We talked with a COO recently who ranked his locations every month. Collections per visit, best to worst, one slide. The bottom location got the attention: the extra visits, the process reviews, the coaching. It had been the bottom location for over a year, and in that year it had been fixed four times.
The month the ranking flipped
Then one month the bottom location climbed to third, and nobody could explain it. No initiative had landed. No staff had changed. The room congratulated itself anyway and moved on, but the COO couldn’t let it go, because a number that improves for no reason is a number that can fall for no reason, and he suspected he’d been reading tea leaves for a year.
He had. Two insurers had shifted patient volume between neighborhoods, which moved the blend, which moved the ranking. The slide had never been measuring the locations. It had been measuring what each location’s patients happened to carry in their wallets.
And wallets move for reasons no location controls. Mix is set by the neighborhoods a location serves and the plans local employers buy, which means one big employer switching insurance carriers can shift a location’s blend overnight, no staffing change, no process change, nothing anyone inside the building did. A measure that swings on someone else’s benefits decision was never a performance measure. It just looked like one because it produced a tidy order once a month.
What the gap was made of
The spread from top to bottom was about $18 per visit. We decomposed it, insurer by insurer, and nearly all of it was payer mix, the blend of insurance plans each location happens to serve. The bottom location sat in a neighborhood heavy with one lower-paying plan. Take any single insurer and compare it across addresses, and the bottom location collected as well as anyone in the group, sometimes better. The ranking had been grading geography and calling it performance.
What rankings punish
A ranking doesn’t just misread. It teaches. The bottom location’s manager had spent a year defending her team from a number they couldn’t move, and the fourth fix was received exactly the way you’d expect. Meanwhile the top location coasted, because its rank answered every question before anyone asked one. And that’s where the real problem was hiding. Its biggest insurer had drifted off its own payment pace for five straight weeks, and nobody had looked, because you don’t audit the location the slide calls best. Under a self-baseline, that drift flags in week one or two. Behind a good rank, it had five weeks and was working on more.
How do you compare performance across locations?
Compare each insurer at each location against its own history, and compare locations only insurer by insurer. Blended rankings mostly measure payer mix, which no location controls.
Locations against themselves
The instrument that works is the same one that works everywhere in this business: each thing against its own history. Every insurer at every location, compared to that insurer’s own last twelve months at that address. The monthly ranking got retired. In its place, a short list of breaks, each with a location, an insurer, a start date, and the dollars off pattern. Some months the list was empty. The empty months were the point.
The find, the fix, what stays
Find: an $18-per-visit gap that was mostly mix, four fixes applied to a location that was never broken, and a five-week drift at the top location that rank had insulated from questions.
Fix: per-insurer, per-location baselines, refreshed weekly, with attention routed to breaks instead of positions on a slide.
Stays: the ranking never came back. The bottom-location manager got her year back, and the group’s next real problem surfaced as a flagged break at the former best location, thirteen days after it started, instead of surviving another quarter behind a good rank.
The monthly meeting changed with it. With no positions to defend, location managers stopped arriving armored and started bringing things to the list themselves, an insurer acting odd, a queue getting deep, because the list treats what they bring as a break to resolve instead of evidence against them. The COO said the meetings got shorter and more honest in the same month, which he had not previously believed were compatible.
Your version
Take your bottom location. Pick its three biggest insurers and compute collections per visit for each, separately. Then compute the same three at your best location. If the per-insurer numbers are close, your ranking has been measuring neighborhoods, and somewhere near the top of your slide there’s a location nobody has questioned in a very long time. The blend hides the same way at group level, where the average eats the bad quarter.
We’ll decompose your location rankings insurer by insurer and show you which gaps are real and which are geography. Grab 30 minutes with us. Prep nothing.