Key takeaways

Extra budget lands on the same page. If it converts 1.8%, every new dollar converts 1.8%.
Match the page to the reason someone searched. In healthcare that meant one page per condition.
Test the things that build trust: who's behind this, what happens next, what it costs. Leave the colours alone.
My changes ran together, so I can't split the credit. I can show the order I did them in.
AI can draft and compare page versions. I did this by hand, so that part is a plan I haven't run yet.

In February 2023 I joined Gaba Healthcare. It’s a telepsychiatry company that was spending about USD 20,000 a month on ads. Only 1.8% of the visitors those ads brought in booked a consultation.

More budget would just have sent more people to the same page. So I fixed the page first.

I built a separate landing page for each condition people searched for. Then I tested versions built around trust. Only after that did I shift 40% of the Meta brand budget into search.

The booking rate reached 5.0%. Daily qualified calls went from 12 to 20.

Those changes ran at the same time, so I can’t say how much each one added. I can only give the order: page first, spend second. The case study has the full account.

Try it

Check who’s arriving. If they’re the wrong people, fix the targeting first. If they’re right, the page is the next suspect.
Name the one action you want. A booking, a sign-up, a purchase. Count how many visitors take it today. That’s your starting rate.
Match the page to the search. One page for each reason people arrive.
Put trust where it’s seen first. Who’s behind this, what happens after the click, what it costs. Test versions of those, not colours.
Change one thing at a time where you can. I couldn’t at Gaba. Next time I’d space the changes out.
Only then add budget. A higher rate makes every later dollar worth more.

Where AI fits

I did this by hand. So how I’d use AI is a plan I haven’t run. I’d use it to draft page versions from a clinician’s outline, and to compare a page against the questions people actually search.

A person still checks every claim. In healthcare, a clinician signs off each one. The healthcare workflows guide covers the rules.

Common mistakes

buying traffic for a page you haven’t fixed
one generic page for every reason someone arrives
testing the look instead of the trust
claiming the whole lift for the page when other changes ran alongside it

This week

Find the page that gets most of your paid traffic. Write down how many visitors take the one action you want. Then list the reasons people arrive, and check the page answers the biggest one.

This is part of AI in performance marketing. Next up: Google and Meta run your ads now.

Sources

Gaba Healthcare case study: the booking rate, calls, budget split and timeline, from my own records of the account, February to December 2023.
Gaba’s condition pages: still live, and edited since I left, so the structure is mine and some copy may not be.

Frequently asked questions

How do I know the page is the problem and not the traffic?

Look at who arrives and what they do. At Gaba paid ads brought plenty of visitors and only 1.8% booked. When people arrive and few act, suspect the page first. Check the targeting too, so you're not fixing the wrong thing.

What conversion rate is good?

I don't have a universal number. Compare your page with itself: the rate before a change and after it, on the same traffic. At Gaba the booking rate went from 1.8% to 5.0%.

Did the page work alone?

No. I also moved 40% of Meta brand spend into high-intent search and built organic content. They ran at the same time, so I can't give the page its own share of the result.

Three-minute lesson on this Ads: more variants, better tests Marketing course, medium. Free, no sign-up.

More from the blog

All articles →