- One-off customization per se is bad. Agreed. This is SaaS, not a services business.
- Being paid a lot to build something important already on your roadmap can be magical. It helps you see the future. It is a gift.
So here’s my rule:
- If BigCo offers to pay you a huge sum for a standard contract that also includes a requirement to ship an important feature you don’t yet have (usually > any other customer is currently paying you, or close),
- And they are OK without any true exclusivity on it (sometimes a short period is OK),
- And it’s already on your Next 24 Months Roadmap,
- then build it.
- Because — there almost certainly be another. Another customer that wants this exact, important feature too. And will pay a lot to have it in the feature set.
Being paid to build something that could change the game (because other, big customers will want it, too) is a gift. If one big company wants it AND you already had been thinking about building it, if it’s really already on the roadmap, at least somewhere … then, it’s not an anomaly. It’s the future.
If I hadn’t done this in both of my startups … I would have failed. Twice.
As CEO, you’ll often have to drive this decision, though. It won’t always be clear to everyone, the tradeoffs. Usually the team will not have consensus here. The sales team may want 10-12 of these features a year. You can probably only afford to do 1-2 that radically alter the timing of the roadmap. And even a great product team may have trouble intuiting when to make these bets, and when not to. And customer success sometimes is caught in a whirlwind of customer complaints and feedback. You as CEO have to visit these customers to know. To hear why they need the important feature. Don’t build it if you haven’t met the customer in person. Get on a jet.
And finally — you have to get paid. This rule does not apply if you are not paid. It does not even apply to paid pilots. It has to be a 1+ year contract.
Do not build any feature without getting paid for it. Best case, it means the prospect doesn’t really value it. More likely — they will never use it. Trust me.
More on the topic (and other related ones) here from SaaStr Annual: