A/B testing agency
Convertica's A/B testing services are quoted one by one as a custom proposal. Tell us how many visitors your site gets and what you want to test, and we reply by email with a proposal for your site. Rated 4.7 on Trustpilot from 82 reviews.
More than 2,000 clients since 2017.
A/B testing services plan, build and read controlled experiments on your website. Visitors are split at random between the current page and a changed version, and the version that gets more sales, sign ups or leads, by a margin the data supports, is the one that ships. A/B testing is also called split testing.
The value of a test is the proof. A change that looks better can still lose, and a test shows that before the change goes live for everyone. The cost is time and traffic: every test needs enough visitors to reach a clear result. That is why the first question on the form above is your monthly traffic. For the method itself, read the A/B testing guide.
Proof
These were split tests Convertica's team ran for past clients. Each figure is shown with its metric, timeframe and source, exactly as published in the case study. Results vary by site; they are not promises.
Lead generation
+51.6%
Conversions (valuation tool submissions), 47 days
Source: Empire Flippers, valuation tool submissions, 47 days, split-URL test then rollout
Read the case study: Empire Flippers: a valuation tool submission process rebuilt
Great results. easy to work with, and do proper split-testing. Even noticed when I accidentally made a change to a page they were testing, and they stopped the test and started again to make sure I didn't distort results. The most important thing to know: they had a positive result on revenue.
Working with Convertica was a great experience. They always provided on-time updates, had great communication, and I liked that they were capable of optimizing various types of content through A/B testing. Would recommend!
Traffic first
The traffic you need for A/B testing depends on the tested page's conversion rate and the smallest change you want to detect, not on one fixed number. Fewer visitors means a longer test, and below some point a test cannot reach a clear result in any reasonable time. The worked example below shows how quickly the wait grows.
Worked example, computed with the same code as the A/B test sample size calculator: a page that converts at 3%, a test looking for a 20% relative lift (to 3.6%), 95% confidence, 80% power and two versions. That test needs 13,914 visitors per version, 27,828 in total. The table shows how long that takes at different levels of monthly traffic to the tested page.
| Visitors a month to the tested page | Days to reach the sample | Whole weeks to run |
|---|---|---|
| 5,000 | 170 | 25 |
| 10,000 | 85 | 13 |
| 30,000 | 29 | 5 |
| 50,000 | 17 | 3 |
| 100,000 | 9 | 2 |
This is an example, not a rule: a page that converts better, or a bigger change, needs fewer visitors, and a smaller change needs many more. Put your own numbers into the sample size calculator, and see how long to run an A/B test for why tests run in whole weeks. Count only the traffic that reaches the page you would test, not the whole site.
When an A/B test loses or ends flat, the original page stays live, and the result still tells you something: that change did not help those visitors, on that page, on that device. A loss costs the time the test ran, not the sales a bad change would have lost for good. Convertica's case studies publish these results beside the wins.
That is why results are read by page and by device, and why a loss is the start of the next test rather than the end of the work.
An A/B testing proposal from Convertica starts with the form at the top of this page and ends with a written proposal by email. There is no fixed package: what to test, how and at what price depends on your site, your traffic and what you sell, so each proposal is quoted one by one.
Four quick steps: monthly visitors, what you would like to test, whether you test today, and where to reply. Your website is the only thing we need to look at.
Traffic decides what is realistic. If a test could not reach a clear result on your pages, we say so, and suggest what would work better.
It sets out what we suggest testing and why, what the tests need in traffic and time, how results are read, and the price for your site. Nothing starts until you agree to it.
Not ready to test? The free CRO audit shows what is costing your page conversions in a few minutes, and it is a good source of test ideas.
A/B testing is not the right fit when the tested page gets too few visitors for a clear result, when the change is a fix for something plainly broken, or when you need the improvement live this week. In those cases, building the fix without a test is usually the better use of your time and traffic.
