Following their first two articles on data management and the structuring of a performance analysis department, Six Sports Management and Alexis Barat are now looking at the issue of funding such a department. What are the costs involved, and how can it be financed?
THE RESOURCE COSTS OF A PERFORMANCE ANALYSIS DEPARTMENT
A performance department must be structured with a long-term vision while keeping in mind the recurring need to address short- and medium-term requirements. It is the department head’s responsibility to ensure the sustainability of all projects through an organisation appropriately designed to support management decision-making through Data Analytics.
Structuring such a department over the long term requires the involvement of data experts to implement the necessary foundations. Data engineers build the architecture and data storage systems, data analysts prepare for future reporting needs (visualisations, dashboards, etc.), and data scientists build predictive models. Naturally, hiring such specialists externally comes at a cost: on average, between €400 and €500 excluding VAT per day of work. And with demand for Big Data expertise continuing to grow, these rates have been increasing year after year.
1. Building your analytics infrastructure
The first building block in creating a department is to standardise and consolidate the data already held by the club or flowing in through its data collection tools. Data analysts and data scientists cannot work effectively without a database that has first been standardised and cleaned. This is therefore where the data engineer comes in, building the underlying data architecture. A typical assignment generally consists of five stages:
· Identifying the requirements (defining how the data will be used, security and availability levels, identifying the type of application that will connect to the database, assessing data volumes).
· Defining the database model (transactional model and multidimensional model).
· Selecting the database management system (DBMS) and infrastructure (choosing the provider and deciding between external or internal infrastructure depending on the budget).
· Optimising the database (pre-production testing and adjustments).
· Maintaining the platform and anticipating future developments.
The first four stages require between 3 and 6 months of full-time work, depending on the quantity of data and number of data sources. Such an assignment is estimated at €8,200 per month. Once the system goes live and enters the run phase, regular maintenance is required to prevent the quality of the collected data from deteriorating. Maintenance typically requires around 3 days per month, i.e. approximately €1,230.
2. Implementing visualisation tools
Once the infrastructure has been built, the involvement of a data analyst is required to streamline the decision-making process. Creating automated dashboards makes it possible to present decision-makers with clear, structured data following extraction and analysis. The data engineer and data analyst can work together for a period of time, as the latter acts as the direct interface with the end user of the data — in this case, the coaching staff. Their understanding of the requirements is crucial in helping the data engineer build the data warehouse.
Unlike the data engineer, however, the data analyst does not need to work full-time, and the duration of their involvement varies significantly depending on the automation and reporting requirements. Take the example of a strength and conditioning coach who needs a range of dashboards to monitor players’ physical performance based on data from numerous tools. This would require between 1 and 3 months of work, including knowledge transfer and training on how to use the tools. Here again, the budget would be approximately €8,200 per month.
Finally, it is possible to bring in a data scientist to work on predictive models and statistical analysis. These specialists tend to work on longer assignments, generally lasting more than six months, and are relatively expensive (around €10,000 per month for an external consultant). Nevertheless, the results can have a significant impact on a club, although they require a substantial volume of data to have been collected beforehand.
Big Data experts’ daily rate (TJM) benchmark in February 2023
No alternative text was provided for this image
Source: Malt – Technology rate benchmark
3. Running your department on a day-to-day basis
Once the foundations have been implemented, a department must be able to rely on permanent resources capable of making the most of its datasets. These are primarily sport scientists (video or GPS analysts, strength and conditioning coaches, etc.). The most important point is to involve these stakeholders throughout the entire data project, as they represent what is known as “the business”: their work directly addresses management’s needs.
Ideally, they work closely with data analysts to automate reporting and create dashboards. They are also the people that data analysts will train to use the visualisation tools. Together with the department head, they form the permanent team responsible for reporting to decision-makers (manager, head coach, sporting director, general manager, etc.).
When creating such a department, there is no particular need to recruit additional resources. Rather, the objective is to develop or replace the existing analytical resources.
The overall idea is to combine different areas of expertise in the most effective way in order to use data sustainably. The annual cost should therefore include the salary of the department head and those of the sport scientists (interns, apprentices, junior and senior staff). These professionals are then supported occasionally but regularly by Big Data experts: data engineers, data analysts or data scientists, depending on how the business requirements evolve.
