, but the underlying statistical problem is well understood and points in one direction. Football has few scoring events, high variance and long stretches where a player is uninvolved. A single match is an extremely small sample of anything. Any honest process therefore wants a player seen across a substantial part of a season, in a range of fixture types, before it will trust its own conclusion. The clubs that get this wrong are usually the ones under time pressure, which is why signings made in the last seventy-two hours of a window have such a poor reputation. The mechanics of that deadline are covered in the piece on how the football transfer window is structured, and the compression it creates is a scouting problem before it is a negotiating one.

Recruitment analysts and traditional scouts do different halves of the same job

The public version of this is a fight. Inside a functioning department it is a division of labour, and the division falls along a clean line: the analyst is better at search and comparison, the scout is better at explanation and fit.

An analyst can rank four hundred right-backs across nine leagues by a defined set of adjusted metrics in an afternoon. No scout can do that, at any level of talent, because the volume is simply beyond human viewing. The analyst can also answer questions about repeatability that a viewer cannot: whether a striker's finishing is running ahead of the quality of the chances he is getting, which is the whole point of the framework described in what an expected goals model is actually counting, or whether a midfielder's pressing numbers reflect his own habits or his team's instructions, which is the distinction the metric in the pressing measure built on passes allowed per defensive action tries to isolate.

What an analyst cannot do is explain why. A model can say that a player's output dropped in the second half of last season. It cannot say that his father was ill, that the manager changed the system in October, or that he was playing through a problem he did not report. A scout can find that out, usually by talking to people, and it changes the interpretation completely.

The productive arrangement puts the analyst at the front of the funnel and the scout at the back, then treats disagreement between them as a signal. A player the data loves and the scouts dislike is worth an argument, because one side is seeing something the other cannot, and the argument is cheap compared with the transfer. A player both sides love is usually expensive, because everyone else's data and everyone else's scouts have reached the same conclusion.

The broader structure this sits inside, the reporting lines and where the analysts actually sit, follows the pattern set out in how a sports analytics department is put together, and football's version arrived later than most.

1. 1The role briefThe head coach and sporting director agree what the team is missing, described as a function within the system rather than a shirt number. The brief carries an age band, a wage ceiling and a fee ceiling, and it is written down so that it cannot be quietly rewritten later to justify a signing.
2. 2The data screenAnalysts turn the brief into filters over event and tracking data and produce a longlist. Leagues the club believes it can buy from without a premium are weighted up; leagues with thin data are handed straight to human scouts instead.
3. 3Video reductionAnalysts and video scouts work through the longlist with clip reels for specific questions and full matches for general ones. Most names die here, and the reasons are recorded so the same player is not longlisted again next window.
4. 4Live viewingScouts watch the survivors in person, several times each, deliberately including away fixtures and matches the player's team is losing. Reports go in on a fixed template so that candidates watched by different people remain comparable.
5. 5The non-football checksMedical history, injury pattern, contract length, agent, wage demands, work permit or visa status, and character reference from people who have worked with him. Any of these can end a candidacy on its own.
6. 6The recruitment meetingScouts, analysts, the sporting director and the coach rank the shortlist and attach a maximum price to each name. The ranking matters more than the top name, because the top name is usually unavailable.
7. 7Valuation and offerThe finance side converts the fee into an annual accounting cost across the contract, checks it against the wage bill and the profitability rules, and the club bids. A rejected bid moves the process to the next name on the ranked list rather than back to the beginning.
8. 8Registration and integrationThe player signs, is registered inside a registration period, and is handed to the coaching staff with the scouting file attached. The best departments then keep tracking him for two seasons and grade their own decision against what actually happened.

The sequence a structured recruitment process runs through. Every stage is designed to remove candidates before the next, more expensive, stage begins.

Signing a role, not a player

The single most common technical error in recruitment is evaluating a player against a position instead of against a job.

Two central midfielders in the same league can have almost nothing in common. One receives the ball facing his own goal in a team that plays out under pressure, and his defining skill is turning out of contact. The other receives it facing forward twenty metres higher up in a team that wins the ball back high, and his defining skill is arriving in the box late. Both appear in the same database category. Neither can do the other's job well.

Role profiling fixes this by writing the specification in terms of actions and locations rather than labels: where does this player need to receive the ball, what does he need to do with his first touch, what does he need to do when the team has the ball on the far side, what is he responsible for when possession is lost. A candidate is then measured against that description.

