The Average Is Hiding Your Worst Location
David Selva · · 3 min read
Reporting & Dashboards, Cleaning Companies, Multi-Location
Roll three branches into one dashboard and the numbers look fine. On average, they almost always are.
If one of your branches stopped returning calls tomorrow, how long would it take you to find out?
For most multi-location owners I ask, the honest answer is about a month — when revenue comes in soft and somebody finally goes looking for a reason.
An average is a device for hiding your worst performer
Three locations. One answers 95% of its calls, one answers 80%, one answers 55%. The dashboard reports 77% and nobody panics, because 77% doesn't look like an emergency.
There is no such thing as a 77% customer experience. Every caller reached exactly one branch, and a third of them reached the bad one. The average describes a company that doesn't exist.
This gets worse as you add locations, which is precisely the opposite of what owners expect better reporting to do for them.
What has to be split by location
Every number a local manager can personally influence:
- Calls answered and missed, by hour of day
- Speed to first response on new inquiries
- Booking rate from inquiry to scheduled job
- Reschedules and cancellations
- Review count and rating over the last 90 days
- Revenue per crew, not just the branch total
What can stay company-wide
Marketing spend by channel, brand-level reputation, overall payroll — anything no single manager owns or can move on her own.
The test is simple. If a location manager could change the number by doing her job differently, she needs to see her own version of it every week.
Comparison does something a target can't
Here's the part that surprised me the first few times I set this up.
Telling a manager she is at 55% against a target of 90% produces a defensive conversation about how her market is different, her staff is newer, her phones are busier.
Showing her that the branch fifteen minutes away is at 95%, with the same script, the same pricing and the same weather, produces a completely different conversation — usually one where she is the one asking what they do differently.
Rankings are uncomfortable and they work, as long as you are comparing things that are genuinely comparable.
The failure mode: numbers that stop being true
Once branches know they're measured, some of them begin managing the measurement instead of the work.
Jobs get marked complete before they are. Inquiries that were never going to book get deleted rather than marked lost. The dashboard improves while the business doesn't.
The defense isn't more metrics. It is choosing numbers that are hard to fake — call recordings behind the answer rate, review counts that originate outside your system, revenue that has to reconcile against what was actually invoiced.
Look at your worst location, not your average
Open whatever reporting you have and split a single number by location for the last thirty days. Call answer rate is the easiest one to start with.
Then ignore the company figure completely and look only at the lowest branch. That number is what a third of your customers experienced this month, and it's the only one on the page worth doing anything about.