Fewer leads. The same number of patients.

BioID Health is a private online hormone clinic in the UK, offering hormone replacement and testosterone replacement therapy. They compete with the NHS, which provides HRT for free. Every single person who lands on their site has a free alternative available. That is an unusual amount of pressure to put on a landing page.

Their existing page was long. Expensively long.

“The extremely long version we invested a large amount of money on and unfortunately it just has not worked. So, we are very cautious as you can imagine that when it comes to dealing with specialists who build landing pages, we want to be 100% certain we are getting what we ask for. We in all honesty don’t have money to waste at this point.”

BioID Health, first email

That was the first email, and it is worth reading twice, because it tells you what the real brief was. They had already been sold a landing page once.

The problem was not traffic. Their paid ads were working.

“Currently we are seeing a lot of people fill out our lead forms but not actually book an initial consultation which is the first step in our business.”

People were filling in the form and then evaporating. The clinic was paying for leads that never turned into patients, and the gap sat between the form submission and the first consultation. By the time they came back to us in September, they had put a number on it: the page was converting visits to patients at 0.5%, and they needed it in the 1 to 2% range.

The obvious move is to make the form easier. Fewer fields, softer commitment, bigger button. Submissions go up, the dashboard looks better, and everyone is pleased for a month. It would have made BioID’s problem worse. They did not have a submission problem. They had a problem where submissions were not turning into patients, and a lower-friction form produces more of exactly the leads that were already failing to convert.

For a private clinic competing with a free alternative, the page has a harder job than collecting an email address. It has to get someone to the point where they have decided that paying is worth it, before they fill anything in. So the rebuild qualified harder rather than softer. It did more of the convincing on the page, so that fewer, more decided people reached the form.

We ran the rebuild as a head-to-head A/B test in HubSpot against the live original, from 25 September to 14 October 2025, on real paid traffic. Here is the whole picture.

The rebuilt page pulled fewer views, fewer submissions, and fewer contacts, and it produced the same 28 paying patients. Time on page and bounce rate were effectively identical, which tells you the difference was in who converted, not in how the page was consumed. BioID promoted it to the live page.

“I completely agree, let’s go ahead with the winning variant.”

Patients show up weeks later in a different system, so most landing page work optimizes submissions instead. Read on those default metrics, the rebuilt page looks like a failure: fewer leads, less activity, worse numbers on every chart a marketing dashboard shows. The only reason this test could be read properly is that the funnel was instrumented all the way through to customer, not just to form fill.

A page that generates fewer leads and the same number of patients is not a worse page. It is a cheaper one, because every submission that never becomes a patient costs somebody a follow-up call. That is the whole argument for treating the website, the CRM, and the sales process as one system rather than three.