Chapter 28 Key Takeaways: Building an Analytics Department

Top-Level Summary

Building a successful analytics department in professional football is fundamentally an organizational challenge, not a technical one. The clubs that derive the most value from analytics are those that invest as deliberately in people, processes, and culture as they do in data and technology.

Section-by-Section Takeaways

28.1 Organizational Structures

  1. Three viable organizational models exist --- embedded (within football ops), centralized (independent unit), and multi-club (group level) --- and the optimal choice depends on club size, ownership philosophy, and organizational context.

  2. Reporting lines determine influence. The Head of Analytics should have a "seat at the table" in senior decision-making forums. Without direct access to decision-makers, the department risks becoming a reactive service desk.

  3. Analytics departments mature through four recognizable stages (Ad Hoc, Foundational, Established, Advanced), each with distinct characteristics, capabilities, and budget requirements. Understanding your current stage enables realistic goal-setting.

  4. Personnel costs dominate the budget at approximately 55% of total spend. Data acquisition is the second-largest expense at roughly 20%, followed by technology infrastructure at 15%.

  5. Structural resilience matters. Departments that depend entirely on a single manager's support are vulnerable to regime change. Reporting to the sporting director or CEO provides greater durability.

28.2 Hiring and Team Composition

  1. The Head of Analytics should be the first hire. This person establishes the vision, builds credibility, and shapes the department's culture and strategic direction.

  2. T-shaped analysts are the most effective. Deep expertise in one domain (data science, football tactics, engineering) combined with working knowledge across adjacent areas creates versatile team members who can bridge the gap between data and football.

  3. Hiring sequence matters. Prioritize roles that deliver visible value to stakeholders quickly (match analysts) before investing in technically sophisticated but slower-to-deliver capabilities (data scientists, engineers).

  4. The biggest hiring mistake is over-indexing on either football knowledge or technical skill. The ideal candidate has both, but when trade-offs are necessary, hire for the skill that is harder to teach (typically technical ability) and invest in football education.

  5. Compensation for analytics roles is rising but remains below technology industry benchmarks, creating retention challenges particularly for data scientists and engineers.

28.3 Technology Infrastructure

  1. The technology stack has four layers: data sources, data infrastructure, analytics tools, and delivery/communication. Each layer must integrate seamlessly with the others.

  2. Single Source of Truth (SSOT) is non-negotiable. All stakeholders must access the same canonical data. Multiple conflicting spreadsheets destroy trust in analytics.

  3. Buy commodity services; build where differentiation matters. Most clubs should purchase event data and use commercial dashboards while investing in-house development only for tools that provide genuine competitive advantage.

  4. In-house development carries hidden costs including maintenance, documentation, key-person risk, and opportunity cost. Build only when you have the engineering capacity for long-term maintenance.

  5. Data provider selection is a critical early decision. Use a structured evaluation framework covering coverage, granularity, timeliness, accuracy, API quality, cost, and exclusivity.

28.4 Workflow and Process Design

  1. Analytics work follows cyclical patterns aligned with matchday preparation (weekly), transfer windows (biannual), and the competitive season (annual). Workflow design must accommodate all three cycles simultaneously.

  2. Automate routine tasks aggressively. Data ingestion, report generation, dashboard updates, and quality checks should be automated to free analyst time for high-value creative work.

  3. Knowledge management prevents institutional amnesia. When analysts leave, their knowledge should stay. Version-controlled code, documented methodologies, and searchable analysis archives are essential.

  4. Maintain a structured prioritization framework. Allocate at least 20-30% of capacity to research and development work, even when reactive demands are high.

28.5 Stakeholder Management

  1. The relationship with coaching staff is the single most important factor in analytics effectiveness. Building trust requires starting with their questions, speaking their language, being honest about uncertainty, delivering quick wins, and being physically present.

  2. Every analysis must pass the "So What?" test before reaching a stakeholder. If you cannot explain why it matters, what action to take, what evidence supports it, and how confident you are, it is not ready.

  3. Lead with conclusions, not methodology. Non-technical audiences want the answer first, then the evidence. Reverse the academic convention.

  4. Resistance to analytics is natural and should be expected. Effective strategies include framing analytics as augmenting (not replacing) expertise, acknowledging limitations honestly, finding early adopters, and letting results speak.

28.6 Measuring Analytics Impact

  1. Measuring analytics ROI is inherently difficult due to attribution challenges, the absence of counterfactuals, long time horizons, and qualitative benefits. Structured measurement is essential despite these challenges.

  2. Use a multi-dimensional impact framework spanning financial impact, decision quality, process efficiency, and stakeholder satisfaction. No single metric captures the full picture.

  3. The Points Above Replacement (PAR) model provides a useful (if imperfect) framework for estimating analytics' contribution to sporting performance, aggregating across recruitment, tactical, and set-piece decisions.

  4. Long-term value creation compounds. Analytics departments build institutional knowledge, improve asset values, reduce decision-making risk, and create competitive moats that are difficult for competitors to replicate.

28.7 Case Studies

  1. Ownership commitment is the common denominator across every successful analytics operation. Without top-level championship, cultural resistance will overwhelm even the most talented analytics team.

  2. Context matters. No two clubs should implement analytics identically. FC Midtjylland's lean, set-piece-focused approach differs fundamentally from City Football Group's large-scale, multi-club operation, yet both are successful because they are tailored to their respective contexts.

  3. Patience is a competitive advantage. Building an analytics culture takes years. Clubs that expect immediate results and abandon the effort after one disappointing season forfeit the compounding benefits of sustained investment.

The Five Principles

If you remember nothing else from this chapter, remember these five principles:

  1. Analytics is an organizational transformation, not a technology project. The hardest problems are cultural, not technical.

  2. Start small, deliver value quickly, and earn the right to expand. Trust is built through demonstrated impact, not promises.

  3. Hire people who can communicate as well as they can code. The best model is worthless if no one understands or trusts it.

  4. Invest in relationships before investing in technology. A $50,000 analyst with the coach's trust delivers more value than a $500,000 platform no one uses.

  5. Measure your impact, even imperfectly, from day one. What gets measured gets improved, and what gets improved gets funded.