The deal slipped on Tuesday. We did something about it the next Wednesday.
By Meet Patel · 2026-08-31 · updated 2026-08-31
Summary
Management latency is the time between something changing in a business and the person able to act knowing about it. One example ran to eight days, in a company that pages an engineer for 400ms. Measure it with four timestamps: changed, known, decided, acted.
The deal slipped on a Tuesday.
It surfaced in the forecast call on Friday. The response was agreed the following Monday and shipped on the Wednesday after that.
Eight days.
The same company alerted on four hundred milliseconds of API latency and would page a person at three in the morning to fix it.
Nobody found that strange, including me, for quite a long time.
One of these is instrumented and one is not
Engineering has spent thirty years learning to measure how long things take. There is a number, it is watched, and when it degrades somebody is woken up. The whole discipline of running systems assumes latency is a first-class property you can see.
The management layer of the same company has none of that. There is no number for how long it takes us to find out something changed. Not a bad number — no number. It has never been measured, so it cannot get worse, which is a comfortable position and a false one.
Four timestamps
You need exactly four, per incident:
- Changed — when the thing actually happened in the business.
- Known — when a person capable of acting on it knew.
- Decided — when a decision was made.
- Acted — when something shipped.
Most companies can produce none of the four. Take ten recent material changes, reconstruct them from calendars and message threads, and take the median. It is an afternoon of work and it is the most useful afternoon I have spent on our own operations.
The total is the headline. The interesting part is which gap is widest, because each one has a different cause and a different fix.
Changed to known is a noticing problem. Nothing in the company was responsible for looking, so it waited for a human to wander past.
Known to decided is a cadence problem. Somebody knew on Tuesday and the meeting where it could be raised was Friday.
Decided to acted is an ownership problem. The call was made and it was not clear whose week it landed in.
In our eight days, the biggest single gap was the first one — three days before it reached anyone who could act. I had assumed it would be the last one. It was not close.
Cadence is a floor you cannot argue with
Here is the part I find genuinely hard to get around.
If the only forum where a class of decision gets made meets weekly, then latency for that class cannot go below about seven days. Not on a good week. Not with a motivated team. The cadence is a hard floor, and no amount of dashboard freshness lowers it.
Which means the warehouse being current to the minute is beside the point if the only moment anyone looks is Friday. The pipeline is fresh; the organisation is not. Those are different properties and only one of them gets monitored.
Measure it once
I am not going to tell you what a good number is. I do not know, the benchmarks do not exist, and anyone quoting one is guessing.
What I will say is that almost everyone I have asked to do this exercise came back surprised, and always in the same direction. The number is bigger than they thought, and the gap that dominates it is not the one they expected.
Measure yours. Then measure whether it fell. That is the whole method, and it beats any benchmark somebody sells you.
The definition, and how it differs from data freshness and reporting frequency, is written up here: management latency.
Frequently asked questions
How do you measure management latency?
Take ten recent material changes and record four timestamps for each: when it changed, when a person able to act knew, when the decision was made, and when the action shipped. The median gap from changed to acted is your management latency, and the widest interval tells you whether the problem is noticing, cadence or ownership.
What is a good management latency?
There is no benchmark worth quoting and anyone offering one is guessing. The useful comparison is internal: measure it once, then measure whether it fell. Most companies find their first number is in days for things they assumed took hours.
Written by Meet Patel — founder of Company 8, building Dan (usedan.com). Dubai, UAE.