Data for startups
What does a data function actually do?
I like to describe a startup’s data maturity / journey using “Maslow’s hierarchy of data needs”:
Layer 1Reconciling multiple sources and providing key metrics
Bring your data from different systems (product, marketing, finance, support and so on) into a single data platform and define the key business rules. This becomes the reconciled “source of truth” and provides a view on critical metrics around unit economics and growth. It is decoupled from the production systems, so you can move fast without impacting customers.
Layer 2Scaling decision making
The data platform is democratised to the broader business with an intuitive serving layer and strong governance. With AI especially, coaching the company on critical thinking and analysis is the bulk of the job here.
Layer 3Enriching production systems
Reconciled data is sent back to the product and business tools to enable enrichments, automations and customer-facing data products. Examples include training and running customer-facing ML models, real-time operational dashboards, and a more detailed view of customers in the CRM.
I’ve done this journey end-to-end a few times now. For anyone starting on this journey, these are my core principles that helped deliver high-performing data functions:
- Focus on influencing commercial outcomes. Delivering data isn’t the same as delivering a result.
- Think of internal data users as customers and the data platform as a product.
- Always do the right thing — data compliance and security are P0.
If you want to learn more about how to do this data thing right, I’d be happy to chat.