Build or Buy? Should Your Club Build Its Own Scouting Software?

Building the first version is now the easy part. Keeping it running through every transfer window is the job.

With AI coding tools, an analyst who can code can build a working scouting tool in a few weekends: a player database, a report form, a shortlist page, maybe a chatbot on top. It looks good in a demo and it fits your club exactly. So more clubs are asking a fair question: why pay for software we could build ourselves?

We have a biased answer, because we sell scouting software. But we also built it, so we know what the work looks like after the demo. This post is about that work, so you can decide with the full picture.

Why clubs want to build

The reasons are good ones. A tool built in house speaks your club's language, with your game model, role names and ratings. Your reports stay on your own systems. There is no yearly licence. And with AI, one person can start in a month what took a team of developers five years ago.

All of this is true at the start. What matters is whether it is still true in the third season.

The demo is about 10% of the work

The first version covers the easy part: a happy path with clean data. What comes next is the work nobody demos.

Data that does not agree with itself

Every data provider has its own player ids, team names and definitions. StatsBomb and Sportmonks will not always agree on who a player is, which team they play for or what counts as a "progressive pass". Someone has to match records across providers, and fix them when it goes wrong.

It goes wrong in both directions. The same player appears twice under slightly different names, so their history is split in two. Or two different players share a name and a date of birth, and an automatic merge turns them into one person, with the wrong reports attached. We check for both all the time, and still fix cases by hand. Without the same rules, the first bad merge quietly breaks your shortlists.

Football never stops moving

Squads change every window. Loans start and end, teams go up and down, and providers rename stats or change coverage. Each change has to reach your tool, or scouts work from out of date squads. It happens every week, and the busiest weeks are the transfer windows, exactly when your staff have the least time.

Scouts are not sitting at a desk

Scouts write notes on a phone, in a stand, sometimes with no signal, then finish the report on a laptop on the train home. That means a mobile app, offline handling, sync conflicts and app store updates. A web page that works on the office wifi covers only part of the job.

Bugs, updates and AI changes

Code written fast, with or without AI, usually has no tests and little structure. It works until someone changes it, then fixing one thing breaks another. Bug reports arrive on WhatsApp on a Saturday morning, from a scout in a stadium, and the only person who can fix them is at a match too. Libraries and hosting need regular updates. If the tool uses AI, models and prices change every few months, and someone has to check the reports still sound right. By the second season, much of the builder's time goes on keeping the tool alive rather than improving it.

Built fast, not built safe

This one is easy to miss, because an insecure tool works perfectly until the day it does not. AI can write code that runs, but it does not make that code secure. It often picks the quickest setup that works, not the safest, and nobody notices unless the person building it knows what to look for. The result can be a database that hackers can reach from the internet, with every report, shortlist and player record inside it. Bots scan the internet for open databases all day, so it is a question of when someone finds it, not if. The common mistakes:

• A database open to the internet, with no password or a default one.
• Secret keys for the data provider or AI service written into the app, where anyone can find them.
• A login that hides pages in the browser, but does not stop someone asking the server for the data directly.
• Permissions that are all or nothing, so every scout, coach and outside agent sees every report.
• No backups, or backups nobody has tried to restore.

Scouting data is sensitive. It holds personal data about players, many of them under 18, which in Europe falls under GDPR. It also holds opinions, contract details and sometimes injury notes. A leak in the middle of a negotiation costs a club far more than a software licence.

None of this is a comment on your analyst. Security is a specialist job, in the same way goalkeeping coaching is. If you build, have someone with real security experience review the tool before any player data goes in, and again every time it changes. And ask yourself whether your analyst's time is better spent on servers or on football.

When people leave

This is the biggest risk, and the one clubs talk about least. Home built tools usually depend on one or two people. When that analyst gets an offer from a bigger club, and good ones do, the tool stops improving the same week. Six months later something breaks during a window and nobody knows how it works.

