Name:
Thierry de Reus
Title:
Head of Tech & Data
Country:
Netherlands
Organization:
Fanalists
Description:
Nowadays personalized experiences are the norm. Personal attention, or the lack of, can make-or-break your business. Fanalists supports organizations in sports, media and entertainment to get to know their fans and to get a better gasp of their business. Fanalists breaks data silos, centralizes and enriches data, creates rich fan profiles and makes them available to analyze and communicate with. Fanalists works for event and festival organizers, sports federations like the Dutch hockey federation (KNHB), media companies such as The Walt Disney Company, and sports organizations like cycling team Team Jumbo-Visma and football club Anderlecht.
Awards Categories:
- Organizational Transformation
- AI Democratization & Inclusivity
- Value at Scale
Challenge:
The world of entertainment and sports is inspiring. Talented event organizers create the most creative live concepts and popular sports clubs create stimulating experiences for their fans. But really getting a grip on the actual individual fan? Nah. Data is generally not their comfort zone. Let alone understanding data science concepts like predictive modeling to understand their fans and customers even better.
It turned out to be hard to truly get a grasp on the power of data, customer segmentation and personalization by explaining the theory behind it. Something was missing. Something that makes it transparent, visible and usable for creative marketers. Without that missing link, organizations would never outgrow bulk marketing campaigns and generic strategies. And that would be a shame: the interplay between creativity and data results in such a powerful combination.
Moreover, Fanalists wants to support organizations in sports and entertainment in a scalable manner. Creating ad hoc analysis and customized data models for more than a handful organizations is not feasible. On the other hand, there will always be elements fully tailored to the needs of the individual organization.
How can we create a clear and scalable workflow for our own data analysts and experts? How can we achieve that, while making it understandable for the talented marketers and campaign managers on the other end of the table? How can we use Dataiku to make great use of the best features of the platform? And how can we split project-specific details and configuration from the models in Dataiku so our data flows won't grow into non-transferable messes over time?
Solution:
We created a framework that ensures that our in-house data analysts make use of a data flow that is as standardized as possible, while enabling the individual data analyst to create changes or data models specifically for an individual project.
In short: we created an extensive ETL flow, divided the flow in multiple phases and started managing project-specific settings and definitions externally as much as possible. And we launched an interface that brings transparency to the marketers and specialists at the other end of the table. That interface is called the Fanalists Terminal, where everyone working for the project, both Fanalists team members and our clients, can log in and see what is in the data model.
Phase 1: Integration Layer
This phase consists of bringing in the data and connecting to external sources. Generally this phase contains a mix of sources, such as SQL databases, SFTP files, and datasets retrieved by means of an API. At the end of this phase, all data is combined and standardized according to the conventions we defined within the framework. All columns in the available datasets are shown in the Fanalists Terminal as so-called "data fields" to enable everyone to get a grasp of what actually is in the database.
Phase 2: Context and Settings
Most data in entertainment and sports does not speak for itself. A lot of information and context is in the heads of humans. Keeping it that way does not really work for data models.
That's why everyone involved in the project can use the Fanalists Terminal to add information and context to the available datasets. Which artists were performing at the music festival last year? What type of sports event did we sell for? What are the definitions of our marketing permissions?
All this information is managed in the Fanalists Terminal, and applied later on in the data flow. This way, every project can be unique without making concessions to the standard flow.
Phase 3: Grouping and Modeling
In this phase all data is combined to create 360-profiles. After this phase, everyone involved in the project can analyze and act on rich fan profiles. Moreover, within this phase our data analysts can add prediction and clustering models.
Usually a project starts without advanced predictive models - but with a growing maturity of the organization, models could be added when the project is ready for it.
At Fanalists, we defined multiple business models and have multiple predictive models on the shelf that we can implement to match specific business models. For instance, when our client has a membership model, we can add churn prediction to the mix. But when we come across a client who is selling tickets for a yearly festival, we can add a model predicting repeat customers to the stack.
Phase 4: Segmentation
Based on the created 360-profiles, everyone involved in the project can use the Fanalists Terminal to define fan segments. Again, all so-called "data fields" are available to use and configure segments with. Because of phase 2 and 3, the existing data fields are enriched with additional information and data points related to the predictive or clustering models.
Impact:
With a clear data infrastructure, Fanalists is able to support organizations in sports and entertainment. By making it as transparent and comprehensible as possible, everything clicks into place, while also giving them the tools to play with their data. Through this combination, our clients can get the most out of their data driven marketing strategy.
1. Easily setting up the baseline infrastructure for data analysis and dashboarding
The data infrastructure makes it possible to both analyze and act on the data of their fans. With an implemented data infrastructure, including a Fanalists Terminal, marketers and strategists can capitalize fully on the segments they created by analyzing them through dashboarding, i.e. Qlik Sense and Looker. On the other hand marketers can use this information to create personalized marketing campaigns and communication flows, by syncing this information to marketing platforms like email services. Using segments based on predictive models is therefore effortless and comprehensible. And after Fanalists helped a client reach that stage, the fun begins.
2. Enabling organizations of all sizes to leverage data insights, without the need to hire specialists
It speaks for itself that analyzing and understanding their business and fans, leads to better decision-making, marketing efficiency and eventually revenue increase. Rolling out a data-driven marketing strategy not only leads to great fan experiences, it also results in more loyal fans and more valuable customers. Fanalists makes it possible for organizations without huge budgets and in-house data specialists to implement this innovative way of working.
3. Reducing efforts (and cost) to take the plunge toward data-driven marketing
The described solution is beneficial for Fanalists in improving efficiency, reducing the necessary capacity, and improving quality of delivery in the long term. Since the data flow is as standardized as possible, new projects are kicked off faster. And because enrichments, definitions and segmentations are configured by the clients themselves, there is significantly less back and forth communication. Essentially it means hitting two targets with one shot: better internal workflows result in lower costs for the client and therefore lower the risk to take the plunge. So we can create great personalized fan experiences together.
