Practical advice on how to build your data team
Explore the whys and hows of building the data team at Libeo
For several months now, we’ve been building the data team at Libeo.
During early calls with recruiters or candidates, our data ambitions were often met with surprise given Libeo’s development stage.
So we wanted to share the whys and hows behind our data team.
Why invest in a data team early on?
Most startups do not invest as early in their data team.
At Libeo, our product is at the center of everything we do. We rely on it as the primary driver of acquisition, activation and retention. We believe we will not be able to fulfill our ambition for Libeo without a strong data informed culture.
Try this little experiment: ask around your company what is the definition of an active user and how many they currently are, and you’ll probably be surprised by the different answers (some really out of the ballpark) and the number of ‘I don’t know’.
This has been a challenge at Libeo since we’ve been more than 5 people. And it only grows.
The root causes for this active user problem are the lack of a centralised and verified source of data accessible and siloed teams. When the data is not centralised and verified, it is impossible to know where and how to query the right tables to generate the accurate metric. Siloed teams create bias in the metric definition. Growth teams will move the definition closer to the activation objective while success teams closer towards retention.
Here is where the data team comes in. It has one simple and ambitious objective: empower teams across Libeo with better insights to solve their daily business problems and make better decisions.
What are the data team’s missions?
The data team missions are twofold: business intelligence and data products.
Provide Intelligence across all teams
The first and foremost mission is to be the dedicated data partner for key topics around Libeo.
We have organised those topics in 4 areas:
- Growth: From marketing and growth to activation and sales.
- Users & usage: User knowledge, product and design decisions, engagement, retention and partners’ usage.
- Ops: User care, internal ops, fraud detection and data-enabled processes.
- Business monitoring: finance, internal performance monitoring and investors reporting.
Libeo teams need to quickly make strategic and impactful decisions, and for that being data-informed is critical. We build, analyse and share the minimal data stories that maximise those decisions.
As we are growing fast, monitoring and reporting is also key. This allows us to proactively spot problematic trends and iterate quickly to improve.
Build great data products
Data products are about creating value with either tools for internal users or product features for our users.
The first product we are responsible for is (obviously) our data architecture, from extraction to loading, transformation and provision to all teams. Going back to our initial lack of centralised and verified data, we built a simple, scalable and self-serve data architecture (but more on that in another article).
They are 2 main areas of applications for data products:
- Data-enabled tools and processes for internal teams: our goal is to automate data-related tasks that we used to do manually from fraud detection to Know Your Customer requirements and activation or churn predictions.
- End user-facing data features: in close collaboration with product and engineering teams, we build data-powered features to improve the user experience such as document tagging or relevant data-sharing with our partners.
How is the data team organised?
One key question we faced was how to structure the team to answer Libeo business needs at scale.
Most data teams follow one of two models: centralised or decentralised — if you’re interested in this topic, here is a great read from dbt. Both have advantages and drawbacks, and as often, the optimal answer is the best of both worlds.
Our data team is a centralised team with decentralised senior data members integrated within squads related to the 4 areas of application detailed earlier (Growth, Usage, Ops and Business monitoring).
The centralised component is key to tackle the siloed teams and provide a single source of truth to the whole team. It is responsible for coherence, good practices and knowledge sharing across all teams.
The decentralised component in each area of application gives us a clear understanding of data problems on specific business use cases. The truth lies within the end user, and that’s why it requires senior data members to be directly involved in the business areas. They typically meet on a weekly basis to share insights and feed our data roadmap continuously. This relationship makes them a privileged point of contact. They are responsible for understanding the business questions the data team must solve, prioritise them according to our OKRs and provide analysis. The data team as a whole makes sure the data roadmap is aligned with our objectives and vision.
Iterate is key to improve
This is our first iteration. We anticipate this organisation to change as Libeo grows to answer new challenges, and learns from our wins and failures. Learning by doing is a core value at Libeo and it has been a key enabler of our strong growth.
To keep on growing and meet our ambition, we’re actively recruiting (more than 80 open positions), have a look at libeo.io/jobs to join the adventure.
If you’re interested in discussing data topics with us, drop us an email at data@libeo.io and make sure to follow Tales of Libeo where we will be sharing more insights.
