Who should be your first data hire and when should you hire them?

Before we dive in, a bit about the authors and why our perspective may be relevant:

• Marc is CEO and co-founder of Fabi.ai, and in his day to day role has spoken with hundreds of data leaders across all company sizes and industries. This gives Marc a unique vantage point on what generally does and doesn’t work for data teams. Marc has also led product teams at Clari and Assembled pre-first data hire and has experienced first-hand and assisted in the set up of the data team
• Aditya is the data leader at Shogun, and has led data teams at companies such as Poshmark and StrongDM. He comes with a wide variety of experiences and a deep experience with the modern data stack.

So… you’re a founder, maybe a CTO, or perhaps an engineering or product leader, and you’re starting to ask yourself if you should hire your first full-time data “person”. This is fantastic. If you’re asking yourself this question, it means you’ve likely made it past the chaos of the first few years as a startup and you’re starting to think about scale and operational efficiency.

But the question of who you should hire for this role can be complicated. Do you hire a data scientist who can perform some magic on your data and uncover insights no one thought about? Or perhaps a data engineer who can actually clean up the data so that you can even try and start making sense of it?

In this post, we’re going to first touch on what signs you should be looking for that indicate that you might be ready for your first “data hire”, and then once you’re ready to proceed, what experience that individual should come with and what you can realistically expect of them.

‍

When is the right time for your first data hire?

What should happen prior to your first hire

Long before you’ve hired your first data employee, you’ve likely already been doing some form of data analytics, you just haven’t had a person doing it full time. Generally speaking, there are two stages an early stage company goes through before they think about “analytics” as a full-time role:

1. Basic product & marketing analytics: On the product front, you’re using a product analytics solution that stores events, such as Posthog, Amplitude, Mixpanel or Pendo. This gives you some basic user metrics to help you understand how customers are using your product. In a separate thread, you’ve set up some web analytics (likely in GA4) to start tracking website traffic and attribution, and if you’re following more of a B2B sales motion, you likely have a CRM (Hubspot, Attio or Clarify) that provides basic reports. At this stage, you should be embracing canned reports as much as possible since you’re mostly trying to get directional insights from your data. You likely don’t have too many users or such a large sales pipeline that you don’t already have a good feel for what’s going on. This is also the stage where you can, and should, embrace spreadsheets as much as possible. Your business is evolving so quickly that you should expect what you care about to continually change.
2. SQL queries = BI: At this stage you want to start digging into more custom reports that are unique to your business. This is when data stops fitting nicely into canned reporting solutions. You may have a certain type of record in your customer database that’s powering your product that you want to better understand. For example, going back to our fictional Superdope company that sells widgets, you may want to see how many customers have more than X widgets in their cart but haven’t checked out. The best way to do this is likely just to write a SQL query against your production data. If you’re just doing one-off queries and are comfortable in a SQL IDE, that will probably get the job done, or if you want to start building dashboards or board reports, you might adopt a light-weight BI solution.

In other words, before you hire your first data person, you should already have some data (it doesn’t need to be all unified and cleaned in a data warehouse), and you should already have a sense of the types of questions you’re trying to answer with data.

Signs that you’re ready to hire your first full time data employee

Deciding to hire a full time data person isn’t a light decision. WIth your typical senior data scientist coming in at roughly $240k in total comp and the tools they will likely ask for easily running you $10k-$20k a year, you should expect a real return on investment.

Note that your first data hire likely should not be a data scientist, we touch on this more below, but this salary should give you a good gauge of what to expect in terms of compensation.

The types of questions that you should be asking yourself that likely indicate that you’re ready are some flavors of the following:

1. How can we pull all our data together to get a big picture of the business and tell a cohesive story to the board and investors?
2. What should we be working on to better serve our customers and accelerate growth?
3. Should we be doing more A/B testing and quick experimentation to figure out what might help us hit our growth plans?

You may have noticed something about these questions: for the most part, these questions are not just the output and end of the journey. They’re actually the input into decisions that the business then has to make. So a very large part of being ready for your first data hire, is also being ready to support them both financially and operationally. We touch on this below.

