About 19 minutes. Measured on production on 11 October 2026: the 81 comparables runs that finished and were charged in the previous 30 days took a median of 18.7 minutes each.
That is the time from the moment the run starts to the moment its result is delivered.
The measurement
| Measure | Comparables runs |
|---|---|
| Finished and charged runs in the 30 days to 11 October 2026 | 81 |
| Median time, start to delivery | 18.7 minutes |
| The middle half of runs | 8.1 to 37.4 minutes |
| One run in ten took longer than | 57.9 minutes |
One query against production, run on 11 October 2026 at 10:12 UTC. It counts every comparables run started in the 30 days before it that finished and was charged with credits, ours and our clients' alike. The time runs from the start of the run to the delivery of its result; where a run carries no delivery time, the time of its charge is used (4 of the 81). Runs that were stopped or failed are not in it.
Why the spread is wide
The middle half of runs spans about half an hour, and one run in ten takes close to an hour. We have not measured what drives the difference, so we name no cause. Plan for the long end of the table, not the short one.
What comes back
A board of comparables. Each comparable carries its address, rent and size, with a source on every figure, so you can open the record behind a number instead of trusting it.
What it does not mean
- It is not a promise for your property. The table is what happened, not a service level.
- A fast board is not a finished opinion. Choosing which comparables count is a judgment, and it stays with you.
- The output is a draft for a licensed professional on your team to review and sign.
How to see it on your own property
Ask for a demo and name one property you know well: request a demo