Blocker latency
Last updated: 19 October 2026
Blocker latency is how long work stays stuck before anyone who can help knows about it. It is the gap between “we cannot move” and “somebody has been told”, and it is where a lot of a team’s days quietly go.
The definition we use
For one blocker, latency = reported time − started time. For a workspace we report the median over its recent blockers. The median, not the average, because a single blocker that sat for a month would otherwise hide how the usual case behaves.
- Started time is when the worker says it began (“stuck since Tuesday”). If they do not say, it is the time they reported it.
- Reported time is the timestamp WhatsApp put on their message, never the time our server got round to processing it. A message sent during a network outage and delivered a day later still counts from when it was sent.
- Because an unstated start defaults to the report time, the figure can only under-state latency. It is a floor.
A worked example
Brian is waiting on a customs certificate. It stopped his work on Tuesday at 10:00. He mentions it on Thursday at 09:00. That blocker’s latency is 47 hours. If four other blockers in the workspace had latencies of 2, 5, 30 and 96 hours, the median is 30 hours: the middle value once they are ordered.
What it does not measure
It does not measure effort, speed or output, and Chaos does not rank or score people. A blocker is a fact about a piece of work: it cannot move until something outside its owner’s control changes. That something can be a colleague, an outside office, a missing document, or the manager, and a blocker waiting on a manager is shown as plainly as any other.
Why a team would want it lower
People rarely announce a blocker. They say “still waiting”, “I’ll see”, or nothing at all, and bad news travels slowly when reporting it feels risky. Making it safe to report, asking one specific private question when a signal appears, and routing the blocker to whoever can clear it all shorten the gap. Blocker latency is the number that tells you whether they are working.
Questions
What is blocker latency?
Blocker latency is the time between a piece of work becoming stuck and someone who can help knowing about it. Chaos measures it as the median of (time reported − time the blocker started) across a workspace’s blockers.
How is blocker latency different from cycle time or lead time?
Cycle time and lead time measure how long work takes from start to finish. Blocker latency measures only the gap inside that time where work is stuck and nobody has been told. A team can have a short cycle time and still lose days to blockers that surface late.
Is blocker latency a measure of how hard people work?
No. It describes how quickly a blocker becomes known, not how productive a person is. Chaos does not score individual pace or productivity, and a blocker waiting on a manager is shown as plainly as one waiting on anyone else.
How do you know when a blocker started?
From the worker’s own words, such as “stuck since Tuesday”. When they do not say, the start is taken to be the time they reported it, so the measured latency is a lower bound, never an overstatement.
More on how we study this: the Chaos blocker research programme.