Every enterprise we've worked with measures revenue, cost, headcount, NPS, churn, and a hundred other things. Almost none of them measure the one number that quietly governs all the others: the time between a meaningful signal arriving in the organization and a meaningful action leaving it.
We call it decision-to-action latency. It is the hidden tax on every enterprise.
Where the latency hides
It rarely shows up as a single delay. It accumulates across handoffs that no one owns:
- A signal lands in a system. It waits to be ingested.
- It is ingested. It waits to be modelled.
- It is modelled. It waits to be reviewed.
- It is reviewed. It waits to be discussed.
- It is discussed. It waits to be approved.
- It is approved. It waits to be executed.
Each step looks defensible in isolation. Compounded, they turn a real-time signal into a quarterly retrospective.
The market does not wait for your approval chain. It moves while you are scheduling the meeting.
Why dashboards make it worse
Dashboards were supposed to compress this loop. In practice, they often extend it. A leader sees a number, schedules a session to interpret it, asks for a deeper cut, waits for the cut, schedules a second session, and finally decides — by which point the underlying reality has moved on.
The dashboard did not produce a decision. It produced a request for more analysis. That is the opposite of what the moment demanded.
How to make latency visible
Three practices, in order:
- Name the decision domains. Not departments. The ten to twenty recurring decisions that actually move the business — pricing changes, inventory rebalances, hiring approvals, incident responses, capital allocation reviews.
- Instrument the loop end-to-end. For each domain, log the timestamp of the originating signal, every intermediate handoff, and the final action. You will be uncomfortable with what you see.
- Set a target latency per domain. Some decisions warrant deliberation. Others are bleeding value every hour they sit. Treat the two differently and budget the loop accordingly.
What collapses when the loop closes
When latency is measured and managed, the organization stops needing most of its dashboards. They are replaced by decision systems — agents that reason from a shared semantic ground, propose actions with provenance, and route to a human only where judgment, taste, or accountability genuinely require it.
The work of leadership shifts. Less interpretation. More design. Less reviewing what happened. More shaping what gets to happen automatically next time.
That is the real ROI of the operating model we are building. Not faster reports. Faster reality.