Hiring a data scientist to fix broken billing data is an expensive mistake. Here is the order that actually works.
Most founders hire backwards. They raise a round, look at their disorganized dashboards, and decide they need a Head of Data or an expert in predictive modeling. What they usually need is someone to explain why their Stripe export shows $412,000 in monthly revenue while their accounting ledger shows $398,000.
Last year we worked with a 45-person logistics software company that spent five months and $160,000 hunting for a machine learning engineer. When the engineer finally arrived, he spent his first six weeks asking where the cleaned warehouse tables were. There were none. He was handed a raw Postgres read replica and a mess of Hubspot webhooks. He left four months later.
Start with the plumbing and the definitions
Your first hire should not be building recommendation engines. They need to sit between your operational software and your leadership team, untangling definitions and building clean, reliable tables.
- First hire: A pragmatic analytics engineer or full-stack analyst who writes strict SQL and understands basic relational modeling.
- Second hire: A domain-specific analyst embedded directly within product or sales operations.
- Third hire: Specialized roles, whether that means data engineering for scale or data science for complex inference.
The generalist test
Look for low-ego pragmatism. Your first analytics hire will spend half their time as a data custodian and the other half as an internal journalist. They must be comfortable auditing missing foreign keys in your CRM on Tuesday and explaining margin decay to the executive team on Thursday.
If an applicant asks during the interview what your advanced experimentation stack looks like, they are probably the wrong fit for a company under 100 people. You need someone who is content using a warehouse, dbt, and a simple reporting tool without over-engineering the architecture.
Delay the manager
Do not hire a Director or VP of Data to manage zero people. We routinely see companies hire an expensive data leader who immediately purchases a six-figure enterprise tool stack, drafts complex governance documents, and then demands a team of four to actually execute the work.
Hire the builder first. Let them establish your source of truth, document the core metrics, and demonstrate that data can influence a roadmap. When that single person is visibly underwater with valid, high-impact requests from every department, then—and only then—do you hire someone to manage the function.
Want this kind of clarity on your own reports?
We rebuild executive packs and dashboards for a living. Send us what you've got. We'll tell you, honestly, what we'd change.