Customer fit: what it is, how to measure it, and how to spot the iceberg customer

Customer fit is how well what you sell matches what a specific customer needs, and how well that customer matches what your company can keep delivering. Both halves count. A customer who loves the product but needs a custom build, weekly calls, and a discount to stay is a bad-fit customer with a good opinion of you.

When the fit is real, you can see it in the numbers. Good-fit customers understand the product on the first call, buy in weeks instead of quarters, open fewer support tickets, and refer the next customer. April Dunford's positioning process in Obviously Awesome starts with exactly these people, and her description of them is the best definition of customer fit I've read: "They understood your product quickly and bought from you quickly. They became raving fans, referred you to other companies and acted as a reference for you."

What customer fit is, and how it differs from product-market fit

Product-market fit is a property of the market: enough people want the thing that the company can grow. Customer fit is a property of each account. You can have product-market fit and still sign a customer who will never work out, and most companies do, because revenue looks the same on the invoice whether the customer fits or not.

A good-fit customer has the problem you solve, in the form you solve it, at a price that leaves you a margin. A bad-fit customer sees some value at first and then finds the edges of the product. They ask for customization, they take more support than the account is worth, and they churn anyway, usually with a review that says the product "didn't do what we needed." Tom Randle, the CEO of Geckoboard, put it plainly: "Churn is totally governed by the kind of customers that are coming in and whether they're a good fit." No onboarding program fixes a customer who should never have been sold.

The cost of getting this wrong is well documented. Harvard Business Review puts the cost of acquiring a new customer at five to 25 times the cost of keeping an existing one, and Bain's research found that a 5% improvement in retention raises profits by 25% to 95%. Both numbers only hold for customers worth keeping. Retaining a bad-fit account at any cost is how you lose money on a customer who pays you.

Beware of the iceberg customer

Revenue is the number that hides bad fit best. I learned this at a marketing agency I worked at earlier in my career. Our largest account, the one we called our most valuable customer in every internal meeting, turned out to be the account that consumed the most hours, forced the most exceptions to how we worked, and left us so dependent on one contract that losing it would have meant layoffs. On paper it was the best customer we had. In practice it was running the company.

I call these iceberg customers. The revenue is the part above the water. Five things sit underneath it:

1. The support load. Iceberg customers need a share of support and custom work far out of proportion to their share of revenue. The cost of serving them can pass what they pay.
2. The roadmap. A customer that size gets listened to, and the roadmap drifts toward features only they will use. Every sprint spent on them is a sprint not spent on the next 20 customers.
3. The margin. Extra support, custom features, and the discount they negotiated at renewal add up, and the account that looks like your most profitable is often near break-even.
4. The dependency. When one customer is 30% of revenue, they set the terms, and you say yes to things you'd say no to anyone else.
5. The exit. They leave eventually, because the product was never built for them, and they take the 30% with them in one quarter.

The way to avoid signing the next one is to measure customer fit with something other than the size of the contract.

‍

Five customer fit metrics

Each of these is useful on its own, and none of them is enough on its own. The pattern across all five is what tells you who your good-fit customers are.

1. Revenue. What the account pays, as MRR or ARR. A threshold like "at least $5,000 MRR" is a fine starting filter. Revenue on its own is the iceberg problem, so it's the first metric and never the last one.
2. Cost to serve. The hours, tickets, and custom work the account consumes. A threshold like "fewer than 10 support tickets a month" turns a vague sense that an account is heavy into a number you can compare across customers. Revenue minus cost to serve is the number that matters.
3. Satisfaction. CSAT or NPS, tracked per account rather than as a company average. A CSAT above 70% means the product is doing what they bought it for. A high-revenue account with a low score is an iceberg customer that hasn't left yet.
4. Engagement. How often they log in and which features they use. The important part is which features. Heavy use of the core product means the fit is real. Heavy use of one workaround feature you built for a single customer means the fit is with the workaround, and that's a different product.
5. Growth potential. Whether the account can add seats, teams, or products as the customer grows. A good-fit customer at $5,000 MRR with three more departments who share the same problem is worth more than a bad-fit customer at $15,000 who has already bought everything they'll ever buy.

Put all five in one sheet, one row per customer, and sort. The accounts that score well on four or five are your ideal customer profile, and the description of them (industry, size, the problem they had, who signed) is what marketing and sales should be aiming at. The accounts that score well on revenue and badly on everything else are the ones to watch.

The three mistakes that make the numbers lie

The first mistake is weighting revenue over everything, which is the iceberg again. The fix is to never look at revenue without cost to serve next to it.

The second is skipping cost to serve because it's hard to measure. Support tickets, hours logged by CS, and the number of custom requests in the last quarter are all good enough proxies. Any of them beats not measuring it.

The third is treating all engagement as good. A customer who logs in daily to use the product as intended is a great sign. A customer who logs in daily to export data into a spreadsheet because the reporting doesn't do what they need is a support ticket that hasn't been filed yet. Look at what they're using as well as how often.

How positioning brings in good-fit customers

The cheapest way to improve customer fit is to stop attracting the bad-fit customers in the first place, and that's a positioning job. When the homepage says exactly what the product is, who it's for, and what problem it fixes, the right customers recognise themselves and the wrong ones leave before they ever book a demo. Fewer of the wrong demos means shorter sales cycles, fewer custom requests, and a support queue full of questions the product can answer.

Messaging that names the specific pain does the filtering for you. It's the difference between "a project management tool for teams" and "project tracking for construction firms that are losing money on change orders." The second one gets fewer leads and closes more of them, and the ones it closes stay.

Once you've measured customer fit, take the description of your best five accounts and check it against your homepage. If someone who looks exactly like them wouldn't know within 10 seconds that you're for them, that's the next thing to fix.