A group asks each site for monthly figures. Ten spreadsheets arrive in eight formats, three of them late.
One site counts an enquiry when it arrives and another when somebody responds. One includes walk ins and one does not. Two changed how they record something in March and did not mention it.
The group total is assembled anyway and presented at a meeting. Nobody can act on it, because nobody can tell whether a difference between two sites is performance or definition.
Start by defining the measures, not collecting them
Four or five measures, defined precisely, agreed with the sites. Definition is the whole exercise, and collection without it produces confident nonsense.
Define an enquiry exactly
When it counts, from which channels, and what happens with duplicates and walk ins. This single definition causes more inconsistency than anything else.
Measure response speed centrally
Time to first response is the most comparable and most actionable measure across sites, and it is the one most likely to explain a difference in conversion.
Measure conversion on a common basis
From the same starting point, over the same period. Sites measuring from different points produce numbers that look comparable and are not.
Capture the data at the source rather than by return
A monthly request for figures produces late, reconstructed and inconsistent numbers. Data captured automatically as things happen is the only reliable route.
Allow for legitimate differences
Catchment, size, competition and what each site can do. A league table ignoring these produces resentment and drives behaviour that games the measure.
Compare a site against itself over time
Frequently more useful than comparing sites against each other, and considerably less contentious.
Report the few things that change decisions
Demand, response speed, conversion and satisfaction. Reports with thirty measures are read by nobody and acted on by nobody.
Show sites their own numbers regularly
Managers act on what they can see. A monthly group report nobody at site level receives changes nothing on the ground.
Investigate outliers rather than ranking them
The site at the bottom usually has a reason, and finding it is worth more than the ranking. Frequently the reason is a definition rather than performance.
Record what enquiries were about, not just how many
The mix by reason is what tells a group where to fix something, which is the audit method behind keeping every location saying the same thing.
Include the out of hours share
What arrived when each site was closed, by site. This is usually the largest and least examined number in the whole report.
Agree who owns the report
One person responsible for the definitions, the collection and what the numbers mean, rather than a document assembled by whoever is free at month end. Reporting owned by everybody is reporting nobody will defend when a figure is challenged.
Watch for the variation you chose
Hours and prices that differ deliberately will show in the numbers, and the report should account for it, which is the ground in opening hours that differ by site.
Capture it automatically at every site
An assistant handling enquiries at every location records the same fields, in the same way, at the same moment, which produces comparable data as a byproduct rather than as a monthly request. The consistency work sits in keeping answers consistent across franchise locations and the multi site handling in ai receptionist multiple locations. See pricing, run the numbers on the chatbot ROI calculator, and the sector view sits under b2b.