Multi-Site Oversight When You Are Between the Buildings

Multi-site oversight is usually designed for a person sitting still. Reports get built for a Monday review, dashboards assume a wide screen, and the whole apparatus quietly requires that somebody be at a desk with an hour to spend. Regional managers do not work that way. They work in cars, in waiting rooms, and in the twenty minutes between one site and the next.

The result is oversight that is technically available and practically unused. The dashboard exists. Nobody opened it Tuesday because Tuesday was three buildings and a payer meeting.

What gets lost is not the reporting. It is the window where a problem is still cheap. A site whose phone coverage collapsed on Monday morning is a fixable staffing conversation on Monday afternoon and a month of lost bookings by the time it shows up in a monthly roll-up.

Pulling is a desk activity, pushing is not

Every reporting system built for a multi-site group assumes the reader goes and gets the number. That is a pull model, and pull requires attention you do not have on a Tuesday.

The part of oversight that travels is the part that arrives on its own. A short, standing message at a fixed hour with one number per site. An exception alert when a site crosses a threshold you set. A link that opens the underlying case if you want to look, and does not demand that you do.

That is not a smaller dashboard. It is a different job. The dashboard answers questions you thought to ask. The push answers the question you did not know to ask, which is the only one that matters when you are not looking.

Everything else can wait for the office, and should. Trend analysis, rate comparisons across a quarter, and rubric changes are genuinely better done sitting down.

One number per site, and it has to mean the same thing everywhere

The hard part of multi-site measurement is not collecting numbers. It is that the same label means different things at different sites, so a comparison across ten locations is comparing ten definitions.

Standardizing that is an athenaOne configuration exercise before it is a reporting exercise. Departments have to map cleanly to the physical sites you actually manage, which is not automatic in a group that grew by acquisition. Practice-level settings and department lists are the ground truth for what a site is, and they need to be reconciled once, deliberately, before any cross-site number is trustworthy.

Once that is true, one rate per site is enough for the road. Not a scorecard. One rate, in the same units, with a threshold you set. Everything richer belongs to the weekly review.

The practices that get this right pick the rate that is closest to revenue rather than the one that is easiest to collect. Calls that ended without a booking, cases opened and not closed, and requests that aged past a day are all closer to money than average handle time will ever be.

The site that changed something and did not tell you

Here is the complication that makes remote oversight worth building rather than merely nice. Sites change their own configuration, and central operations finds out late.

One practice retired its entire cosmetic and procedure appointment-type catalog overnight and folded everything into a single fifteen-minute follow-up type. That type cannot hold a forty-five minute service. Nothing broke loudly. Bookings kept happening, they were just wrong, and the schedule started absorbing visits it could not deliver.

A manager sitting in that building would have heard about it by lunch. A manager two hundred miles away learns it from a metric, if the metric is watching. Case volume against a retired type drops to zero. Case volume against one surviving type spikes. Nothing in a monthly report shows that until the month is over.

So the alert fires on the shape of the change, not on somebody remembering to send an email. And then it stops. The system does not decide what the new mapping should be. It says which site, which appointment types moved, and when, and a human decides whether the change was intentional. That handoff is the whole design. Detection travels well and judgment does not.

Automation does not remove the manager, it changes what they read

There is a persistent assumption that adding automation to a multi-site front office reduces the oversight burden by itself. The evidence from practice leaders is more modest than the pitch.

A June 2, 2026, MGMA Stat poll of 260 applicable responses found that 68% of practice leaders say their organization has not redesigned a role or adjusted staffing with the help of AI in the past year, while 26% say they have. Most groups added capability and kept the org chart, which means the manager is still the manager. What changes is what lands in front of them.

That is the honest version of the promise, and it is the useful one for an MSO. Automation handles the volume that is repetitive and complete. The exceptions, the site that behaves differently from its peers, and the judgment calls still route to a person. Your job is to make sure the routing reaches them wherever they are.

Redeploying attention is the actual win. A regional manager who is not reconstructing what happened gets to spend the drive thinking about the two sites that need a decision.

Set the thresholds centrally, let the numbers differ locally

A rural site with four providers and an urban site with twenty will never produce the same rates, and forcing them to share a threshold produces alerts that everyone learns to ignore.

The workable pattern is central rules and local levels. The definition of an unclosed case is the same across the group. The count that triggers an alert is set per site, from that site’s own recent baseline. That way a genuine deviation fires at every site and normal variation fires at none.

Access pressure is why this is worth the setup. When MGMA asked practice leaders where they would focus on patient access in 2026, no-shows were the largest single answer at 27%, with online scheduling at 24% and phone access at 22% close behind. Those three compete for the same front-desk hour, and they compete differently at each of your locations.

Review the thresholds monthly, at a desk, with the full picture in front of you. Then let them run, and let the road version be exceptions only.

Key Takeaways

Design oversight around what arrives on its own, because anything requiring you to go and look is a desk activity.
Reconcile athenaOne departments to the sites you actually manage before trusting any cross-site comparison.
Publish one rate per site in the same units, and pick the one closest to revenue rather than the easiest to collect.
Alert on configuration changes a site makes without telling you, since remote managers learn about them from metrics or not at all.
Keep the rule central and the threshold local, so a real deviation fires everywhere and normal variation fires nowhere.
Leave trend analysis, rate comparison, and threshold tuning for the weekly review, and keep the travelling version to exceptions.

The reason multi-site oversight breaks is not that the numbers are hard. It is that the person who needs them is moving and the system was built for someone who is not. Fix that by inverting the direction: one comparable number per site pushed at a fixed time, exception alerts on thresholds you set locally, and a link that opens the case in athenaOne when you want the detail. Keep the analysis for the desk. The road version only has to tell you which building to worry about.

Sources