It was late in a conference room that had run out of coffee, and I had asked the founder across the table for something that sounded easy. List every decision from last quarter that would still matter in a year. Not the meetings. Not the fires. The calls that would still be shaping the company twelve months out.
He thought about it longer than he expected to. Then he wrote them down, one at a time, and stopped at nine.
Nine decisions. A whole quarter of a growing company, and the number of choices that actually mattered fit on a single page with room to spare. He looked at the list a little surprised, the way people are when a thing they carry every day turns out to be smaller and heavier than they pictured.
Then I asked the harder question. For each of the nine, what was in front of you when you made the call. What did you actually know.
We went down the list. And the pattern that came out of that page is the reason I run this audit before anyone in the room is allowed to say the word “tool.” The nine decisions that would shape the next year of the company had been made on memory, a competitor’s press release, and one spreadsheet that was eight months old.
I want to be honest about something before I walk you through it, because the audit only works if you go into it with your guard down. The first time I did this, I did it on myself. I was advising three companies and running the strategy for a fourth, and I was certain I was operating on good information. I was not. Two of my highest-stakes calls that quarter were running on nothing but the confident feeling that I had seen this before. That feeling wearing the costume of experience. The audit does not care how many years you have. It just shows you the page.
Here is how to run it. It takes an afternoon.
Step one: list the decisions, not the meetings. Your calendar is not your decision calendar. Most leaders have their meetings mapped in obsessive detail and have never once written down the actual calls they make that set direction. Pricing. Hiring plans. Where engineering points for the next two quarters. Which market you chase and which you politely ignore. What you stop doing. Write them in your own hand, in one place, and do not pad the list. When the founder did this, he kept trying to add operational decisions to make it feel more substantial. Whether to renew a vendor. Which candidate for a mid-level role. I made him cross those off. Those matter, but they do not still matter in a year. The list is supposed to be short and heavy. That is the point of it.
Step two: score each decision on leverage. For every item, ask how much of the company’s next year rides on getting it right. Be honest about the spread, because the spread is brutal and it is supposed to be. A few of these decisions are worth more than all the others combined. On the founder’s list, two of the nine carried more weight than the remaining seven together. One was a pricing model change that would either open the mid-market or quietly wall them out of it. The other was a bet on which product line got the engineering team for the next year. Everything else was real, but it was not that. You are looking for the two or three that, if you get them wrong, no amount of execution downstream saves you.
Step three: score each decision on the quality of the information behind it. Not whether you felt confident. Confidence is not the measurement. The measurement is whether the call was actually informed by current data, or by instinct dressed up as analysis. Go line by line and grade the inputs the way an outsider would. The pricing decision, the one that could wall them out of the mid-market, had been made off a single conversation with one loud customer and a number the founder half-remembered from a panel. That was the whole evidentiary basis for a call worth the better part of a year.
Now lay the two columns next to each other and look.
The pattern is almost always the same, and it is the entire reason I bother with the exercise. The highest-leverage decisions are the ones running on the thinnest information. The person making the eight-million-dollar call is working from less data than the analyst three floors down who has AI helping him with an eighty-thousand-dollar one. The stakes and the rigor are inverted. The more a decision matters, the more likely it is being made on a hunch, because the big calls are the ones nobody has time to build a foundation under.
I know the objection, because I have made it myself and I have heard every sharp founder in the room make it. Great companies are built on instinct and speed. Data slows you down. If you tried to fully inform all nine of these decisions you would still be gathering inputs while a faster competitor took the market. The founder’s gut is the thing that got the company this far, and now I want to put it on a leash. There is truth in that, and it is worth sitting with rather than swatting away. Speed is real. Judgment is real. Some of those nine decisions are genuine judgment calls that no model on earth improves, and treating them like data problems would be its own kind of failure.
But the audit is not an argument for informing everything. It is the opposite. It is the tool that tells you where to stop trusting your gut and where to keep trusting it. You do not need AI on all nine. You need it on the two or three where the leverage is high and the information is thin, and you need to leave the rest alone. That is where a strategic briefing earns its cost in a single quarter, not because it is clever, but because it is pointed at the one place the stakes and the rigor are furthest apart.
The founder did not buy a platform. He picked the three decisions that scared him most, the ones with the widest gap between what was riding on them and what was under them, and he built the data foundation beneath just those. The other six he left exactly as they were. The pricing call that had been running on one loud customer got a real read on the mid-market before he pulled the trigger. It changed the number. Not dramatically, but enough that on that decision alone the work paid for itself.
Here is what I did not expect the first time I watched someone go through this. The lasting change was not the three foundations he built. It was that he started catching the pattern in advance. A quarter later he told me he now flags a decision the moment he feels it is high-leverage and thinly informed, before he makes it, not after. The audit stopped being something I ran on him and became something he ran on himself, in real time. That is the version that compounds.
I have also watched what happens without it. Another founder I know, sharper than most, skipped this step entirely and went straight to buying the tooling. He put AI against a dozen dashboards that were already well-fed and well-understood, the eighty-thousand-dollar decisions, because those were the easy ones to instrument. His three biggest calls that year stayed exactly where they had always been. On his gut. Six months later two of them had gone sideways, and the tooling had produced beautiful reports about decisions that were never going to hurt him.
You can run this audit yourself this week. The scoring is not the hard part. The hard part is writing the decisions down and seeing, in your own handwriting, how little was under the ones that mattered most.
At Kennesys, the fractional CIO and CTO work starts here, at the strategic layer, mapping where your leverage actually is before a dollar goes toward tools.