A study of a clinic's ad budget efficiency: 2,698 chats, 494 clicks and 44 real leads

The marketer reported about six hundred leads a month. About twenty people made it to the doctor's chair. That is a thirty-fold gap, and in the aesthetic medicine clinic this case is about, it was explained the way it is explained almost everywhere: the front desk supposedly fails to close, the ads supposedly bring the wrong people. We analyzed six weeks of this clinic's chats, and it turned out both sides were wrong - the mistake was in the comparison itself.

This is a real case, the clinic is anonymized, and the numbers are given as they are.

Why this analysis was needed

The argument between marketing and the front desk is familiar to almost anyone who pays for advertising. Usually nobody has an answer, because nobody sees what happens between the click on the ad and the booked appointment.

The ads manager shows clicks. The booking system shows appointments. The gap between them stays a black box: how many people actually wrote, what they asked, at which message they went quiet and why. As long as the box stays closed, the argument runs on impressions, and whoever is louder wins.

In this clinic, the WhatsApp front line was handled by a WhaleBiz AI agent. It answered every enquiry and saved the whole conversation, together with the code of the ad the person came from. That is why the box could be opened.

The first finding: there were not six hundred leads

Of the 494 conversations opened from ads, exactly half looked the same: the person tapped the ad, a chat opened with an automatic text, and they did not add a single word of their own to it. Not one question.

From there, the funnel looked like this:

The "600 leads versus 20 appointments" comparison turned out to be wrong: 600 are clicks, not leads. Twenty bookings out of forty-four real leads is a close rate of about 45%, and that is a good number. The failure was not where the front desk receives the lead, but before it.

This finding is unpleasant for marketing but useful for everyone. The front desk turned out to be the healthiest link in the chain. Had the clinic solved the problem the "obvious" way, by replacing or adding staff, it would have spent money on a stage that already worked and left the broken one untouched.

Second: two ads were eating half the budget

Every conversation carried the code of the ad the person came from. That made it possible to calculate the return of each creative separately, not of the campaign as a whole.

The gap in return between the best and the worst ad was four-fold. The best creative, with a third of the traffic, produced more leads in absolute numbers than the highest-volume one.

Why this happens is clear enough. Ad platforms optimize for what they can see, and what they see is the click and the opened chat. An ad that collects lots of cheap clicks looks like a winner by their logic, and the algorithm happily gives it more and more impressions. The platform does not know that there are no conversations behind those clicks, because the conversations happen outside it.

Third, and most useful: what made the difference

The most valuable part of the analysis is not which ads worked, but why. The pattern held across the whole set of ads: explanation brings curiosity, a promised result brings intent.

Ads that explained how the procedure works collected many clicks and almost zero conversations. The person watched the video, tapped and stayed silent: their curiosity was already satisfied, and they had no question left. Ads that said what the person would get collected fewer clicks, but people wrote and booked.

A particularly clear example was a pair of creatives for exactly the same procedure: one had 22.7% leads, the other zero. It was not about the procedure or the audience, but about what the ad promised.

One caveat matters especially in medicine. "Promising a result" does not mean guaranteeing an outcome to a specific patient: that is not allowed in medical advertising, and good creatives do not do it. The point is to talk about the goal of the procedure and why someone should come to you, rather than how the device works or how long a session lasts.

Fourth: the ad radius makes no difference

The first thing people usually do with "low-quality" leads is narrow the geography. This clinic's data shows it does not help. The breakdown covers 493 of the 494 conversations:

A person who lives eight kilometers away behaves the same as someone who lives around the corner. Narrowing the radius does not buy quality, and widening it does not hurt. The real spread sits in the ad itself.

Fifth: 42% of enquiries arrive when the clinic is closed

Almost half of all messages arrived outside working hours, including at night, on Friday and on Saturday. All of them got a full answer with a price, an explanation and photos of results. Before the agent was connected, these enquiries would have waited until morning or the next working day.

This matters for the ad math too. The clinic pays for the click when the person taps the ad, not when the front desk shift starts. If almost half of enquiries arrive when nobody can answer, a large part of the budget is wasted before anyone even looks at ad quality. What response speed looks like across the market, we covered in our response-time audit of 98 private clinics, and Saturdays and holidays specifically in the piece on answering customers over the holidays.

What the clinic did with these findings

The analysis did not end as a report on a shelf, but as four concrete decisions:

1. Moved budget from the two high-volume, weak ads towards the two that performed.
2. Asked for budget broken down by ad code, to get a cost per lead for each creative rather than a campaign average.
3. Stopped blaming the front desk: its stage turned out to be the healthiest in the whole chain.
4. Got a rule for new creatives: promise a result, do not explain the procedure.

What this means for any clinic with an ad budget

Data like this does not come from the ads manager, and it does not come from the CRM. It only exists where every enquiry is handled by a system and saved in full, with the ad code and the whole conversation history.

That is why a clinic with an automated front line has something the others do not: it sees not only how much it spent, but exactly what happened to each person who came from an ad. The analysis is done monthly and turns into concrete budget decisions.

If you want to check your advertising without any system, here is the minimum you can do right now:

• Separate clicks from leads. A chat opened with an automatic text is a click. Only someone who wrote at least one word of their own can become a lead, and even among them only a minority become real leads.
• Measure return per ad, not per campaign. An average cost per lead hides the gap between the best and the worst creative.
• Check what your ads promise. If most of them explain the procedure, you are probably buying curiosity.
• Leave the radius alone until you have sorted out the creatives.
• Look at how many enquiries arrive outside working hours and how many of them get an answer right away.

Doing this by hand takes a long time: the chats have to be exported, tagged and matched to ad codes. Where an AI sales agent handles the front line, this data is already in the CRM and collected without staff involvement.

How to get the same data for your clinic

If a WhaleBiz agent handles your front line, all the data is saved automatically. On request we will turn it into an analytical report with clear recommendations for your advertising specialists. If you do not have an agent yet, plans and setup costs are published on the pricing page.

You can write to us on WhatsApp or through the form above.

Questions and answers

Where do the numbers in this case come from?

From the clinic's own chats over six weeks: 2,698 conversations with 1,789 different people, 494 of them opened from ads. The WhatsApp front line was handled by a WhaleBiz AI agent, so every conversation was saved in full, together with the code of the ad the person came from. The clinic is anonymized and the numbers are given as they are.

Why can't you see this in the ads manager or the CRM?

The ads manager sees the click and stops there. The booking system sees the appointment and starts there. Everything in between, meaning whether the person wrote a single word, what they asked and at which message they went quiet, reaches neither of them. This data only exists where every enquiry is handled by a system and saved together with the ad code.

So should ads that explain the procedure be switched off?

Not necessarily switched off, but they should not get most of the budget. These ads are good at collecting attention and clicks, but they rarely lead to a conversation. The clinic in this case moved money towards ads that say what the person gets, and made that the rule for new creatives. The point is to name the goal of the procedure, not to guarantee an outcome: medical advertising must not promise a specific result.

Should I narrow the ad radius to get better leads?

According to this case, no. The share of leads within 1 km, 1-3 km and 3-10 km is almost the same: 9.3%, 9.2% and 8.6%. Narrowing the radius does not buy quality, and widening it does not hurt. The spread in results comes from the ads themselves, not from the distance to the clinic.

How do I get an analysis like this for my clinic?

If a WhaleBiz agent handles your front line, all the data is saved automatically, and on request we turn it into an analytical report with recommendations for your advertising specialists. If you do not have an agent yet, start with setup: plans and what each includes are published on the pricing page, and you can discuss your case through the form on this page or on WhatsApp.