Methodology
How these case studies are counted
Every number on the eight case studies follows the rules on this page. Where a rule makes a figure less impressive, the rule wins.
8 accounts · Published 2026-08-05
Source
Where the numbers come from.
These are accounts I personally managed. The work was delivered inside an agency team, and the analysis, the decisions described and the conclusions drawn are mine. Client names, campaign names and creative are withheld.
Spend, leads, cost per lead, clicks, CPM, CTR and the video engagement rates come from the Meta Ads platform export for the account, unmodified. Two figures per account do not: the qualification rate and the close rate are the clinic's own, reported from their records. The qualified-lead and patient counts are then derived from those two rates, leads multiplied by the qualification rate and that figure multiplied by the close rate, rounded once for publication. They are arithmetic on a reported rate rather than a headcount, and nothing on these pages should be read as a person-by-person audit of a clinic's patient records.
The clinic is anonymised and the market is not named. What transfers to a New Zealand clinic is the method and the shape of the cost, not the cost itself: auction prices move with the market, the season and the offer. Treat every dollar figure here as evidence of how the account was run, not as a price you should expect.
Definitions
What each word means here.
Advertising vocabulary is loose, and loose vocabulary is how misleading numbers get published without anyone lying. These are the definitions used on every page in this set.
- Lead
- A person who completed the form or the messaging flow the ad pointed at. Counted once.
- Qualified lead
- A lead the clinic judged worth booking: correct service, contactable, and not already a patient. The clinic makes that call, not the platform and not the agency.
- New patient or new client
- A qualified lead who attended and paid for a first appointment.
- Cost per lead
- Spend divided by leads, for the account and date range shown.
- Leads per click
- Leads divided by link clicks. The share of people who clicked and then actually completed the form. It is the metric CTR is most often mistaken for.
- Form conversion rate
- Leads divided by the number of people who opened the form. It separates a weak form from weak traffic.
- Hook rate
- A custom metric, not a column Meta ships by default. It is built in the ad account from Meta's three second video plays, which count plays rather than people and include autoplay, so a high hook rate is not evidence that anyone chose to watch. The formula behind the figures on these pages is not recorded in the account material this set was built from. A diagnostic for the first frame, not a result.
- Hold rate
- A custom metric, not a column Meta ships by default. It is built in the ad account from Meta's video watch figures and describes how far into a video the audience stayed. It scales inversely with video length, so a short clip reports a higher rate than a long one with no difference in quality, and the creative runtimes for these accounts are not recorded. A diagnostic for the edit, not a result.
- Frequency
- Average number of times one person saw the ad over the date range. Rising frequency on a flat budget usually means the audience is exhausted.
- CPM
- Cost per thousand impressions. An auction price, not a performance measure.
- CTR
- Click-through rate: clicks divided by impressions. Appears here only where it contradicts the cost, which is most of the time.
Currency
United States dollars, converted once.
These accounts were billed in United States dollars, so United States dollars are what the pages report. Every dollar figure in the headline metric strips carries a New Zealand dollar figure underneath it in smaller type, converted at a single fixed rate so the two never drift apart from one page to the next. Figures in headings, prose and charts are United States dollars, unconverted.
The rate is 1 USD = 1.7015 NZD, taken from ExchangeRate-API open endpoint (open.er-api.com/v6/latest/USD) on 5 August 2026. It is a fixed figure written into the site, not a live feed. The real rate on the day any of this money was spent was different, and the New Zealand figures are therefore an indication of scale, not a record of what was billed.
Rounding
Why the arithmetic sometimes lands a few cents out.
If you divide a published spend figure by a published count, you will occasionally get a result a few cents away from the published rate. That is not an error, and it is worth explaining rather than quietly smoothing over.
Two figures per account come from the clinic rather than the platform: the qualification rate and the close rate. The qualified lead count and the patient count are derived from them, leads multiplied by the one rate and then by the other, and rounded once for display. Cost per qualified lead and cost per patient are calculated on those derived counts before the rounding, which is why dividing a published spend by a published count can land a cent or two away from the published rate. Cost per lead is the one rate that is always exactly the published spend divided by the published lead count, and it is checked automatically every time this site is built.
Totals across accounts work the other way. The portfolio and per-industry figures are sums of the whole-number counts published on each of the eight pages, so they reconcile exactly and the site refuses to build if they do not. Worth being plain about what that does and does not prove: the lead counts being summed are platform figures, and the qualified and patient counts being summed are the derived ones described above. The arithmetic is checked. The rates underneath it are the clinics'.
Limits
What this set does not prove.
These are not New Zealand accounts.
No account in this set ran in New Zealand. New Zealand clinic advertising is regulated differently, priced differently and competed for differently. Nothing here should be read as a forecast of what a New Zealand clinic will pay per lead.
Eight accounts is eight accounts.
It is enough to show a method and a pattern that repeated. It is not a sample large enough to support a rule. Where a pattern held in some of these accounts and not in others, this set says so and shows which, rather than flattening it into a claim about all eight.
A lead is not a patient, and a patient is not a course of care.
Everything published here is lower funnel: leads, qualified leads, patients, and the cost of each. Reach, impressions, engagement and video views appear only as diagnostics, and never as a result. An account that bought a great deal of attention and no patients is a failed account.
Nothing here is a clinical claim.
These are advertising accounts. No figure on any of these pages says anything about treatment, outcome or clinical benefit, and none should be read that way.
The free Ad Check
Two written pages on your clinic's advertising. Free.
Your ads, site and Google profile wording, reviewed against your profession's advertising rules, the ASA Therapeutic and Health Advertising Code and the Fair Trading Act. In your inbox within two NZ business days. No call, no follow-up sequence, no obligation.
Free · Async · Two NZ business days