Prerequisites before you make the hire

To ensure fertile ground for your first data hire, you need three things:

1. Data
2. Willingness to support them financially
3. Willingness to execute on their insights

Going back to what we discussed above about what should happen before you hire your first data employee, you need data. The data doesn’t have to be pristine, well defined and centralized. That’s part of their job. But it does need to exist and have some kernel of usefulness. If you’re not sure if you have this, circle back around with your sales, marketing, product and engineering teams to see what you have today.

The second point is critical. As we touched on, even though a data scientist wouldn’t be your first hire, using their salaries as a measuring stick, you can easily expect to have to pay them $240k. The real cost of an employee is approximately 115% of their salary, and you need to be able and willing to purchase some of the tools they need to do their job. If we do some basic math:

• Salary: $240k
• Benefits and other costs: $36k
• Data warehouse: $10k
• ETL tooling: $2k
• Lightweight BI solution: $3k (a more heavy-duty BI solution typically starts around $10k)

You may already have a data warehouse and some of the other tools already, so adjust as you see fit, but the general idea being that there are very material additional costs and that needs to be part of the plan. 

There are a lot of great “free”, open source alternatives for some of the tools as well, but if that’s the plan, it needs to be factored into the hiring profile and does require more technical expertise which does come at a cost. “Free” in this case doesn’t literally mean free, there are a lot of hidden costs to self-managing data platforms.

Finally, you need to be ready to support them following their insights. An insight is only useful if it influences a decision. If your roadmap and plan is set in stone for the next 6-12 months, it may not be that valuable to hire a data person. If you think some of the data may help you make better decisions in the next few months and you’re open to experimentation or adjusting plans, you’ll be setting them up for success.

You’ll notice here that your first data hire should be providing insights that feed into your corporate strategy. For this very reason, you likely don’t want to hire a junior individual who doesn’t have executive exposure. More on this in the next section.

‍

What background should your first data hire have?

Technical expertise and experience for the first hire

As we saw above, your first data hire should be able to work with data at a very technical level but also be a strategic advisor to your leadership team. That sounds like a unicorn and may feel a bit daunting, but fret not, they’re out there! And with the rise of powerful AI tools built specifically for data analytics and reporting, the profile of this individual has changed a bit.

Starting with technical expertise. You may or may not already be extracting data and pulling it into a centralized data warehouse, but either way, this individual will be owning that process going forward. Being able to do this requires advanced SQL and Python at the very least. In addition to this, they should have experience with data modeling (dbt or Coalesce), version control (Github), and data warehousing (example: PostgreSQL, Snowflake, Redshift, MotherDuck, BigQuery…)

This type of technical expertise is particularly valuable in the age of AI. AI is only as good as the underlying data, and with AI transforming the way reporting is done, specific skills at the reporting level tend to matter less. That said, if you already have a BI solution that you’ve spent a lot of money on and started building out (we don’t recommend doing this before you hire your first full-time data employee), then this person should ideally have experience with that specific tool. But again, if you had to choose, you should lean more heavily on experience deeper in the stack since everything else is built on top of that.

On the non-technical front, you should be looking for someone who you would invite into executive discussions. This is a person who has been an individual contributor, but also grown a team and ideally worked at a startup. If you have to pick one: focus on finding someone who has worked at a startup that grew its data function and had a strong mentor at the very least. During the interview process, you should consider asking candidates about some of their past projects which should ideally include some of the following:

• Using event data to understand a customer’s journey
• Building a unified customer model
• Leveraging historical data to trace customer or user changes over time

For all the reasons above, we highly advise against trying to save a few dollars by hiring a relatively inexperienced individual. Someone with this profile, without proper guidance will struggle to draw useful insights and will likely cause you to rack up tech debt that will only cost you more down the road.

At this stage, we also feel that it’s important to call out that this person likely isn’t your finance person. Although you may be lucky enough to find someone who is deeply technical and can also help with financial planning and forecasting, these skillsets tend to be very distinct. Data can be very forgiving, but not when it comes to the financial health of the company, so best to leave that to someone who is experienced in that specific area.

A note on titles: Titles in the data space are notoriously confusing and interchangeable. You may see someone with a “Senior Data Scientist” title whose experience is mostly building out BI while you may find someone with a “BI Lead” title who is actually quite knowledgeable on the machine learning and statistics front. But to attract the best candidates, this role should be titled “Head/Director of Data”.

