Your first data hire
I've made the first data hire at one startup and advised several others as they did the same. Each time we've had basically the same conversation, and it looked something like what you’ll find here.
A few assumptions about your company
You're somewhere between 20 and 50 people.
Most of your data consists of whatever is native to your existing tools (a Stripe dashboard for revenue, a Facebook dashboard for ad performance, etc.). Some enterprising engineer or marketer also probably writes SQL written against your production database and pastes it into Excel or Sheets.
You probably have Google Analytics set up, and you might also track events with Mixpanel or Segment or something similar.
You probably don't have a data warehouse (Snowflake, BigQuery, etc.), a BI tool (Tableau, Looker, etc.), a data ingestion tool (Fivetran, Stitch, etc.), or a transformation tool (dbt).
What will your first data hire do?
This person is going to be working with folks across the business to better understand what's going on. That means building dashboards, doing ad hoc analysis, and communicating insights.
But they're also going to be building some of the foundations of the data warehouse. The data models they build to enable their own analysis will be around for years.
You're looking for someone whose responsibilities will straddle two realms: Analysis and Analytics Engineering.
What skills should this person have?
Business understanding. This person will be working with colleagues across the business to help them understand what is going on. They will be performing analysis on everything from marketing efficacy to retention to product usage. They should have good instincts about which insights are valuable and which are not.
Data modeling. This person needs to be able to decide how to represent the business in concepts and tables. The company will depend on these tables for years, so the costs of messing them up are high. Include an engineer in the hiring process to test for data modeling skills.
SQL. Can't model data without SQL.
Basic comfort in git. Git can be learned quickly, but it is a very effective litmus test in the interview process. If a candidate hasn't worked in git before, they probably haven't done the other kinds of work you'll need them to do. They won't be technical enough to make substantial technical choices around tools, data models, and more.
How senior should this person be?
You should hire a senior individual contributor with the potential to build and lead a team. This doesn't mean you'd want them to manage five people immediately, but you should be confident that when it's time to hire the second data person, they'll be ready to manage.
Hire someone with at least four years of experience. I almost never rely on years of experience as an input, but do not compromise here. Less than four years of experience means this person will likely struggle to navigate complex organizational dynamics, make good hiring decisions, or speak the different dialects of their stakeholders.
Who should manage this person?
Someone technical. This could be a head of engineering or CTO, but it might also be an unusually technical person on the marketing or growth side. It might be a technical founder, but they likely have bigger fish to fry. Maybe a head of finance could be the manager, but I’d put that option last on the list.
The key is that this manager needs to be able to recognize whether this first data hire is making the right foundational technical choices. The manager will be signing off on major tool purchases and helping prioritize work, and it's hard to do this without some technical background. As the team grows you can shift it to a less technical part of the org, such as under a COO or CFO, but I’d advise against doing this too early.
What should this person's title be?
Honestly, this is the least important question as long as you get the above right.
In precise terms you're hiring a Senior Analytics Engineer or a Senior Data Analyst, but in practice this person might prefer a Senior Data Scientist title, or Analytics Lead, or something else.
Any of the above is fine as long as everyone is on the same page about the work this person will be doing.
More reading
The Founder's Guide to Analytics. This post from Tristan is great. He wrote this years ago but it holds up so well. I essentially agree with all of his advice at the 50-150 employee stage.
Analytics Job Descriptions. dbt Labs, characteristically, has a really helpful nuts and bolts guide here. You'll want to add more detail on your own company and the specific problems this person will be working on, but your first hire should see a JD that is somewhere at the intersection of the "analyst" and "analytics engineer" JDs in the dbt Labs post.