The second half of the same problem is transplantation. A player's numbers are produced by the interaction between his ability and his team's structure, and moving him changes the second term. A full-back who looks outstanding in a side that builds patiently from the back may look ordinary in a side that goes long, not because he got worse but because the situations that suited him stopped occurring. The style of build-up he is arriving into, of the kind described in playing out from the goalkeeper and what it demands of defenders, is part of the assessment rather than context to be noted afterwards.

The best defence against this is a specific and uncomfortable question that good departments force scouts to answer in writing: which of this player's current strengths depend on his current team, and which would survive a move to ours? Scouts who cannot answer that are describing rather than assessing.

The medical, physical and age filters that end candidacies quietly

A large share of shortlist deaths have nothing to do with football, and they happen in a stage that is invisible from outside.

Injury history is examined for pattern rather than volume. A player with one long absence from a single traumatic injury is a different risk from a player with six short absences from recurring soft-tissue problems, even if the total days missed are similar. The second pattern is the one that tends to repeat.

Physical profile is checked against the demands of the destination. A league that plays more matches, at higher intensity, in a different climate, with a winter schedule the player has never experienced, is a real change in load rather than a detail.

Age is a financial filter as much as a sporting one. A fee is spread across the length of the contract in a club's accounts, which is the mechanism explained in how transfer fees are written down over a contract, so a long contract on a young player produces a low annual cost and a resale asset, while a short contract on an older player produces a high annual cost and nothing at the end. Clubs operating under strict profitability rules of the sort covered in the Premier League's spending regulations end up with age bands in their briefs for reasons that are entirely accounting-driven.

Contract status decides whether the club is buying a player or negotiating with one. A player with eighteen months left is a discount; a player with four years left is a full price and a long negotiation.

None of this is scouting in the romantic sense and all of it removes more names than the scouting does.

Character work, and the part of the file nobody puts in writing

Every recruitment department does character checks and almost none of them describe how, because the process is informal, occasionally unfair and legally awkward.

The mechanism is a network. Scouts and sporting directors accumulate contacts over decades: former teammates, academy coaches, physiotherapists, kit staff, journalists in the player's home city, agents who represent other players at the same club. A quiet conversation with three of those people produces a picture that no data source contains.

What the club is trying to establish is narrower than "is he a good person". It is closer to a set of practical questions. Does he train at the same level he plays at? What is he like in the eighth week of a bad run? Will he accept a season on the bench, and if not, will he say so honestly? Is the family move going to work, and has anyone actually asked the family?

That last one is underrated and expensive. A significant number of signings that are described afterwards as adaptation failures were nothing of the sort. The player settled fine and his family did not, and no amount of video work would have found it. Clubs that take relocation support seriously do it because they have paid for the alternative.

The obvious hazard is that informal reference-gathering is exactly the mechanism by which a reputation, deserved or not, follows a player for a decade. A single dressing-room fallout can turn into a permanent label applied by people who were not there. Departments that are honest about this weight character intelligence carefully, ask what the source's own interest is, and treat one negative voice as a prompt to check rather than a verdict.

Youth scouting, and the rules that fence it in

Academy recruitment is a different job with a different clock, and the regulations do a great deal of the shaping.

The central constraint is FIFA's protection of minors. International transfers of players under eighteen are not permitted except in a small number of defined situations, principally where the player's parents move to the new country for reasons unrelated to football, where the move is inside the European Union or European Economic Area for a player aged sixteen to eighteen and the club meets specified obligations around education and accommodation, and where the player lives close enough to a national border that the move does not involve leaving home. The effect is that scouting a fifteen-year-old abroad is mostly a matter of building a relationship with a player a club is not allowed to sign yet, and waiting.

Domestic youth recruitment is fenced differently, by national rules on how far a boy of a given age may travel to train, on compensation between clubs, and on how and when a club may approach a player registered elsewhere. Those rules differ by association and change often enough that a specific number here would be out of date quickly.

What does not change is the underlying difficulty, which is that youth scouting is prediction over a longer horizon than anything else in the sport. A fourteen-year-old's current performance is heavily contaminated by his physical maturity relative to his peers, and the boys who are twelve months ahead in growth dominate age-group football for reasons that will have vanished by twenty. Academies that have thought about this look at technical habits, decision speed and the ability to solve unfamiliar problems, and they explicitly discount the physical dominance that a scout's eye is naturally drawn to.