30/60/90 plan and ROI

If you’re considering hiring your first full time data employee, you’re likely wondering what the ramp up period should look like and when you can expect positive ROI.

Let’s start with the more difficult question: How do you define ROI for this role and when should you expect positive ROI? Unfortunately, ROI on data projects remains nearly impossible to measure despite much debate (try searching “ROI on data teams”). The reality is that you simply can’t expect to put in $X and get out $Y when it comes to reporting. You’ll know your investment is worthwhile when it feels like you couldn’t operate without the insights that the team delivers. Put another way: following ramp up time, if your executive team is not turning towards the data team to get answers to help guide them, you may not be getting your value.

Now for a ramp up plan for this first data hire:

• First 30 days - The first month is about two things: Understanding the business and understanding the state of the data stack. The very first thing this hire should be doing is gaining a deep understanding of how your company makes money and how that’s measured today. Every other question branches out from there, and a failure to understand the fundamentals of the business model will cause this individual to pursue vanity metrics. They should meet with every executive, starting with the CEO, and acting heads of marketing and sales.
• First 60 days - At the end of 60 days, your data hire should understand the top priorities for all key functions of the business and how they can support them. Within those priorities, they should have identified the top 3 priorities across the entire business and have developed some form of reporting for those areas. At this point, you should reasonably expect this person to have developed any new data models (tables) they need for reporting and delivered useful and accurate reports in the executive team’s tool of choice - spreadsheets or lightweight BI are totally fine and perhaps even encouraged. A first data hire that comes in and immediately dives into BI and spends the first few months just getting set up is likely not going to be a good long term fit.
• First 90 days - By the end of the first 3 months, this person should have a deep understanding of what makes the business run, they should have already delivered reports for the top areas of focus and they should have started putting in place an operational heartbeat. They should be driving meetings or updates (new technologies make this incredibly easy) and be actively participating in strategic discussions. They should also have a full roadmap that encompasses the needs of all key executives.

Biggest mistakes when hiring your first data person

Although we’ve already touched on this above, it’s worth calling out specifically the biggest mistakes we tend to see.

The first is not setting clear expectations. Either with yourself or with the hire. If all you have is a general sense that “we need more reporting”, it might be a rough ride for all parties. If you’re able to clearly articulate why you’re hiring for this role and what you hope to see in the first 30/60/90 days, all parties will be much better off.

The second mistake is not providing the proper support. We touched on this in detail above, so we won’t linger on this too much here, but not building in the monetary or resource budget to provide them the tools they need and the time to react to their plans or suggestions will render this first hire completely ineffective. Make sure you’re ready to provide more than just their salary.

The third mistake is hiring someone who is very specialized in a certain department. For example you may see someone with extensive experience as both an IC and a leader in “Marketing analytics”, but unfortunately, if this is the individual’s only experience, they may have a really difficult time working across functions and will likely default back to what they’re most comfortable with: marketing. This is great for your marketing department, but they’re surely not the only ones in need of data support.

Finally, the most common mistake is thinking this person just needs to write some SQL, and hiring someone with very limited experience. Without proper mentorship, they will likely get stuck in the technology, miss the forest from the trees, and ultimately end up costing you much more in both strategic direction and tech debt than you might expect.

‍

Your first hire should be deeply technical with experience providing strategic guidance to executives

If you’ve outgrown your pre-canned product and marketing analytics platforms, and you’re starting to wonder if there are insights in your data that could help drive the company strategy, you’re ready to start thinking about building out a data team. But before you do so, make sure you’re ready to provide the financial support they will need along with a willingness to experiment and adjust the plans based on their feedback.

You’ll want an individual with deep SQL and Python experience, who has ideally led and grown data teams in previous startups. This person should be someone that you expect to turn towards for strategic advice, but they’re going to be on their own initially so they should be self-sufficient. Make sure you don’t hire too junior or too much of a specialist.

It may feel daunting to find someone who fits the bill, but they’re out there, and once you find the right person, they will be a force multiplier on your business. Data can truly deliver competitive insights, and conversely, it can be a money pit, so it’s worth taking the time and waiting for the right moment to find the right person. Ultimately, you’ll need to consider the unique nature of your business and team and weigh your priorities to determine which traits are the most important for this role.