Run history — opening a finished run and checking its work
Where a completed run keeps its plan, how to check the result against it, and the two scopes of change available afterwards.
The project, and which files were selected
A project page is a header, a toolbar and the file list. Selection is what a run reads, so the first question about any run is which files were ticked when it started — Deselect all and Select all are in the toolbar for exactly that reason. A surprising result often turns out to be a correct result over the wrong file set.
The run history drawer
History opens the project's runs, newest first, each with a status. Scroll it before opening anything: the pattern is more informative than any single row. Several runs in quick succession on one project usually means assumptions were being tuned, and only the last one counts — so check you are reading the run somebody actually sent.
Inside a run: the Proposal tab
Opening a run gives you tabs, and Proposal is where the narrative lives — strategy first, then results. Read it in that order. A result you have not read the plan for is a number with no provenance, and it is the plan that tells you what the number is a total of.
The strategy, in the agent's own words
The plan states how it grouped the files, which version it kept, what it excluded, and which rating method it chose for each dataset with a reason attached. This is the part to read closely, because every later figure inherits these decisions. If a grouping looks wrong here, nothing downstream will correct it.
Checking the result against the plan
Then apply the arithmetic. Total divided by lanes should equal the stated average; the freight mix should sum to the lane count exactly; the lanes rated should match the dataset the plan named. Each check takes seconds and each one has caught a real problem: the first finds a total computed over a different population, the second finds double-counted lanes, the third finds a grouping that quietly took in more than it said.
Read the anomaly line against the totals
Runs report anomalies as well as prices, and the two can contradict each other — a summary can state that every lane priced successfully to a large positive total and, a sentence later, flag every row as zero-or-negative. Those cannot both hold, and the way to tell which is wrong is the cost range: a set whose cheapest lane is a healthy positive number is not entirely zero-or-negative. Check the anomaly rule before sharing a run that says something like this, because it is exactly the sentence a customer will quote back at you.
The deliverable, and the sheets you did not build
The customer workbook carries the rated data alongside the shared reference tables — fuel surcharge schedules, rules and accessorials — at the same row counts the Reference page publishes. Confirming that from inside a delivered workbook is how you know the tariff is genuinely inherited rather than described somewhere and hoped for.
Steering it from here
Two scopes of change. Edit a sheet to fix one thing in place — the file viewer guide covers what that touches and what it does not. Or reiterate the whole project to run it again under different instructions. Prefer the narrow one: a re-run replaces every number in the output, including the ones that were already right and already quoted.
Ready to build your own?
Start free and ship your first app in minutes — or have us set it up with you.
More guides in Rating & Proposals
More resources