01
Build your versions
Create up to three challengers in the campaign editor, then preview each one on your page.
A/B testing
Stop guessing which version of your site converts better.
Test headlines, offers, and calls to action on the pages you already have, with no code. RightMessage checks the results every week and tells you in plain English when a version wins.
No second testing tool. No statistician. No annual contract.
Original page · A/B testStarts a trialSplit 34 / 33 / 33
Acme
PlatformCustomersPricing
Log inOne workspace for every project, team, and client.
Acme helps teams plan, track, and deliver work without juggling five different tools.
Start your free trial
Most A/B tests find a winner that doesn’t exist.
Refresh a dashboard every morning and, sooner or later, random noise hands you a “winner.” RightMessage checks once a week instead, and its verdicts stay valid however often you look.
02
Know how long it’ll take
Before you start, see the time to a verdict based on your last 28 days of traffic.
03
Get a weekly verdict
Results are checked every Friday. A version wins only when the evidence is clear, however often you look.
04
Keep the winner
Tests never end on their own. Pick the version to keep, and RightMessage can show a page winner until your site has the change.
Plain-language results
See what won, by how much, and what’s still too close to call.
Every version gets its visitors, conversions, difference from Control, and a likely range. A version wins only when its whole range beats Control. Every version lands in GA4, Google Tag Manager, or PostHog too.
Campaigns/Homepage testIllustrative results
Page test
Original page
Goal: Starts a trialWeek 2 of up to 26Split 34 / 33 / 33
VersionVisitorsConversionsRateDifferenceLikely range vs Control worse · 0 · betterResult
Control
Your page as it is · 34% of visitors
Visitors1,695
Conversions183
Rate10.8%
Difference—
—
Outcome headline
33% of visitors
Visitors1,638
Conversions203
Rate12.4%
Difference+14.8%
22% worse to 62% better
30-day trial offer
33% of visitors
Visitors1,651
Conversions193
Rate11.7%
Difference+8.3%
33% worse to 58% better
Difference from Control by week
Descriptive. Only the weekly checks can call a verdict.
Weekly trendHow verdicts workSent to your analytics
What you can test
Test the page everyone sees, each audience, or the whole campaign.
The page everyone sees
Test your page as it is against up to three versions. No audiences needed.
Each audience’s message
Test an audience’s changes against up to three challengers, like a different offer for returning visitors.
The whole campaign
Hold back a percentage of visitors (10% by default) to measure what the whole campaign adds.
Test where you personalize
Find out what your visitors actually respond to.
Start your trial, or book a demo to see testing on your own site.
Common questions
Questions about A/B testing.
Which plan do I need for A/B testing?+
Campaigns and the campaign holdback are on Scale ($249 per month, or $199 per month billed yearly). Starting or restarting page and audience A/B tests, plus question, form, offer, and flow split tests, needs Pro ($499 per month, or $399 per month billed yearly). Ending a test, Deploy, and Undeploy don't need Pro, and a running test keeps running if you move from Pro to Scale. Grow doesn't include campaigns, so campaign tests stop showing on your site. Sending tests to GA4, Google Tag Manager, or PostHog needs no extra plan beyond the tests themselves. Wait for cookie consent and Exclude this browser are on every plan.
How long until I get a verdict?+
It depends on your traffic and conversion rate. The earliest verdict comes at the week 2 check, about three weeks after the start. Before you start, RightMessage shows the time to a verdict and the smallest difference it can spot within 26 weeks, based on your last 28 days of traffic. On low traffic, verdicts can take months, or may not come. You still see a descriptive weekly trend from the first check.
Why don’t I see p-values or a confidence percentage?+
Campaign tests use weekly checks that stay valid when you look every week: when versions truly perform the same, about 1 in 40 tests names a false winner. Instead of p-values, each version gets a likely range and a plain result: Winner, Did worse, or No clear winner. Goal value is shown but never decides a winner. Legacy question, form, offer, and flow split tests still show confidence and p-values.
Will visitors see the original page flash before a test version?+
Not if the tested elements carry the rmcloak class. Those elements stay hidden until RightMessage decides which version to show, then reveal one by one, with a 3-second fallback on new snippets. The editor marks each tested element Protected or Not protected. Server-side rendering of tested campaigns through RightMessage Edge needs Edge 0.3.0, which isn’t on npm yet.
Can I analyze tests in my own analytics?+
Yes. RightMessage sends an impression for every version to each enabled and published GA4, Google Tag Manager, and PostHog app, with a version string such as RM-cpn_5didDYbc-control. Holdbacks and legacy split tests send them too. The “Send to your analytics” panel lists each version’s string.
Does it respect cookie consent?+
Yes. With Wait for cookie consent on, RightMessage reads TCF v2.2, Google Consent Mode, or your own consent call. Until consent, nothing is stored, visitors see Control, and no stats are sent.
Can my AI assistant run tests?+
Righty and the RightMessage MCP can draft versions, set up tests, and read verdicts. Starting, restarting, ending, and publishing stay in your hands: Righty never does them, and MCP changes are drafts until you publish.
When should I choose an enterprise experimentation suite instead?+
If you need multivariate or split-URL tests, server-side or feature-flag experiments, bandits, real-time results, or mutually exclusive experiment groups. RightMessage is a personalization platform with built-in A/B testing, not an enterprise experimentation suite.
One platform, from the first answer to the next change worth making.
Ask useful questions, recognize known leads when they come back, build audiences, personalize what each visitor sees, test what works, and learn what to change next. Each part feeds the others.
Website Personalization
Change headlines, CTAs, forms, and offers for each visitor without duplicating pages.
Learn moreA/B Testing
Test the page everyone sees and each audience’s changes, then read a plain-language verdict every week. You decide when to end a test and which version to keep.
You are hereAudience Segmentation
Build audiences from behavior, traffic source, quiz/survey answers, on-site activity, and customer data in your ESP/CRM, then reuse them across pages, forms, and follow-up.
Learn moreQuizzes, Surveys, and Forms
Ask useful questions, qualify each lead, and keep using the answers on later pages and in follow-up.
Learn moreVisitor Identification
See which known contacts from your ESP/CRM are on your site right now and what they’re doing, and alert your team when a hot one comes back. No IP guessing.
Learn moreInsights & Recommendations
See which audiences convert and where leads drop off, then get a specific recommendation for what to test next.
Learn more