The department head must ensure proper documentation and knowledge transfer so that changes in personnel do not negatively affect ongoing projects.
Overall, based on the figures discussed above, structuring a performance analysis department therefore represents a cost ranging from €30,000 (for a club with little collected data) to €85,000 for a club looking to begin conducting its first predictive analyses based on its collected data.
Maintaining the structures put in place costs approximately €2,000 per month, or €24,000 per year — less than the cost of a junior video analyst. For clubs with limited resources, it is therefore clear that deploying such a department is accessible.
However, once again, not all clubs will have the same requirements, and therefore they will not incur the same costs. Every year, Arsenal reportedly spends around €2 million on its performance analysis department, hiring some of the most highly qualified analysts to extract insights from the most advanced tools available on the market.
The choice of tools is also a strategic consideration, and anticipating future requirements must be incorporated into project management. While some solutions are prohibitively expensive, it is also important to highlight that many high-quality tools are free or very affordable, making it possible to build such a department at a reasonable cost.
THE TOOL COSTS OF A PERFORMANCE ANALYSIS DEPARTMENT
Digital transformation has been underway for many years. As the organisation of work continues to evolve, companies have been forced to adopt increasingly powerful tools. Digitalisation is now essential for multiple reasons: time savings, improved efficiency, better communication, centralised information, improved user experience, and more.
The world of data is no exception, and the tools used to process it have developed rapidly in response to growing demand. We will identify the potential “data pathways”, i.e. all the tools put in place to manipulate data, from its creation through to its use.
First, the data needs to be generated. This can happen in different ways. Data can be generated manually — for example, when doctors take notes during consultations with players — or it can come from connected technologies (IoT), such as GPS or video.
GPS and video data analysis software is already widely available and commonly used within clubs’ performance departments. These tools make it possible to generate exploitable raw data.
Data can also come from business applications such as ERP (Enterprise Resource Planning) or CRM (Customer Relationship Management) systems. All these software solutions come at a cost, but they allow data to be generated internally.
It is also possible to obtain raw data externally through data providers such as Opta, whose business consists of compiling data on teams and players.
Secondly, the data needs to be transformed and moved. This is where the support of a data engineer is required. This process is known as “ETL”, standing for “Extract, Transform & Load”.
The “Extract and Load” processes make it possible to move data from point A to point B — for example, from GPS data analysis software to a Data Warehouse used for storage.
Open-source (and therefore free) programmes such as Airbyte or Apache Airflow can be used, although they require development and maintenance by a data engineer. Alternatively, specialised tools include Informatica PowerCenter, Microsoft SQL Server Integration Services (SSIS), Hadoop, AWS Glue, SAP BusinessObjects Data Services, Google Cloud Dataflow, Datameer, Stitch, Fivetran, and many others.
There are numerous tools available on the market, and the final strategic choice should be based on three considerations:
· Integration. ETL tools can connect to different data sources, but the choice should take into account the tool that can connect to the largest number of different sources already used within the club, or at least to the desired sources.
· Level of customisation. Some tools are more customisable and flexible than others and are therefore more or less user-friendly.
· Cost. The cost of the resources required around the tool itself must also be considered. A free, open-source tool, for example, may potentially require significantly more maintenance.
In conclusion, choosing the right tool is essential. The ETL process is mandatory in order to retrieve raw data from multiple sources and standardise the database. A paid tool generally costs between €200 and €1,000 per month, depending on the package selected (number of connectors, data processing capacity, etc.).
Third, the data needs to be stored, generally in a “Data Warehouse” (a database hosted in the cloud). The development of cloud technology in recent years has made the use of Data Warehouses simple, fast and relatively inexpensive.
Market leaders include Google BigQuery, OVHcloud, Snowflake, Azure and Amazon Redshift. There are five key points to consider when selecting such a tool:
· Integration. It is important to select a solution that can leverage the entire existing or future ecosystem.
· Reliability and support. Cloud technology has many advantages, but hosted data can still be subject to incidents. It is therefore important to ensure that the provider offers strong support and good communication.
· Performance. Data access and processing speed are important considerations when making the final choice. For example, Amazon Redshift offers a column-oriented format and massively parallel processing, whereas BigQuery uses as many resources as necessary to deliver results within seconds.
· Flexibility.
