Should the group take over lead handling from the locations entirely?

Once the spread becomes visible, the tempting fix is total: route every call to a central team and take the locations out of the loop. Sometimes right, often overcorrected — the honest answer is a layered one.

What centralizing does well

A central intake layer ends the roulette at the first touch: every inquiry, any location, gets the same answer speed, the same script quality, the same capture. It also creates the group's first real funnel data, because everything flows through one measured pipe. For the floor — the instant response, the missed-call recovery, the after-hours coverage — central is simply better, because floors are pure structure and structure centralizes cleanly.

What full centralization breaks

Local knowledge lives at the desks: the doctor's quirks, the parking situation, which hygienist is out Thursday. Callers eventually need the LOCATION, and a central team that can't answer local questions creates a new failure mode — the transfer loop — while the desks, stripped of intake, disengage from lead outcomes entirely. Fully centralized groups often win the first touch and fumble the handoff.

The hybrid most groups need

Centralize the floor, standardize the human layer. Central: the instant text-back, missed-call recovery, after-hours capture, and the measurement pipe. Local: the human conversations, running standardized sequences with per-location SLAs and monthly scorecards. The desks keep ownership of their patients; the group keeps ownership of the standard; nobody's lead survives or dies by local habit anymore.

Sequence it by evidence

The scorecard decides the order: if probes show first-touch failure (rings-out, slow responses), build the central floor first. If first touch holds but follow-up scatters, standardize sequences first. Guessing the order costs quarters; the free score estimates it in three minutes, and the $500 discovery measures it site by site, timestamped.