From Customer-Led Development: How to Build What Your Customers Want When AI Can Build Anything — the operating manual for building what customers want when AI can build anything. See every chapter.
Sales makes promises. Product breaks them. The customer leaves. Then everyone blames everyone else.
Sales says product never builds what deals need. Product says sales promises things nobody agreed to. Customer Success asks why they keep inheriting angry accounts. The customer only says one thing: "You told me this would work."
Then they churn.
This is the promise-to-delivery gap. It's where a lot of churn begins, and almost none of it has to happen.
Six months, one sentence
Here's how it plays out, month by month.
Month 1. Sales is working a big deal: Acme Corp, $150K ARR. The champion loves the product but needs custom reporting. The AE says, "We can do that." The deal closes.
Month 2. The promise gets logged somewhere. Maybe a Salesforce note, maybe a Slack thread that sinks within a day. Maybe only in the AE's head.
Month 3. Product reviews the backlog for roadmap planning. The reporting request isn't there. Or it's there with no context, one line among hundreds.
Month 4. Acme's champion asks their CSM when custom reporting is coming. The CSM asks product. Product says it's on the roadmap but not prioritized. The CSM passes that along.
Month 5. The champion was told four months ago this would happen. They start looking at competitors.
Month 6. Acme churns. The exit survey says "missing features." Nobody connects it to a sentence spoken on a sales call five months earlier. Sales blames product. Product blames sales. The $150K is gone.
Then it happens three more times that year. Different customers, same pattern. At year-end review, $600K has left through four accounts, every dollar traceable to a promise that fell through the cracks.
It sounds like a horror story. For most B2B companies, it's a Tuesday.
The gap is bigger than any one deal
An IDC estimate, via LXA Hub, puts the cost of sales and marketing misalignment at 10% or more of revenue a year for B2B companies (see also Kixie). The gap between what sales promises and what product delivers is the same disease in a different organ.
At $10M ARR, that's a million dollars walking out the door. Not because the product is bad or the market moved. Because the left hand didn't know what the right hand promised.
And it compounds. Churned customers talk to prospects. Lost revenue means less to invest in product. Sales, still chasing quota, keeps promising. Product keeps building what it planned rather than what was promised. The gap widens.
This isn't a people problem
The easy story is that sales overpromises and product doesn't listen. Swap a few people and it's fixed.
It isn't. The problem is that there's no system, so three predictable things happen.
Promises are invisible
The AE closes, writes some notes, moves on. Nobody reviews those notes systematically. Product can't see them. CS doesn't check them during onboarding. The promise lives in one head or one buried record, which means, organizationally, it doesn't exist.
Nobody is paid to keep them
- Sales earns commission on closed deals. More promises, more closes.
- Product is rewarded for shipping its plan, not someone else's commitment.
- CS is measured on renewals. They inherit the problem; they can't prevent it.
- Nobody is compensated on promises kept.
Promises are an externality. Someone else's problem, right up until the customer churns and it becomes everyone's.
Even documented promises get distorted
Every handoff adds noise:
| Who | What they have |
|---|---|
| Customer says | "I need custom reports with our branding." |
| Sales logs | "Custom reporting requested." |
| Product reads | "Reporting feature." |
| Engineering builds | Export to CSV. |
| Customer receives | Something that doesn't solve their problem. |
By the time the signal reaches the builder, it's unrecognizable. You can ship something and still break the promise.
Redefine what counts as a promise
Most teams think of promises as the risky stuff an AE blurts out on a call. That's too narrow. Customers make commitments based on everything you told them: roadmap hints, pricing terms, the demo, the contract. If they built a decision on it, you're on the hook for it.
Anything a customer relied on when they signed is a promise, whether an AE said it on a call or it's written into your terms.
Once you see it that way, "missing features" in an exit survey stops being a product-market-fit signal and starts being an accounting error. Something was owed. Nobody tracked the debt.
The fix is a system, not more communication
The instinct is to fix this with meetings and Slack channels. More talking. But the problem isn't that people don't talk. It's that promises have no home, no owner, and no finish line.
A real fix gives every promise three things:
- A home. One visible place where every commitment lives, with the customer's name and who made it.
- A neutral owner. Not sales, not product, not CS. Someone who sits between all three.
- A finish line the customer controls. Shipping isn't done. Done is the customer saying it solved their problem.
Rule of thumb: A promise closes when the customer confirms it works, not when you ship it.
Want to know how big your gap is? Pull last quarter's closed-won deals, read the notes, and list every promise you find. Then count how many actually shipped. The difference is a number most companies have never looked at, and it's usually uncomfortable.
Customers remember what you promised them long after your team has moved on. The only question is whether you remember too.
The full promise system, including the tracking board, the weekly review, the pre-deal check, and how to get sales and product to actually adopt it, is in Customer-Led Development.
If this changed how you read your next churn report, subscribe to the Context Limit newsletter for more ideas on closing the gap between what you say and what you ship.