Why Adgent
Product All features It reads Account memory Daily verdict Ground truth Trust Gate Creative intelligence Change Ledger It builds Campaign build Media plan Live dashboards Agents & reports
Solutions By team For agencies For in-house By industry Ecommerce & DTC Lead generation Travel & hospitality Marketplaces Local & multi-location Mobile apps & games
Pricing About Blog Request demo

For clinics, restaurants & multi-location brands

The conversion isn't on the internet.

Google does not measure store visits. It models them — extrapolated from a panel of volunteers, with thresholds it does not disclose and up to five days of delay [Google Ads Help, current].

In September 2025 those modelled visits started being counted as conversions by default, at a value Google assigns. Modelled foot traffic now flows into automated bidding next to real revenue, and nothing in the interface distinguishes them.

capacity over it, budget buys appointments nobody keeps
Not diminishing returns. Negative ones.
The other half of the conversion

Only 61% of callers reach a person.

In local businesses the phone is not a secondary channel, it is the conversion. Across more than 60 million calls, 37% of phone leads convert — against 2–5% for a form [Invoca, Jun 2025].

But well over a third of callers never reach anybody. No ad platform reports that — it happens after the click leaves their view — yet it sets the ceiling on everything the ad spend can achieve.

In healthcare it gets worse before it gets better

59% of callers reach a person, and of those, only 43% of agents ask the caller to book [Invoca healthcare report, 2025].Two filters, applied after every dollar of media spend, invisible to the account that spent it.

Inside one clinic’s account
Cost per patient here looks fine. Anything I should know?
The account average is hiding two campaigns. The service-based campaign is at roughly double this account’s cost per booked appointment, and its radius runs four miles past the catchment — into territory your other branch is targeting too, going by the geography in this account.

Separately, any call over thirty seconds is counting as a conversion here, so the bidding is learning to buy long calls rather than booked appointments — wrong numbers and reschedules clear that bar.I am reading this account only. I cannot open your other branch’s account or compare the two.
What would you do first?
Tighten the radius before you touch the budget — the overlap is costing you more than the spend level is. Show rate needs your booking system, and comparing two locations needs cross-account reading. Both roadmap, not shipped. What I can see today is this account’s targeting overlap, its geographic bleed and its call-quality signal.
Per-location read12 separate reads
twelve locations, twelve separate reads three came back flagged each on its own · never ranked

Each location is read separately. Adgent does not compare one to another.

Three mistakes that cost real money

Each one trains the algorithm to do more of it.

01

Call duration as a quality signal

A twelve-location dental group counted any call over thirty seconds as a conversion[Groas, Jun 2026]. That signal feeds automated bidding — so the algorithm went hunting for whatever produces sixty-second calls: wrong numbers, reschedules, vendors, job applicants. It did exactly what it was told.

02

Locations bidding against each other

Campaigns structured by service rather than by location put twelve offices into the same auction against one another. You pay a premium to compete with yourself, and the account-level report shows nothing unusual because in aggregate the numbers are fine.

03

Geographic bleed, on by default

The default targeting setting reaches people merely interested in your area. For a clinic with a ten-mile catchment, an out-of-area click is not low-probability — it is deterministically 100% waste, and it is the setting nobody checks.

The constraint no bidding algorithm has

Near full capacity, more budget buys no-shows.

This is the structural difference between local businesses and everything else on this site: you cannot serve unlimited demand. Once the schedule is close to full, incremental spend produces leads booked into distant slots — and distant slots no-show more. That is not diminishing returns. It is negative returns, bought at full price.

All three get worse with time rather than staying flat, because every one of them trains the bidding. A quarter of leaving it alone is a quarter of the algorithm getting better at buying sixty-second wrong numbers, better at putting your own branches into the same auction, better at finding people merely interested in your area.

Meanwhile the modelled store visits Google began counting by default in September 2025 keep flowing into that bidding at a value Google assigns.

The account does not drift toward neutral while you decide.

It compounds in the direction it was last taught — every day, on the budget you already approved.

  • 60sthe call length the bidding learned to buy, because duration was the conversion
  • 39%of callers never reach a person, and no ad platform reports it
  • 12locations read on their own, never ranked against each other
Where Adgent stands today

Two honest limits, stated up front.

Show rate and revenue per patient live in your practice-management or booking system — Adgent does not read those today. And cross-account comparison, which is what would let it spot two of your locations bidding against each other, is on the roadmap rather than in the product. We would rather you knew that before a demo than during one.

01

It reads every location's account daily

Including the ones nobody opened this week — which in a multi-location group is most of them, because attention follows whichever site is currently on fire.

02

It refuses targets your own history rules out

In a capacity-constrained business this stops being a trust feature and becomes a core one: a target that can only be reached by buying leads you cannot serve is a target worth arguing with before the budget goes out.

03

It audits geography and placement on demand

Radius bleed, placement drift and demographic waste are all questions you can ask in one message rather than build a report for — which matters when the answer changes per location.

Next

The two mechanisms that carry the most weight here.

A tool that refuses a target you cannot serve, and a read that covers every location rather than the one currently getting attention.

Running these accounts for clients?

The same read on every client you run — including the ones nobody opened this week, because attention follows whichever account is currently on fire. One senior-level pass per account per day, at the same price whatever they spend.

How this works across a book of accounts
Free per-location read

Find out what each location’s account is actually buying.

We will read every location’s account — one at a time — for the geographic bleed, the call-quality signal feeding your bidding, and the targeting default nobody checks.

  • Radius bleed and placement waste, per location
  • Conversion signals checked for what they actually count
  • Show-rate and cross-location comparison need systems we don’t read yet
What comes back

Per-location findings rather than a blended average — because a healthy group average routinely hides three sites losing money.

We usually reply within a couple of working days.

Thanks — we’ve got it.

We’ll be in touch within a couple of working days.