· Cost. Amazon Redshift charges based on the type of instance supporting the Data Warehouse. BigQuery charges both for the amount of data stored and the amount scanned for each query. The final cost can therefore often be difficult to predict.
Choosing a data storage solution is therefore simpler and less expensive than choosing an ETL tool. A data engineer or data analyst responsible for connecting the Data Warehouse to reporting tools can also advise on this choice. Data storage should cost between €20 and €60 per month.
Finally, the data needs to be visualised, interpreted and turned into added value. This is where the data analyst comes in. They will identify use cases with sport scientists and work on the most appropriate charts and visualisations for interpreting the data.
There are many visualisation tools available on the market, including free, open-source tools such as Google Data Studio or Metabase. There are also many paid tools, with varying price points: Power BI, Tableau, Sisense, Looker, Qlik, Toucan Toco, OpenDataSoft, and others.
The key criteria for selecting these tools are the same as those mentioned above for storage solutions. Visualisation tools generally cost between €10 and €30 per user per month.
It is also possible to use R or Python (open-source programming languages) to create visualisations. Excel also remains a long-standing tool for most organisations. However, these solutions are less easily transferable over time because it is difficult to train new resources to use them effectively. They are practical basic tools for specific use cases, but they are not necessarily “user-friendly” solutions for a structured, scalable department.
It is important to keep in mind that the choice of all tools enabling a decision-maker to visualise and interpret data must be made consistently and in a “scalable” way. In other words, the solution must remain viable, efficient and sustainable as the organisation grows.
Thus, the tools required to establish a reliable data infrastructure — once again, we are not referring here to the way clubs such as Manchester City use it — represent only between €3,000 and €15,000 per year. This is not an especially high cost considering the long-term added value of such tools.
Of course, the costs of tools such as RPE systems, data providers and others must also be added, but here again, affordable solutions are available.
FUNDING A PERFORMANCE ANALYSIS DEPARTMENT
Funding a performance analysis department is not an easy task. Such a department is not exactly a research department, but rather a means of optimising production processes. This is not generally an area for which companies receive substantial levels of financial support.
The most effective way to finance such a department is undoubtedly through a corporate partner. Whether a large company or a technology company, many potential partners could be interested in such a project.
Firstly, this is because they are likely to be receptive to an initiative of this kind if they themselves use data or if innovation is at the heart of their business model. A straightforward sponsorship agreement may therefore be sufficient to finance such a department.
Alternatively, the sponsor can support its deployment in many ways: by providing software that it already uses as part of its own business, by providing employees to help create the department’s architecture or carry out regular maintenance, or even by contributing to the analysis itself depending on the type of partner — for example, if the partner is a sports betting platform.
Furthermore, if part of the department is operated by the supporting association and used for performance development and recruitment within the youth academy, the partner may be eligible for a 66% tax reduction.
Another approach that can provide access to substantial financial support is to create a department that also has a research objective focused on sports data.
The corporate partner could potentially benefit from a Research Tax Credit equivalent to 30% of the amounts invested in the project, while the club could benefit from funding available through the CIFRE scheme (Industrial Research Training Agreements).
Such support can reduce the effective annual cost of a PhD student’s gross salary including employer contributions from €33,000 to €8,300. The company may also receive certain rights to patents or other intellectual property discovered as part of the research work.
Finally, in exceptional cases, certain regions provide support to local SMEs as part of their digital transformation. They may finance up to 80% of the amounts allocated to the acquisition of new software, for example, depending on the project’s impact on the local economy.
As an example, the Île-de-France region, through its PM’up Relance programme, can finance the strategic repositioning of an SME struggling to reach the next stage of its development, for amounts of up to €250,000 (50% of the total investment).
While financial support can appear relatively accessible when it comes to purchasing software licences, the more challenging issue remains the structuring of the department itself. For this, a technology partner appears to be the most relevant option for financing the project if the club’s cash position does not allow it to hire a specialised provider such as Six Sports Management.
We have now reached the end of our series of articles on optimising the use of sports data within clubs.
We hope you have found this series informative, and the team at Six Sports Management will be delighted to answer any questions you may have on the subject and support you in deploying such a project or assessing its feasibility.
The secret to performance in any industry is to stay ahead of the pack when it comes to adopting new tools and methods. In sport, data clearly represents a competitive advantage today, and all clubs are gradually embracing it. Getting ahead of the competition can therefore be a significant advantage.