How to analyse
Wyscout player data.

Start with the recruitment question

Write the brief before opening the data. A useful brief describes the role, competition level, age window, expected playing time, financial constraints and the actions the player must perform. “We need a midfielder” is too broad. “We need a press-resistant number six who can progress possession and defend transitions” creates a testable question.

Translate the brief into observable traits

Separate essential traits from supporting traits. For a progressive midfielder, essential evidence might include progressive passes, passes into the final third, passing accuracy under an appropriate volume and ball retention. Defensive actions, interceptions and duel success may support the picture. The exact combination depends on the team’s game model.

Build the Wyscout export

Run a Wyscout player search for the relevant season and competitions. Export the result as an XLSX or CSV file. Include player name and minutes played as a minimum, plus position, age, club and the performance metrics linked to the brief.

Use a broad first export

It is often useful to include every plausible position and then create narrower cohorts during analysis. This keeps the original dataset intact and lets you test how conclusions change when the comparison group changes.

• Use one consistent time period.
• Include minutes so low-sample players can be filtered.
• Prefer per-90 and percentage metrics when comparing opportunity-dependent actions.
• Keep raw contextual fields such as age, team and market value.
• Do not remove inconvenient metrics before examining the profile.

See the step-by-step export and upload walkthrough →

Create a fair comparison cohort

A percentile only describes a player relative to the other players in the selected group. Change the cohort and the percentile changes. That makes cohort design one of the most important analytical decisions.

Test the stability of the result

Run the analysis with more than one sensible cohort. If a player remains strong when the minutes threshold or age window changes, the signal is more stable. If the conclusion disappears immediately, investigate why.

Interpret percentiles correctly

An 80th-percentile result means the player’s raw value is equal to or higher than 80% of valid players in that active cohort. It does not mean the player is 80% “good”, nor does it automatically mean the underlying action helps the team.

Read raw values beside ranks

Percentiles make differently scaled metrics easier to compare, while raw values show the size of the underlying difference. Two players can sit in adjacent percentile bands even when their raw values are almost identical. Always examine both.

Use radar charts as summaries

A radar is useful for recognising profile shape, not for calculating overall quality. Check the axes, cohort and raw numbers before drawing a conclusion. A large shape can still reflect the wrong role or an unbalanced metric selection.

Move from data to shortlist

Use the data to create a review order. Start with players who meet the essential parts of the brief, then record why each player is interesting and what still needs to be tested.

1. Screen: remove players outside the genuine constraints.
2. Profile: compare role-relevant metrics and identify outliers.
3. Challenge: look for weaknesses, missing data and alternative explanations.
4. Compare: place credible candidates side by side using the same cohort.
5. Validate: review video, tactical role, physical qualities and decision-making.
6. Document: save the evidence, risks and next action in a concise report.

Talent DNA can perform the cohort filtering, percentile calculation, profile comparison and PDF reporting locally in your browser. The resulting report should support the scouting process, not replace it.

Avoid common mistakes