$58.03
Average cost per enquiry across the full eight months: $117,607 spent, 2,027 enquiries. The average hides the shape. The first five months ran at $76.32 and the last three at $42.92.
Cosmetic and plastic surgery practice, New York City
Twice the patients on a third more budget
Eight months, $142,799 in managed spend, and 2,654 enquiries. The interesting part is not the total. It is that the second half of the period produced a booked enquiry for 44 percent less than the first half, on roughly the same monthly budget.
2,654
enquiries in eight months
44%
cheaper per enquiry, second half
2.3x
monthly enquiries, January to July
$142,800
managed across both channels
The account
A working account that had stopped improving
This was not a rescue. The account was profitable when we took it over and it stayed profitable throughout. That is a harder brief than a turnaround, because there is no obvious thing to fix and no low bar to clear.
What it had was a plateau. Through the first five months of 2026 the account produced between 176 and 198 enquiries every month, at a cost that drifted from 65 dollars up to 91. More budget went in and roughly the same number of enquiries came out. Spend rose 35 percent across those five months and monthly enquiries did not move.
In aesthetic surgery that is an expensive place to sit. Clicks run above four dollars, the procedures are considered rather than urgent, and a practice that plateaus is usually paying more each month to reach the same people.
Test wide, then commit to what the data picked
Most accounts at this size are optimized by adjusting what already exists. Bids get tuned, negatives get added, budgets shift a few percent at a time. That is safe, and it is why plateaus persist: nothing changes enough to learn anything.
We did the opposite. May was run as a test month with the budget deliberately spread across procedures we had not proven, accepting a worse month in exchange for a clear read on which ones could scale. It cost 91 dollars per enquiry, the worst figure of the year.
June spent that answer.
The work
Test wide, then commit to what the data picked
Five steps across eight months. The one that mattered was the decision to stop guessing which procedures deserved the budget and let two months of deliberate testing answer it.
Step 01 of 05
01
Establish what a plateau actually costs
Before changing anything we set a baseline across January to April: 185 enquiries a month at 73 dollars each, on a budget climbing steadily. Written down, that made the problem legible. The account was not failing, it was absorbing more money without producing more patients.
02
Run May as a test, and accept the cost
May was deliberately the widest month of the year. Budget was pushed across procedure lines we had not proven at scale, knowing the blended figure would suffer. It came in at 91 dollars per enquiry against a 73 dollar baseline.
That is the part most agencies will not do, because a monthly report showing a worse number is uncomfortable to send. It is also the only way to find out which procedures have room in them rather than guessing.
03
Read the test and reallocate hard
The test separated procedures that scaled from procedures that only spent. In June the budget moved decisively toward the first group, with spend cut by more than half on the lines that had not responded and concentrated on the ones that had.
Nothing about the total budget changed. June spent slightly less than May.
04
Hold the line while it compounds
June produced 282 enquiries at 57 dollars. July produced 429 at 38. August held at 399 and 39. The temptation at that point is to raise the budget immediately, which usually pushes cost back up before the account has settled. We left it flat for three months and let the improvement establish itself first.
05
Keep the two channels honest about each other
The account also runs Local Services Ads, and we report it separately rather than blending it into the total. Across the same eight months it stayed flat, between 37 and 45 dollars per enquiry, while paid search improved by 44 percent.
That separation matters. Blended into one number the result would look better than the work deserves, and it would credit a channel that did not change for an improvement that happened somewhere else.
Outcome
What eight months produced
44% cheaper
January to May against June to August, paid search only. Monthly spend was almost identical across both halves, so this is not a budget effect. The same money bought 44 percent more.
183 to 429
Paid search enquiries in January against July, the strongest month. Spend rose 36 percent over the same period, so volume grew roughly twice as fast as the budget behind it.
$40.15
The account is not finished. Paid search is winning roughly a third of the impressions available in its market, which means two thirds of the demand for these procedures in New York is still going somewhere else. The next twelve months are about closing that gap without giving back the cost per enquiry, which is a harder problem than the one we just solved.
We are also winding down the Local Services Ads side. After the dispute rules changed we stopped recommending the channel, and this is the last account of ours still running it.
Start here
See what your budget can actually buy.
Model your campaign before you commit a dollar, or send us the account and we will tell you what we would change first.