The deal is nearly there. They like the product, the price works, and there is one thing they need that you do not have. Build it and you win. Everyone in the room can see the path. So the question feels like "can we build this?", and the answer is usually yes. That is the wrong question. The real one is what you are agreeing to own for the next five years.
This is one of the few decisions a founder makes inside a sales cycle that outlives the sales cycle by years. It deserves more than the twenty minutes it usually gets.
The build is the deposit, not the price
Founders cost the sprint. Two weeks of engineering, maybe three. That number is real and it is also the smallest part of what you are signing up for.
What follows is the actual cost: the tests that now have to pass, the documentation, the support people who have to understand it, the onboarding that has to explain it, and every future change to the product that has to work around it. A one-customer feature does not sit quietly in a corner. It takes a seat at every roadmap meeting from now on.
Most of what gets built is never used
This is worth sitting with, because it is not intuition, it is measured.
Pendo analysed feature usage across 615 of its customer accounts and found that 80 percent of features in the average software product are rarely or never used. Break that down and it gets sharper: 56 percent were never used at all, 24 percent rarely. Just 12 percent of features generated 80 percent of daily usage. Pendo put a number on the waste too, estimating that a software company with $50 million in revenue might spend $8.4 million building features customers rarely or never touch.
Now remember that every one of those features had a business case, a product manager and a roadmap slot. A one-customer custom build starts from a worse position than that: an addressable audience of one.
Should you build it? Four questions
Run the ask through these before you answer. If you cannot get through them in the meeting, that is your answer for the meeting.
Would you build this if nobody was asking? If it is already on the roadmap and this deal just moves it forward, that is not a custom build, that is good timing. Say yes quickly.
Can you name the second customer? Not "the market" or "others have mentioned it". A name. If the second customer is hypothetical, you are building a product for one account and calling it a feature.
Are they paying for it separately? A customer who will fund the build is telling you it is genuinely valuable. A customer who wants it included is telling you it is worth something less than the discount they would otherwise have asked for.
Who owns it in two years? The person who championed it will move on. The engineer who built it will move on. If nobody in the room can say who maintains it after both of those things happen, nobody does, and it rots in place.
The answers that are not yes or no
Most of the time the right response is neither. Three that work.
Configure rather than build. A surprising amount of what gets asked for is a settings change, a template, a different default or a report. It looks like less of a win. It is the same outcome without the permanent resident.
Do it manually first. If the customer needs the output, deliver the output. A person running it by hand every month tells you within a quarter whether the demand is real, and costs nothing to stop.
Say yes, with a price and a date. Scoped, paid, contracted custom work is a legitimate business. It only goes wrong when it is quietly folded into the licence fee and then called roadmap.
What a free build teaches your buyer
The same thing an unearned discount teaches them, which is that the number was never fixed. Concede a custom build inside a sales cycle and you have shown the buyer your roadmap is negotiable under deadline pressure. They will remember at renewal, and their procurement team will hear about it before that.
That is not an argument for saying no. It is an argument for saying yes on terms, so the yes reads as a decision rather than a flinch.
When to build it anyway
Sometimes the ask is the market telling you something and the right move is to listen fast. If three prospects have raised the same gap in a quarter, if it unlocks a segment you already wanted, or if the absence of it is the reason you keep losing to one competitor, build it - and build it as a product, for everyone, not as a private favour for one logo.
The distinction is not whether you build. It is whether you are choosing or being pushed.
So take the last custom thing you agreed to build. Would you have put it on the roadmap if the deal had not been on the table?
Is your roadmap being set by your last deal?
In about 6 minutes, the free Growth Scorecard shows where your sales are leaking and what to fix first.
Get your free scorecardSources. Feature usage: Pendo analysed feature usage across 615 Pendo subscriptions, limited to customers who had used Pendo for more than a year, over a three-month window, and found 80 percent of features in the average software product are rarely or never used - 56 percent never used at all and 24 percent rarely, with 12 percent of features generating 80 percent of average daily usage. The same report estimates a software company with $50 million in revenue might spend $8.4 million developing features customers rarely or never use, derived from Gartner's 2018 public cloud revenue forecast and average R&D spend across 54 companies on the Bessemer Venture Partners Emerging Cloud Index (Suja Thomas, "2019 Feature Adoption Report," Pendo, 5 February 2019). Two caveats worth stating: the data is from 2019, and Pendo sells feature-adoption tooling, so it has a commercial interest in the finding. I have used it because the methodology and sample are published in full, which is more than can be said for the widely-quoted "45 percent of features are never used" figure - that traces to a 2002 conference talk about four internal applications, and I have left it out for that reason. The four-question test, the three alternatives to building, and the parallel with discounting come from my own experience advising NZ tech founders, not from a study.