There's a layer underneath every expansion argument I make, and it's the one almost nobody has.
It's called behavioral engineering, and it decides whether any of the rest of it produces a number.
The Definition
Behavioral engineering is designing the conditions that make a customer's next action more likely.
Read the verb, because it's carrying the whole idea.
You can't actually manage another person. You can have the title, Customer Success Manager or Account Manager, and it describes a job rather than a relationship, because the customer doesn't report to you and never will. You have no authority over them, and every plan that quietly assumes otherwise fails on contact with a human who has other priorities.
What you can do is change the conditions around a decision, so the thing you need becomes easier, clearer, more obviously worth it, or simply the path of least resistance.
You can't actually manage another person. You can influence the conditions that make a behavior more or less likely. That's the whole shift, and it's most of the work.
And a behavior, for this purpose, is something you can observe. Not a feeling, not an intention, not a health score. Somebody did a thing or they didn't.
The Growth Stack
It sits in a specific place, and the place explains why it gets skipped.
Service delivery is the floor. You produce what you sold. Nothing above this works until it's true, which is why delivery failure precludes expansion and always will.
The expansion motion is the orchestration layer. Knowing what a customer could buy next, when it becomes useful, and putting it in their head before they arrive.
Behavioral engineering is what makes the first two produce results faster. It's how a customer actually gets from where they are to where the next thing makes sense, rather than getting there eventually, or not at all.
Most companies have some version of the first. A few have started on the second. Almost nobody has the third, which is why the second underperforms and gets blamed for it.
The Question It Answers
Every service business runs into the same wall, and it sounds like a complaint until you take it seriously.
How do you get a customer to do something they don't want to do, can't easily do, or won't get around to on their own, when you need them to do it and you have no authority over them at all?
That's not a motivational problem and it isn't solved by caring more. There is real research and real practice on it, and none of it is taught in Customer Success. So teams improvise: they send another reminder, escalate to the champion, put it on a QBR slide, and describe the customer as disengaged when none of that works.
The customer usually isn't disengaged. The action is ambiguous, or it's boring, or it requires a decision they'd rather not make yet, or it's genuinely hard and nobody acknowledged that.
What It Actually Looks Like
Four moves, none of them exotic.
Name the behavior, not the outcome. "Get more adoption" isn't a behavior. "The ops lead uploads their first list" is. If you can't observe it, you can't design for it and you can't tell whether it happened.
Find what's actually in the way. Usually it isn't motivation. It's an unclear next step, a missing permission, a dependency on somebody who hasn't agreed, or a task that's genuinely unpleasant. Each of those has a different fix and only one of them is a nudge.
Reduce what you're asking for, or absorb it. The most reliable way to make a behavior more likely is to make it smaller. And a growing share of what customers were expected to do is work a system can now do for them, which is the honest version of the agentic argument.
Expose uncertainty instead of collecting agreement. A customer who says yes when they aren't sure has handed you a problem that surfaces months later as blame. Asking which parts they have questions about costs one question and prevents the most expensive kind of failure there is.
Why This Is a Revenue Argument
Because velocity is the lever nobody prices.
A customer who reaches their first milestone in month three instead of month five doesn't just get value sooner. They reach the second milestone before the year is out, which means one more expansion event inside the same contract year, on the same product, at the same price. Nothing about what you sell changed. What changed is how long they spent stuck.
So behavioral engineering isn't a soft skill bolted onto the side of the commercial argument. It's the thing that moves the number, because the delivery engine you already run gets faster when the humans in it stop stalling, and every stall is a behavior nobody designed for.
Where I've Been Using It
Most of this work I've done privately, and a lot of it in Brazil, where I run it as a workshop. Where it has shown up in writing it's been the defensive half: the psychology of a customer deciding to leave, and short-term interventions that shift behavior enough to buy time.
The more interesting half points the other way, at getting a customer to the place where they can buy more, sooner, without anybody selling them anything.
Same discipline. The difference is which direction you aim it, and almost everyone aims it at the exit.
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.