The economics also differ. An academy is a long, slow, expensive bet in which most of the intake will not make it, and the honest justification is not that it produces first-team players in useful numbers. It is that it occasionally produces one, that a home-grown player carries no transfer fee in the accounts, and that a sale of one is close to pure profit under the spending rules. That is a financial argument, and it is worth being clear that it is one.

Players in the target leagues and age band900

Surviving the data screen against the role brief140

Surviving video reduction to a working shortlist25

Watched live at least three times9

Cleared on medical, contract and character5

Bid for2

Signed1

Illustrative arithmetic on stated assumptions, not a measurement and not any club's figures. A search for one role begins with every registered professional in the club's target leagues and ends with one registration. The point is the ratio between stages, not the values.

Show the numbers

Constructed example: the shape of a single position search

Item	Value
Players in the target leagues and age band	900
Surviving the data screen against the role brief	140
Surviving video reduction to a working shortlist	25
Watched live at least three times	9
Cleared on medical, contract and character	5
Bid for	2
Signed	1

Why signings fail, in rough order of how often

Recruitment departments that grade their own past decisions tend to arrive at a similar taxonomy, and it is not the one supporters use.

The role stopped existing. The player was recruited for a system, the head coach changed, or the head coach was replaced, and the new plan has no job matching the one the player was bought to do. This is a structural failure of the club rather than of the scout, and it is the reason clubs with a defined playing identity have better recruitment records than clubs whose identity resets every eighteen months.

The output was contextual. He was good, in that team, doing that job, with those teammates. The assessment failed to separate the player from the environment, and the environment did not come with him.

The sample was too small or too selected. Three matches, a clip reel and a deadline. The club bought a good week.

The physical demand changed. More matches, more sprints, a different intensity, a different calendar. The player is not worse; he is unavailable.

The price became the problem. A large fee and a large wage make a player undroppable for reasons of face and of accounting, so he plays through poor form, the form does not recover, and the club has manufactured a problem it could have avoided by paying less. This failure mode is invisible in the scouting file because it was created after the file was closed.

Nobody managed the arrival. No plan for the first three months, no clarity about what he was expected to do, a language barrier nobody resourced, a family left to sort itself out.

The player was simply misjudged. This happens, and it is the smallest category on the list. Any department where it is the largest category is not really running a process.

A useful correction against most of the above is the loan market, which lets a club test a player in its own competition without buying him, and which is now bounded by its own set of limits and clauses. Those are set out in the rules that govern loan moves and what clubs can write into them, and a loan with an option is often the honest way to price uncertainty that scouting cannot remove.

Judging a club's recruitment from outside the building

Supporters cannot see the reports, but the process leaves fingerprints, and a few of them are readable from a distance.

Look at whether the signings fit a repeated profile. A club with a working process signs players who look like each other in age, role and league of origin, because they came out of the same brief. A club signing a twenty-year-old winger from one country, a thirty-two-year-old holding midfielder from another and a striker whose agent also represents the manager is not running a funnel.

Count the deadline signings. A club whose business is concentrated in the final days of a window is either being outbid all summer or has no ranked list. Both are recruitment failures.

Watch what happens when a bid is rejected. A department with a ranked shortlist moves calmly to the next name. A department without one either overpays or ends the window with the position unfilled.

Check the age curve of the squad, not the last signing. Recruitment is a portfolio. A club that has signed nobody under twenty-five for three years has a problem that will arrive all at once.

Give it two seasons before judging. The honest evaluation window for a signing is longer than a single campaign, and the clubs that grade themselves properly do it retrospectively, against the recommendation that was written before anyone knew the answer.

The uncomfortable truth underneath all of it is that recruitment is a probability business. A department can run a rigorous process, ask the right questions, watch the right number of matches, and still sign a player who does not work, because football has enormous variance and a squad of twenty-five is a very small sample. The point of the pipeline is not to be right every time. It is to be wrong less often than the clubs that guess, and to be wrong more cheaply.

For more on how the machinery around these decisions actually functions, from windows and valuations to the spending rules that bound them, the rest of our football coverage works through it piece by piece, and the wider article archive covers the same ground across other sports.