The renewal notice arrived on a Tuesday. The vendor that ran this company’s core operations wanted 40 percent more, a mid-five-figure jump on what was already their biggest software line item. The CEO called me angry, and his plan was simple: leave.
So we priced the exit, and the exit was worse. Their data sat in the vendor’s proprietary format with no contractual export rights. The migration estimate came back at nine months. About forty people would need retraining, and every integration the ops team had built over three years would have to be rebuilt against a new system. All in, leaving penciled out at several times the price increase, spread over two years of disruption. He paid. He is still paying.
Here’s what I want you to notice about that story. It isn’t a pricing problem, and it isn’t really a vendor problem. The vendor understood that switching would be expensive, and its pricing reflected that leverage. The actual event happened three years earlier, at signing, when the company evaluated the platform on features, price, and a good demo, and nobody asked what leaving would cost. That question was never answered because it was never asked. A decision you don’t make on purpose still gets made. It gets made by default, and in this case the default was written into someone else’s contract.
I think of these as absent decisions: questions the company never formally answered. Every company carries some, and they don’t show up on any dashboard, because a decision that was never made leaves no record. You usually meet them later, at a worse time, at a worse price.
What we did about the vendor
The lock-in was real, so year one was about stopping the trap from resetting rather than escaping it.
Three concrete moves. First, at the next contract amendment we negotiated data export rights, in a documented format, delivered quarterly. That cost some goodwill and a little money and was worth both. Second, the ops team spent a couple of weeks interviewing their own people and writing down every automation and workaround they had built inside the vendor’s platform over the years. Most of that knowledge existed only in a few employees’ heads, which meant the real lock-in was worse than the contract said. Third, they kept one credible alternative warm: a yearly demo, a small pilot in one region, enough that the claim “we could move” stayed true instead of theoretical.
At the following renewal the increase came in at single digits, and the export clause survived. I won’t pretend that was all leverage. Some of it was probably the vendor’s own pricing cycle. But the CEO walked into that conversation able to mean it when he said he had options, and that had not been true before.
Where else absent decisions show up
Once you know the shape, you see it in places that look nothing like a contract.
A COO showed me her reporting stack this spring: about forty dashboards, well built, updated nightly. Her team still couldn’t make decisions at speed, and her instinct was to commission better reporting. When we went dashboard by dashboard and asked “what decision does this drive, and who owns that decision,” most of them had no answer. They were requests for visibility standing in for choices nobody had committed to. The missing thing was never data. Most of those dashboards went away. The survivors each got a one-line label at the top naming the decision they served and the threshold that would trigger it, which sounds bureaucratic and took an afternoon. A month later her staff meetings were noticeably shorter, mostly because arguments about what the numbers meant had been replaced by the pre-agreed answer to what they’d do when a number moved.
One CEO I know ran the blunt version of this test: he turned off a dreaded weekly report for a month to see who noticed. Two people asked. That report cost someone most of a day, every week, and it had been feeding nothing for years.
And I watched a good team nearly automate a fourteen-step invoice approval process this year. When we walked it step by step and asked why each one existed, nine of the fourteen turned out to be scar tissue from problems that no longer existed. A vendor issue resolved two years ago. A check that duplicated what the system already did. An approval for a manager who had left. Automating that process would have made all nine permanent, because automation doesn’t just speed a process up, it commits you to its current shape. The decision they were skipping was “what should this process even be.” They made it, fourteen steps became five, and then the automation was easy.
Where I’ll push back on my own thesis
Not every unmade decision is a failure, and not every dependency is a trap.
Sometimes lock-in is a fair trade. If a system is the only serious option in its niche, or the switching cost buys you integration depth you actually use, then dependence can be the right call. The difference is whether the dependency was understood, priced, and accepted before the contract was signed. The facts can be identical either way. What changes everything is when you found out.
And some decisions are honestly better deferred. Early on, “ship one payment provider and revisit at scale” is often correct. The failure mode is deferral with no owner and no date, which is how a sensible postponement drifts into a permanent absence.
Something to try this week
Pick the report, tool, or process your team complains about most. In your next operations meeting, ask two questions about it. What decision is this supposed to serve? Who has agreed to make that decision?
If the room answers cleanly, you’re fine, move on. If the room goes quiet, you’ve found one of your absent decisions. Then do the only thing that actually converts it: give it a single owner and a date, out loud, before the meeting ends. Not a committee and not a follow-up deck. A name and a day, put somewhere the team already reviews, like the top of the weekly ops agenda. You’ll feel the room resist, and the resistance is the point. It’s the same discomfort that kept the decision unmade in the first place.
One more note from doing this exercise with clients: if the owner you name turns out to be you for more than two items in a row, that’s a finding in itself, and it’s worth asking which of those you’re actually going to make and which you should hand to someone with fewer excuses available.
A question for you
What’s the decision in your business that keeps getting postponed by activity? The hire you keep researching instead of making. The vendor you keep evaluating instead of confronting.
Hit reply and tell me about it. I read every response, and the patterns across them become future issues, anonymized always.
If this was useful, forward it to a leader who has been “gathering more data” on the same decision for six months.
See you next week.
Phillip