People usually come to me with one of two sentences. "We have loads of data and no idea what it's telling us." Or, "Our numbers keep disagreeing and nobody trusts the dashboard." Both are fair reasons to talk to someone. Neither automatically means you need a consultant, though, so here is how I'd think about it if I were sitting on your side of the table.
When it pays off
You have a specific problem and nobody with the time. A question like "why did repeat purchases fall after the pricing change?" is a few weeks of focused work. Hiring a full-time analyst to answer it is slow and a bit absurd. A consultant can do the work, hand over the queries and leave.
You need a skill once. Designing a warehouse model, auditing your event tracking, setting up an experiment process. Most teams do each of these once or twice in a decade, which means nobody in-house has done it before. Borrowing someone who has done it many times is cheaper than learning on your own data.
You want a second opinion. Sometimes you do have an analyst, and the answer they gave doesn't feel right. Or it feels right and the decision is expensive enough that you'd like it checked. A fresh pair of eyes costs little compared with a wrong call.
You're about to buy a tool. Before you sign a two-year contract for a warehouse or a BI platform, an afternoon with someone who has no commission on the sale is worth having.
When it doesn't
Nobody will own the result. This is the failure I see most. A consultant builds a set of dashboards, everyone says thank you, and six months later the data is stale because no one on your side was responsible for it. Before you hire anyone, decide who will look after what gets built. If the answer is "nobody yet," hire that person first or alongside.
You don't know the question. A consultant can help you find it, but that is strategy work and should be bought as something short and cheap, not as a six-month build on a guess.
The disagreement is about priorities, not data. If sales and marketing disagree about who should get the budget, a better dashboard will not settle it. Someone needs to make a decision. Data can inform it but it can't make it for you.
You need someone every day. If analysis is a constant, full-time load, hire. Consultants are best at the bursty stuff.
How to judge one before you sign
Ask what they would do in the first two weeks. A good answer is specific: "I'd read your tracking plan, query the last ninety days of orders, and talk to the person who built the current report." A bad answer is a slide with five arrows.
Ask for an example of past work that you can discuss in detail. You don't need client names. You do need enough detail to tell whether it really happened and whether it resembles your problem.
Ask who will do the work. Sometimes the person you meet on the call is not the person who turns up on Monday.
Ask what they'd say no to. Someone who accepts every request is not exercising judgement. I'd rather lose a project than build something I think won't work.
And be wary of a tool-first answer. If the first thing out of someone's mouth is the name of a product, they are probably selling it, or at least comfortable with it, and neither is the same as it being right for you.
What it costs
I won't quote figures here because they depend too much on the problem. What I'd suggest is comparing the fee with two things: what the problem costs you every month it stays unsolved, and the fully loaded cost of hiring someone, including the months it takes to recruit and the months before they're productive. Often the comparison isn't close either way, and that tells you what to do.
If you can't tell which side of the line you're on, a half-hour conversation usually settles it. I'm glad to have that conversation, and if the honest answer is "not yet," I'll say so.