A startup spends so much energy on acquiring and retaining its customers that voluntarily severing its relationship with any of them feels like self-inflicted harm. Yet, when necessary, acting decisively to “fire” a customer can mean the difference between survival and failure for a young company.

By far the most important time to act is when customer relationships entered into in good faith reveal themselves to be strategically misaligned.

Here’s an example I recall vividly from my own company’s journey.

In the early years, there were several long droughts in which we struggled to make the next sale. Guidewire’s* first product was a claims-management system for property-casualty insurers. This functional domain turned out to be astonishingly large and complex, so our first sales took a long time while our product was still nascent. The early rejections were frustrating enough, but after winning the first few customers, we faced another drought because prospects wanted to see our first cohort succeed before committing to buy. And after that, there was another drought as our end market wanted to see that our platform could serve as a full system-of-record. And after that there were a few more droughts still to come!

Stumbling upon a customer oasis

During one of these parched passages, we thought we found an oasis: an adjacent set of prospects who were interested in our solution. Unlike the first few customers who fit our “ideal customer profile,” these prospects were not insurance companies as such; they were large enterprises that managed their own claims (“self-insureds”) and “third-party administrators” (TPAs), service companies that handled claims on behalf of others.

Within a few months of adapting our positioning and pursuing prospects, we won two new customers—a large grocery chain and a regional TPA. We were ecstatic to have broadened our market and to have added to our (short) roster of customers. Best of all, these customers seemed not to need many of the complex features that full insurance companies did, and the legacy systems they were using looked primitive. We reported progress to the board with pride and started recruiting product management and sales units to focus on the new customer segments.

Within six months, however, we realized we had a problem—or rather, a growing set of problems. First, the new customers found many of our core product concepts misfit to their businesses. Some of the misalignment was subtle, like the naming of a field.

But some of the problems were more fundamental. For example, self-insured companies don’t have insurance policies as such, so parts of the data model and UI confused the users and complicated the accounting logic deep in the system. Also, unlike the claims departments of insurers, TPAs serve multiple clients and need an additional layer of entities in the data model to keep track of financials by client—but these would complicate implementation for the insurers who comprised our core market.

Before long, we were fielding requests in multiple directions for serious product changes. Design sessions started weighing the merits of branching the code base, which we knew in our guts would lead to a crippling tax on a dev team already grappling with a huge scope.

Meanwhile, our messaging got more complicated: instead of “P&C claim system,” our “One Thing” became “a claim system for P&C insurers, self-insureds and third-party administrators.” We had to explain that our ambition to build the other core systems that insurers — but not self-insureds and TPAs — would take priority over other system needs for those segments.

The more complicated messaging and product strategy had a trickle-down effect through the rest of the organization. We needed separate versions of our product collateral and additional navigation branches on our website. Our small pre-sales team had to develop different versions of the demo, which doubled the work with each release. The same applied to our already inadequate documentation and training materials. The legal team had to deal with a new set of terms specific to the new segments, which in turn required new revenue-recognition analysis by the finance team. With great effort and with the best intentions, we were digging ourselves into a deepening ditch.

It was an emotional trial even to acknowledge the truth of the situation. We had worked so hard to win these customers, we had announced them with fanfare, and we had cashed the checks! Surely there was a way forward! Couldn’t we just work harder to somehow serve the superset of these paying customers??

Yes, and yes. But we would have dis-served them all in the end. We would have delayed our product by months, our implementations would have been at risk and we would have imperiled our chances of winning the vital next set of customers waiting for us.

And so we apologized to those misaligned customers, canceled their projects, refunded the hundreds of thousands of dollars they had paid us, reset the scoreboard and foreswore adventures outside of our core market.

Focusing on a strategically coherent path

Easy enough? Well, large corporations “fire” their customers all the time when they divest business units or shut down product initiatives with low traction. We all know the famous decision Steve Jobs made to end 70% of Apple’s product lines when he returned to the company he founded—and the miracles that unfolded afterwards.

But it is different and even harder for a startup. Striving like crazy to make something out of nothing, a young company is trying to (1) stake its claim on a right to exist in the first place, and (2) set forth its identity to the world. Early customers can introduce a painful tension between these imperatives. Each customer contributes positively to the metrics that support the first goal —customer count, installs/usage, transactions, ARR, etc. These metrics substantiate traction, and they unlock the market attention and investor attention a startup craves.

But more important than being on a (seemingly) successful path is being on a strategically coherent path. Without strategic coherence, that traction will inevitably sputter out—and with more painful consequences the longer the reckoning has been deferred.

Another way to put it: A startup’s customer set as a whole defines its identity and strategy more than any marketing statement. That identity and strategy must be coherent and precisely tuned to win against the odds always stacked up against a young company. Thus, wherever the imperfect fit may be—business model, channel conflict, cost-to-serve or functional requirements —misalignment between where the company wants to go and what the customer wants now costs a startup dearly. Those early wins can be Pyrrhic victories in disguise.

My company’s decision at the time hurt like hell, but I am glad the decision was as clear as it was.  A more ambiguous tension might have led us to dissipate more precious energy at that critical stage in our development.

Strategic focus is the startup’s most important asset, and recognizing the necessity of firing a customer is the sharpest test of that focus.  Don’t let mere metrics distract from passing that test.