A/B testing services that end in a clear answer
Our A/B testing services try changes to your pages, forms and offers on real visitors before you roll them out to everyone. Every test starts with a written hypothesis and a visitor target set before launch, runs until it reaches that target, and ends in a clear winner, a clear loser or a documented draw, with the numbers to show which.
What A/B testing services involve
An A/B test splits your visitors at random between the current version of a page, A, and a changed version, B, then compares how many in each group complete the action you care about. Split URL tests do the same with two separate pages, and multivariate tests change several elements at once on sites with heavy traffic. Because both groups visit at the same time, a test separates the effect of your change from seasons, ad changes and luck.
The hard part isn’t the software. It is choosing tests worth running, setting the test size before launch so nobody stops a test early on a lucky streak, and reading the result honestly. Many tests don’t produce a big winner, and a well-run loss is still useful: it stops a weak idea from going live across the site.
Signs you’re ready for A/B testing
- Your key page gets steady traffic every week
- Changes go live because someone likes them
- Past tests were stopped as soon as they looked good
- You send paid traffic to the same landing pages
- A redesign is planned and you want proof first
What’s included in our A/B testing services
The research, the build and the statistics, handled by one team.
Test plan and backlog
Ideas from analytics, heatmaps and customer questions, scored by impact, confidence and effort, so the best tests run first.
Written hypotheses
For every test: what changes, why we expect it to help, and which metric decides the result.
Test size before launch
Test length and the visitors needed, worked out from your traffic and current conversion rate, and agreed before any test starts.
Variation design and build
Variations designed and built in your testing tool, then checked on mobile and desktop and for flicker before launch.
Tracking and quality checks
Each variation checked for errors, with GA4 key events confirmed so results are measured on real leads and sales.
Results and roll-out
A plain report on every test, whether it won, lost or drew, with the numbers behind it and the winning version rolled out.
A/B testing in depth
Three rules that separate a test you can trust from one that only looks convincing.
Set the test size first
Checking a test every day and stopping it the moment B pulls ahead is the most common way to get a false winner. Early results swing widely, and a lucky streak can look significant. So before launch we work out how many visitors each version needs to detect the smallest change worth having, usually at 95% confidence, and we run the test in whole weeks so weekday and weekend visitors are both counted. The test ends when it reaches that number, not when the chart looks good.
One question per test
A test that changes the headline, the image and the form at once can tell you that B won, but not why. Where we can, we keep each test to one clear question, so the result teaches you something you can use on other pages too. When traffic is high enough, a multivariate test can look at several changes together, and we tell you when that is worth it.
Low traffic and search engines
A/B tests need volume. If a page gets too few conversions to reach a valid result in a reasonable time, we say so, then test a busier page or measure a change before and after instead. Tests also follow Google’s guidance for website testing: no showing search engines different content from visitors, a canonical tag on test URLs, temporary 302 redirects for split tests, and tests that run only as long as they need to.
How our A/B testing services work
A steady testing rhythm, with a progress update every week and every result, wins and losses, in the monthly report.
- Step 01
Research
Analytics, heatmaps and recordings reviewed to find where visitors drop off and what might help.
- Step 02
Plan
Tests ranked, each with a hypothesis, a primary metric and a visitor target, approved by you.
- Step 03
Build and run
Variations built, checked on real devices and launched, then left alone until the target is reached.
- Step 04
Decide and roll out
Winners rolled out, losers documented and the next test chosen from what was learned.
Why Griffon for A/B testing
The same team designs the variations, builds them, checks the tracking and runs the ads that send traffic to the page. So a test idea doesn’t wait on three suppliers, and a winning version can go live on the site quickly. You get a progress update every week and a monthly report that includes the tests that lost.
Your testing tool, GA4 and website stay in your accounts, and the test data is yours. Once the work is paid in full, the designs, code and test records are yours too.
Part of our Conversion rate optimization ↗2 hours
our usual reply to a new inquiry, one business day at most
Weekly
progress updates, plus a monthly report with the data and the work done
One team
strategy, design, development and marketing under one roof
Google Partners
our team keeps its Google Ads certifications current
Questions about A/B testing services, answered
What do A/B testing services include?
Research to find what to test, written hypotheses, test size planning, building and checking each variation, running the test, reporting the result and rolling out the winner. We also tell you when a page lacks the traffic for a valid test.
How long does an A/B test take?
Often two to six weeks, depending on your traffic and how small a change you want to detect. We work out the length before launch and run in whole weeks, so a test isn’t cut short by an early lead.
How much traffic do we need?
It depends on your conversion rate and the size of change you’re looking for. Pages with only a handful of leads a week usually can’t support a valid test in a reasonable time. We check your numbers first and suggest before and after measurement if testing isn’t practical.
Which testing tools do you use?
We use the tool that fits your site and budget, such as VWO, Convert or Webflow Optimize, set up in your own account. Google Optimize closed in 2023, so older setups usually need replacing.
Who owns the tests and the results?
You do. The testing tool and analytics are in your accounts, and the designs, code and test records are yours once the work is paid in full.
Get a free testing plan
A 30-minute call. No deck, no pressure. Bring the site, the numbers and the goal, and leave with a plan and a price.