A company tries expansion. It goes badly. They conclude it doesn't work here.
Follow that backwards and there's nothing to object to. Somebody was handed a number under a deadline. The only playbook available was sales pressure, so they used it, on customers picked by the renewal calendar because no other list existed. It didn't work. The conclusion followed from the evidence.
Except the conclusion then goes to work. It removes the reason to instrument readiness, to hold anything back, to put a name on the number. And with the instruments gone, the next renewal that lands higher on a price increase gets logged as expansion, which looks enough like growth to confirm what everybody already decided.
The experiment never runs again, because the result of the broken one closed the file.
It isn't one story
I've now written five of these without noticing they were the same thing.
Compensation. Reps on real recurring commission still front-load the deal, because they've watched delivery and don't believe the customer reaches month fourteen. They're right. And the front-loading, the discount, the thrown-in module, the give-it-three-months, shortens the lifetime they predicted. The forecast and its cause were the same act.
Language. Nobody funds the system that would tell a CSM when a customer is ready, so a capable person compensates from memory and attention. The company calls that great instincts and admires it. Now there's nothing to fix, because a missing system has become a personal quality, and nobody writes a budget line for a personal quality.
Policy. Save the customer at all costs, then stop the free work, then avoid refunds, then avoid disputes. Each reversal is a correction issued by someone reasonable. And because each one feels like a decision being made, nobody notices that no standing rule was ever written. The activity of deciding replaces the decision.
There's a fifth that works differently, and it's the hardest to catch.
Strategy. A company decides what it is. We're a SaaS business, not an agency. That label then quietly sets policy: services capped at some share of revenue, an attach rate nobody exceeds, a multiple somebody once quoted. And it removes the reason to ever test whether the other thing would work, because a SaaS business doesn't do services, so nobody runs the numbers on doing services.
The difference is that the other four are responses to something going wrong. This one isn't a response to anything. It's a premise, adopted early, that makes its own examination unnecessary. There's no incident to trace it back to and no moment where somebody decided, which is why it can sit there for years.
Five departments. Five unrelated problems. One shape.
The shape
Something goes wrong. Somebody responds sensibly. The response resolves the immediate discomfort, and in resolving it, removes the thing that would have shown anyone the underlying problem was still there.
So the problem persists, and it now persists quietly, because the alarm attached to it has been answered.
Two things make this hard to catch.
The first is that every individual step survives scrutiny. There's no bad actor, no lazy quarter, no decision you'd reverse if you saw it in isolation. Ask anyone in the chain to defend their move and they can, completely. The loop is only visible from outside, and by the time you're outside it you're usually explaining a number.
The second is worse. The closing move almost always looks like progress. A conclusion drawn from evidence. A commission plan aligned to retention. A compliment in a review. A policy corrected under new information. Each of those is the kind of thing you'd put in a board update, and each one is the moment the door shut.
One question
There's a test that works on all four, and it takes about a minute.
What would have to happen for us to find out this was wrong?
Then look at whether that thing can still happen.
If expansion doesn't work here, what result would change your mind, and do you still measure anything that could produce it? If the comp plan is aligned, what would tell you the reps don't believe it, and is anyone reading deal structure? If your CSMs have great instincts, what would reveal that they're compensating, and would anybody escalate it? If the policy is right this time, what would show it isn't, and who is allowed to say so?
When the honest answer is that nothing would surface it, you aren't looking at a solved problem. You're looking at a sealed one.
How they open
You can't argue your way out of these, and people try. Arguing is the one thing the loop is built to absorb, because the loop already consumed the evidence you'd argue from.
What opens them is a measurement the loop doesn't control.
Instrument readiness before you run the expansion motion, and now the motion has a result that isn't the sales team's mood. Read deal structure as a forecast, and the reps' private view of delivery becomes visible without anyone having to admit to it. Translate every compliment into the request it's standing in for, and the gap reappears as a line item. Write the standing rule down, and the next reversal has to be argued against something instead of just replacing it.
None of those are clever. They're all the same move: put a number outside the loop, before you need it.
That's most of what the machinery is for. Readiness milestones, an expansion inventory, an owner with a number attached, terms that hold. They look like process and they read like overhead, and what they actually are is the part of the system that can still tell you that you're wrong.
A company without any of that isn't undisciplined. It's just unable to receive bad news, which is a much more comfortable condition and considerably more expensive.
Lincoln Murphy formally named and popularized Customer Success starting in 2010 and has spent 15 years connecting it to expansion revenue and commercial outcomes. Read The Premise.