Landing page A/B testing
- Price
- from $1,500 / cycle
- Timeline
- 4–6 weeks per cycle
- Experience
- 100+ projects · 10+ years · 9+ countries
Trustworthy A/B tests: stable traffic split, cache-aware, judged by lead quality
- WordPress
Our usual stack. We bring in other tools when the task or the project calls for it.
An A/B test is only useful if you can trust it: every visitor must always see the same variant, the cache mustn’t mix variants, and you compare lead quality, not clicks.
We set up tests on your website — from a headline change to completely different pages — and help you draw conclusions.
What we do
- Hypotheses and a prioritized test plan
- Weighted traffic split with sticky variant assignment
- Cache and CDN handling so variants never mix
- Variant data passed to analytics and the CRM
- Results review that accounts for lead quality
Why teams stop trusting their tests
The test crowned a winner, but leads didn't go up
We compare variants by leads that reached the CRM and passed qualification, not by button clicks. A conversion winner sometimes loses on revenue.
Visitors flip between variants on repeat visits
The variant is pinned to the visitor with a server-side cookie before the page renders. No content flicker, no switching on the next visit.
The site sits behind a CDN and the cache serves one version
We split the cache by variant: the CDN stores both versions and serves the right one based on the visitor's tag, with no hit to page speed.
We test everything at once: button color, font, hero image
Hypotheses come from analytics and session recordings, ranked by reach and effort. What changes how people read the offer goes first, not a button shade.
Marketing can't launch a test without a developer
Variants live as blocks in the CMS and traffic weights as a setting, so marketing can start a new headline or offer test on its own.
When A/B testing makes sense
A good fit if
- Your landing page gets steady paid traffic, so small conversion gains show up in the budget
- The site runs on your own stack and you'd rather not add a third-party testing script
- Sales is willing to mark in the CRM which leads became deals
Not a fit if
- You get a handful of leads a month — a test would drag on; a conversion audit and a rebuild are faster
- The page is clearly broken on mobile — we'll fix the layout first and test later
Tests you can trust
One hypothesis per test
We change only what’s being tested, so the result can be explained instead of guessed at.No premature calls
We don’t stop a test on the first spike — we call it when the difference holds, not when it’s noise.Learning that compounds
We log every test’s outcome, so new hypotheses build on what’s already been proven.
Express website audit
We review your site and within 3 business days send a PDF with up to 10 specific growth points, sorted by priority.
- Load speed and the mobile experience
- Structure, copy and the path to a lead
- Technical SEO and search visibility
- Forms, analytics and conversion tracking
For live websites only. No strings attached: you can make the fixes in-house.
How we work
- 01
Funnel review
3–5 daysWe look at analytics, session recordings and CRM leads to find the step where most people drop off.
You get A list of hypotheses scored by impact and effort - 02
Page variants
3–5 daysWe write and design variant B and build it so only the tested element differs.
You get Variant B on a staging URL, approved by you - 03
Traffic split
2–3 daysWe set up a stable, cache-aware visitor split and tracking for leads and their quality.
You get A working 50/50 split with goals in analytics and the CRM - 04
Data collection
2–3 weeksWe run the test until it reaches the traffic it needs and watch that the split stays even.
You get A mid-test progress snapshot - 05
Readout and rollout
2–3 daysWe compare variants by conversion and lead quality and roll the winner out to everyone.
You get A readout, the winning variant live and the hypothesis for the next cycle
What the first test needs
- Access to the site's code or CMS, and to CDN settings if you use one
- Analytics with a goal set on lead submission
- Traffic and lead numbers for the past few months
- A CRM with deal statuses, so we measure quality, not clicks
- Any hypotheses you already have, even the controversial ones
Page audit and a test plan
We review the page and analytics for free and in 3 working days send up to 10 growth points. They become the queue of hypotheses for the first tests.
Not sure which
variant really wins?
We’ll draw up a prioritized test plan, set up a stable traffic split and judge results by lead quality, not clicks.
Many tools swap the page with a browser script — that means flicker and slower pages. We switch variants on the server, so visitors see their variant right away.
From $1,500 / cycle. A cycle covers the hypothesis, the variants, the test itself and the readout. The price depends on whether we change a block or a whole page, how many variants run and whether results are tied to lead quality in your CRM. We quote after reviewing the page and its traffic.
It depends on the expected difference and your conversion rate. Before launch we calculate how many visitors make the result reliable.
Yes: from a single block to two different landing pages on one URL.
Yes, we have a ready WordPress solution that works correctly with caching and CDNs.