Often more than knowledge leaves with them. The hosting account was on their personal email, the cloud bill was on their card, and the only copy of the passwords was in their head. These are common. And it is not only the builder: when a scout leaves a club that runs on spreadsheets and shared drives, their notes often leave too.

What you actually sign up for

A simple way to compare: list the work that happens once and the work that happens every season.

Building wins on fit and control. Buying wins on everything that keeps going.

What building costs over three seasons

Here is a rough example for a club with a small scouting department. The numbers are assumptions, not quotes, so replace them with your own.

The assumptions:

• The builder is a technical analyst costing €60,000 a year including employer costs. They spend half their time on the tool in season one, and a third after that.
• Hosting and AI usage are about €100 a month each for a small department.
• A full security review before launch, and a smaller one each season after changes.

That comes to about €88,000 over three seasons. It leaves out data licences, which most clubs pay either way. It also leaves out the cost that is hardest to price: the builder leaving in season two, and the tool standing still for six months while you recruit. Put that total next to a three-season quote from any vendor, and next to what your analyst could do with the time instead.

When to build, and when to buy

Building is usually the right call when:

• You already have a data department of several people, not one analyst.
• You have a model or method that is your competitive edge, and you do not want it in anyone else's product.
• Your leadership will fund it for several years, including through a bad season.

Buying is usually the better call when:

• Your scouting department is under ten people.
• Your analyst is also doing match analysis, set pieces and opponent reports.
• What you need is a shared place for reports, shortlists and decisions, not a new model.
• You cannot afford for the tool to stall when one person leaves.

The first list describes most top clubs that build in house. The second describes most clubs outside the top divisions, and most agencies and freelance scouts.

The middle option: buy the workflow, build your edge

You do not have to choose one or the other for everything. Many clubs buy the parts that are the same everywhere, like player data, reports, shortlists, permissions and the mobile app, and keep their own people on the parts that are unique to them, like their game model, their role profiles and their own analysis.

This used to mean exports and a lot of copy and paste. MCP (Model Context Protocol) changes that. It is an open standard that lets AI assistants like Claude read data from other tools, so your analyst can work with your scouting data directly, without maintaining the system underneath.

If you go this way, ask any vendor four questions before you sign:

1. Can we set up our own positions, roles and rating criteria?
2. Can our own AI tools and analysts work with our data, for example through MCP?
3. Can we get our reports and data out, in a format we can use, if we leave?
4. Who at the vendor fixes a data problem during a transfer window, and how fast?

The answer to the third question matters most. Your reports are your club's memory. You should never be locked out of them.

Where Athleads fits

Athleads is the "buy the workflow" part. It brings all your data providers into one place, such as StatsBomb, Impect or Noisefeed, connected with the subscriptions you already have, next to official Sportmonks data, with player matching and squad updates handled for you. Your scouts stop switching between tools, and the friction goes with it. Scouts write reports on web or mobile, AI helps draft them, and shortlists and boards keep every decision where the whole department can see it, even after a scout moves on. Reports export to PDF whenever you need them outside. You pay only for the people who run your scouting, like the head of scouting, their assistant or whoever plans the schedule. Scouts writing reports and youth coaches join free, because the more they write, the more your club knows. The price does not depend on which league you play in.

For the "build your edge" part, Athleads has an MCP connection. Connect Claude or another MCP-capable assistant to your workspace, and your staff can ask questions of your players, notes, reports and boards in plain language, or use them in their own analysis. Our post on AI football scouting platforms explains how MCP works, and five signs you have outgrown spreadsheets covers the moment most clubs start looking.

If you want every tool your department uses in one place, join us.

The short version

• AI makes the first version of a scouting tool easy. It does not make the next five seasons easy.
• The hidden risks are a tool that is not built securely, and a tool that stops when its builder leaves.
• Building is not free. Our rough example comes to about €88,000 over three seasons, before data.
• Build if you have a real data department, a genuine edge and multi-year funding. Otherwise, buy.
• Or do both: buy the platform, and build your own analysis on top of it through MCP.
• Either way, make sure your reports can always leave with you.