For those, Convertica's two rebuilds build the fixes without testing them: the free audit finds what to fix, then Convertica's team builds every fix it found, $497 for one page or $1,997 for your whole site. The one page rebuild covers one page, and the full site rebuild covers every page. A rebuild does not run A/B tests, so watch your sales or sign ups before and after the change.
Guides
You can A/B test almost anything a visitor sees or does on the way to buying: headlines, product pages, landing pages, the cart and checkout, forms, prices and offers. Start where traffic is high and drop off is clear. These guides cover the most common places to test.
Landing pages
What to test on a landing page, how to set the test up and how to read it.
Online stores
Test ideas for product pages, the cart and checkout, with the split tests behind them.
Shopify
The ways to A/B test a Shopify store, and what each can and cannot test.
Pricing
How to test prices and how they are shown, with the metric and the cautions that matter.
More on running tests: A/B testing examples, A/B testing tools compared, CRO testing and how to build a testing plan, what AI can do for A/B tests and the statistical significance calculator for reading a result.
Choose an A/B testing agency by asking how it decides what to test, how it sizes and ends each test, who builds the variants, how it reports results, and what you keep at the end. Specific answers are a good sign; vague ones are a warning. The same questions apply to Convertica.
Reviews
Verbatim from Trustpilot, Rated 4.7 on Trustpilot from 82 reviews.
Convertica was a fantastic agency to work with! They ran our A/B tests for over 6 months and helped us greatly improve our conversion rate. So many elements and ideas that I had on my online store were holding us back from optimal conversions and I would have had no idea if it weren't for the great work that they did. They were very communicative, professional, and easy to work with.
I would highly recommend hiring their services if you're looking to improve your conversion rate!
Convertica really goes the extra mile on delivering happy clients and results. Even when initial test didn't prove to be succesful, they'd keep iterating until they found a winner. They push through when needed! Thank you Ayumi & Team.
A/B testing services plan, build and read controlled experiments on your site: visitors are split between the current page and a changed version, and the result decides what ships. Convertica quotes A/B testing one by one as a custom proposal. Monthly traffic comes first, because it decides whether a test can reach a clear result.
A/B testing services plan, build and read controlled experiments on a website for a client. Visitors are split at random between the current page and a changed version, and the result shows which one gets more sales or sign ups. Convertica quotes A/B testing one by one as a custom proposal.
Convertica has no published price for A/B testing. It is quoted one by one as a custom proposal, and the proposal states the price for your site before any work starts. The free CRO audit costs nothing, and the two rebuilds, which build fixes without testing them, have fixed prices on the pricing page.
It depends on your conversion rate and the size of change you want to detect, not on one fixed number. As a worked example, a page converting at 3% needs about 27,828 visitors in total to detect a 20% relative lift at 95% confidence and 80% power. Check your own page with the A/B test sample size calculator.
An A/B test runs until it reaches the sample size planned before it started, counted in whole weeks so every day of the week is included. The time depends on the traffic to the tested page and its conversion rate. Stopping early because one version looks ahead is a common way to get a false result.
No. The one page rebuild and the full site rebuild build every fix the free audit found; they do not test the fixes. A/B testing is a separate custom proposal. If you want a change tested before it ships, ask for an A/B testing proposal.
Then a test would run for months or end without a clear answer. The free CRO audit still shows what is costing you conversions, and a rebuild builds every fix it found. You lose the proof a test gives, so watch your sales or sign ups before and after the change.
No. Nobody can honestly promise a percentage lift before seeing your data, and many tests end with no winner. Treat a promised lift as a red flag. What an agency can show is past results with their metric, timeframe and source, as Convertica's case studies do.
A/B testing and split testing usually mean the same thing: visitors split between two or more versions of a page. Some teams use split URL testing for versions on separate addresses, and multivariate testing for several elements changed at once.
Question not answered here? Ask Kurt Philip, Convertica's founder, on a one-on-one call.
Book a one-on-one call with Kurt Philip (opens in a new tab)
Ask Convertica
Convertica
Hi, I'm Convertica's AI assistant. Ask me anything about the free CRO audit, the one page and full site rebuilds, or any of the free tools.