Completeness is how many of the sheet's topics a tech actually answered - skipping half a sheet is invisible on the board. Photos is the share of photo-required topics that were answered with a picture attached; a finding with no picture doesn't sell. Duration is shop-hours only (Mon-Thu 7-5, Fri 7-3), so an inspection left open overnight reads as the minutes actually worked, not fourteen hours. Median is used rather than average because one run left open over a lunch break drags a mean badly.

Shop-wide, submitted runs only, topics seen at least 3 times. A topic skipped by everyone is usually a sheet problem worth fixing in the Builder, not a people problem.

How long vehicles actually sit in each column, in shop hours (Mon-Thu 7-5, Fri 7-3, holidays from the Hub). The timeline is rebuilt from the board's own step-change history, so a car dropped Thursday evening doesn't read as an overnight bottleneck. Median is what normal looks like; p90 is where the pain is. The bottleneck is the step burning the most total shop time - a quick step on every car can cost more than a rare slow one.

Time a quote is sitting unapproved is the advisor's to chase. Attribution is the advisor currently on the ticket.

What a tech recommended, and what happened to it. Sold and declined come from the closed invoice in job history, not from the board - the board's auth flag is a step rule that authorizes everything on a ticket, so it can't tell you what a client actually approved.

Close rate is sold / (sold + declined) - only recommendations that reached the client. Dismissed and pending are shown but kept out of that denominator, or a desk that dismisses everything would flatter the tech's number. Dismissed means killed on the board before it ever reached an estimate; removed means it made the estimate but came off before close. Invoices sync nightly, so the last day or two sits in pending.

Who converts together. Pairings with fewer than 3 recommendations that reached a client are hidden - that's noise, not a pattern.

Send rate is the share of ROs that got a report at all - auto-send only covers one step, so every other send is a person choosing to. Open rate is of those sent. ARO opened vs not opened is the number that says whether any of this is worth doing.

Total seconds a client spent on the report, in 30-second buckets.