Clubs should usually buy data, video, APIs, and standard workflows, then build the club-specific layer: player profiles, models, style-fit rules, and the scouting process that changes decisions.

The choice should start with components, not a yes-or-no vote on one platform. A Thoughtworks software strategy article argues for deciding build-or-buy at component level: buy repeatable commodity parts, build the parts that differentiate the organisation, and keep services loose enough to replace. Football supplier pages show those parts are already separable: Wyscout sells video, data, search, reports, shortlists, and Hudl integration; Hudl StatsBomb sells event data, models, APIs, and JSON, XML, and CSV delivery; Sportmonks sells API feeds for clubs building internal dashboards or scouting tools.

Buying is for coverage, video and standard workflow

Buy when the club needs reach quickly. Wyscout says it collects data and video from more than 1,000 competitions worldwide and lets clubs flag targets, watch video, build player lists, create reports, and send clips into Hudl. In Hudl’s Urawa Reds case study, technical director Tsutomu Nishino described a small recruitment group using Wyscout to watch matches globally, create clips, share reports, and reduce reliance on agents when evaluating international players.

Buy when the needed feature is common across many clubs. TransferLab markets a recruitment platform covering 90,000 men’s players in 100-plus leagues, 20,000 women’s players, youth data, custom metric profiles, age curves, GBE eligibility, physical metrics, salary data, and cross-league comparison. A club can build some of those tools, but buying makes more sense if the main need is broad search, filtering, comparison, and presentation.

Buy hard-to-collect inputs. Hudl StatsBomb advertises more than 3,400 events per match across 190-plus competitions, plus player-location data for 40-plus key leagues. SkillCorner says it provides 1.2 million-plus data points per game, 30,000-plus games per season, 75,000-plus players, and 120-plus competitions through tracking, physical, and game-intelligence data. For most clubs, collecting that scale of event, tracking, video, and physical data alone is not the edge.

Building is for the club’s recruitment logic

Build when the club has a model or decision rule that changes who it signs. In a Sky Sports interview, former Liverpool director of research Ian Graham said his possession-value model estimated how much each touch improved a team’s chance of scoring, and the article reports that Liverpool’s data work favoured Sadio Mane and Mohamed Salah when Jurgen Klopp initially leaned toward other targets. Graham also stressed coach buy-in, style fit, and the danger of using statistics only to support a player a club already likes.

Even strong models need scouting and context. CIES Football Observatory’s transfer-value model was built from 1,790 paid transfers between July 2012 and January 2020 and reached a correlation above 80 percent between estimated and paid fees, rising to 85 percent when buyer financial strength was added. The same report lists factors outside the model, including liquidity needs, coach disagreement, another signing in the same position, discipline, injuries, and youth-tournament performance.

That is why the club-specific layer should include rules for when data leads and when scouts can overrule it. Brentford owner Matthew Benham told the BBC that Brentford use a mixture of data models and scouting, that models are weaker for young players with limited data, and that the model should like a mid-20s player with hundreds of games before Brentford sign him.

An internal build needs staff and systems

A Training Ground Guru article written by Gemini Sports gives the buy-side warning clearly. It says clubs may be able to build a recruitment dashboard quickly, but the maintenance burden starts when a Wyscout API breaks, StatsBomb changes a data structure, SkillCorner updates tracking metrics, or the system needs fixes at 2am. It gives the example of three data staff spending 40 percent of their time on software maintenance, or £200k-plus in salary cost producing infrastructure instead of insight.

Because Gemini sells the alternative, the article should not be treated as neutral proof that buying is always better. It is useful as a maintenance test: if the internal build turns football analysts into integration and support staff, the club has bought plumbing with recruitment hours.

Club job adverts show what that maintenance burden requires when the club does build. Liverpool’s Lead Data Scientist job advert asked for football analytics experience, Python or R, database knowledge, cloud computing, statistical modelling, machine learning, model validation, productionising models, monitoring model performance, and self-service platforms for departments across the club.

The back end matters too. A Southampton data engineer advert described ownership of a cloud enterprise data platform, robust data pipelines, data quality, governance, compliance, database administration, reliability, issue resolution, documentation, and stakeholder support. If a club cannot staff those jobs, a full internal build becomes a maintenance risk.

The default should be hybrid

A practical default is to buy the inputs and build the layer that decides. Hudl StatsBomb’s API and ready-made file formats make vendor data usable inside internal systems, while Sportmonks advertises API plans starting at €29 per month for smaller coverage and custom enterprise plans for all available leagues. StatsBomb’s open-data repository also provides JSON files exported from the StatsBomb Data API, which lets analysts prototype workflows before a club commits to a larger build or contract.

Before buying, clubs should protect the exit. Turley Law’s SaaS contract guidance says customer-side agreements should specify data ownership, immediate access, portability, export format, termination access, and transition help. A sports-specific Morgan Lewis note says sports data agreements should define data boundaries, ownership, usage rights, anonymisation, purpose limits, and third-party restrictions.

Use this test: buy the part if another club needs almost the same thing, if coverage is the value, or if the club cannot maintain it. Build the part if it encodes the club’s game model, recruitment thresholds, player-profile language, scouting disagreements, or negotiation logic. Do not build unless the football department will use it, the technical team can maintain it, and the club can replace vendors without losing its own work.

Public evidence supports a decision framework, not exact budget bands or ROI. The buy-side evidence is mostly vendor product pages and vendor case studies, so claims about savings should be checked in procurement. The strongest supported answer is still clear: buy scale, build only the edge.