Because for most small and mid-sized companies, a full-time analyst is either too much, too soon, or the wrong tool entirely — and hiring one onto a broken foundation just gives you a very expensive person maintaining spreadsheets.
What a data analyst actually does — and doesn't
An analyst answers questions with data: builds reports, investigates trends, slices the numbers to explain what happened and why. What they're usually not hired to do — and often can't do well on their own — is the plumbing underneath: getting your systems to talk to each other, defining each metric once, and building the model the reports sit on. Drop a talented analyst onto conflicting definitions and disconnected systems, and they'll spend most of their week cleaning and reconciling data by hand — the exact grunt work you were trying to escape. Then they get frustrated and leave, taking the knowledge with them.
The three things people really mean by "we need an analyst"
- "Our numbers don't agree and I don't trust them." That's a definitions-and-integration problem, not a headcount problem. It's the same root cause behind two reports that show different numbers.
- "We spend forever building reports by hand." That's an automation problem — build it once and it runs itself. Worth estimating what those hours are actually costing you first.
- "I have real questions the data can't answer yet." That's the one genuine analyst job — but it only pays off once the foundation underneath it exists.
Hiring an analyst to fix untrustworthy data is like hiring a driver for a car that won't start. The problem isn't who's at the wheel.
The real cost of the full-time hire
A full-time analyst isn't just a salary. It's benefits, the management time to keep them pointed at the right work, months of ramp-up, and the quiet risk that when they leave, every definition and workaround leaves in their head. For a company that needs trustworthy reporting but not daily deep analysis, that's a heavy, permanent, six-figure commitment to solve what is usually a one-time, foundational problem.
What most companies need first
Before — or instead of — a hire, get the foundation built once: definitions agreed in plain language, systems connected so data flows automatically, and one shared model every report draws from. Do that and one of two things becomes true. Either your existing team can now self-serve reports they trust, and you never needed the hire — or you genuinely do have deep analytical questions, and now an analyst (full-time or fractional) is worth it, because they'll spend their time analyzing instead of janitoring. A dashboard has the same trap: it only helps once the foundation beneath it is sound.
When you genuinely do need one
To be fair to the analysts: there's a real moment for the role. When you have a constant stream of novel questions, large or genuinely complex data, or analysis is core to your product, you need dedicated analytical horsepower — and a good analyst earns their salary many times over. For most small and mid-sized companies, that moment simply hasn't arrived yet. A solid foundation plus occasional expert help covers the ground for a fraction of the cost and risk.
Building that foundation — the connections, the definitions, the single model that makes reporting trustworthy without a permanent hire — is the work I do. If someone's told you to hire an analyst and you're not sure that's the real fix, tell me what's going wrong with your reporting.