Process · Web

Why did a landing page need ten people?

At Genomed every new landing page started with a call of about ten people arguing over headlines and button colours. I built a template engine of shared blocks, and from then on I made a landing page alone in one or two days: images, copy and HTML.

Genomed · 2022 to 2024 ·
My role
Senior Product Manager, owner of the web channel
Company
Genomed, the largest genetic testing lab in the CIS, selling tests online
Stack
MODX CMS, shared blocks, HTML
The constraint
Hundreds of tests, each worth a page, and every page treated as a project.

If you only read one thing

A landing page used to go through a ten-person call, mockups, sign-offs and a development cycle. I replaced that with shared blocks in the CMS that I could assemble on my own. A new landing page now took one or two days instead of weeks, and a block that won an A/B test went live on every landing at once.

The situation

When I joined, a new landing page for a test was a project. The first call brought together about ten people: the owner, heads of departments, a developer, a project manager, doctors, the person in charge of processes, a designer and QA. Most of the time went on headlines and button colours. Then came desktop and mobile mockups, sign-offs, front-end, back-end, testing.

Genomed sells hundreds of tests. At that speed only a few of them would ever get a proper page, and every page was built differently.

The fork

Option A

Make the process faster

Fewer people on the call, tighter agendas, more developer time. Still a project per page, just a shorter one.

Option B · chosen

Take pages out of the project process

Build the page once, as shared blocks, and let a product manager assemble new pages without designers, developers or calls.

The meetings were not slow because people were bad at meetings. They were slow because every page was designed from scratch, so everyone had an opinion on everything. Standard blocks removed most of the questions the call existed to answer.

What shipped

A template engine in MODX built from shared, standardised blocks, with prices and turnaround times pulled in dynamically. Knowing only HTML, I could put together a landing page for any test on my own, either from an incoming request or because I saw demand for a test in search data. I made the images in Photoshop, wrote the copy and did the markup myself.

Because the blocks were shared, improvements compounded. When a variant of a block won an A/B test on one page, it went live on every landing page built on the engine.

Before: every landing page as a project Call with~10 roles Brief Desktopmockup Mobilemockup Sign-off Front-end Back-end QA Release weeks per page, ten people on the first call, debates over button colours After: a page assembled from shared blocks Demand ora request Images, copy, HTML from shared blocksall by me Release 1–2 days A block wins an A/B teston one landing Rolled out to every landing struck through: steps a new landing no longer went through
What a new landing page no longer had to go through.

What happened

1–2 daysto assemble a new landing page, instead of weeks
1person instead of about ten
~$0.5K → ~$5KDNA ancestry revenue a month through the web channel

The DNA ancestry page is one example of what the engine made possible. I assembled it on the engine and promoted it through SEO, with no ad budget. In January to November 2024 it was one of the ten most visited pages on genomed.ru, with 57,000 visitors.

Before December 2023 ancestry tests brought the web channel about $500 a month. By the end of 2024 it was $4.5–5.5K a month, and over sixteen months the page earned about $31K above the old level, at monthly average exchange rates. Orders went from 0–6 a month to a peak of 39.

Genomed DNA ancestry landing page: 'DNA test for ethnicity. Learn more about your origins', with buttons 'Choose a test' and 'Discounted bundles'.
The DNA ancestry landing page, assembled from the shared blocks.

The best review the page ever got did not come from analytics. A while later a competitor copied it almost one to one: the structure, the blocks, the copy.

Nobody copies a landing page that does not sell.

DNA ancestry orders through the site's checkout orders a month · December 2022 to March 2025

0
10
20
30
40
Dec 22Jan 23Dec 23Jan 24Dec 24Jan 25
From my monthly sales tracking at Genomed. December peaks: the test is a popular New Year gift.

The takeaway

If every page needs a meeting, the problem is not the meeting. Standardise the parts people argue about, and the argument goes away with them.

All cases