Evaluate one real experiment before changing your platform.
Plan details on this page reflect the published pricing page as of 7 September 2026.
Two questions decide whether a platform evaluation is worth starting: what it costs, and what happens if it goes badly. Both are answered on this page, in full, without a call. Nothing below is held back for a demo.
The button opens your mail client with the checklist already written out. Nothing reaches us until you press send.
What the evaluation costs
Every term below is the same one published on the pricing page. There is no evaluation rate, no pilot contract, and no separate proof of concept agreement to sign.
| Question a buyer asks | The answer, in full |
|---|---|
| How long is the trial? | 14 days, on monthly billing. Full access to every feature on the plan you pick, on your own traffic. |
| Is a card required? | Yes. A card is required at signup. We say so here rather than at the form, because meeting it unannounced is the thing that wastes your afternoon. |
| What is charged today? | Nothing. Nothing is charged today. |
| When does the first charge land? | Day 15, at your plan's monthly price, unless you cancel before then. |
| How do we cancel during the trial? | From your billing page, at any point during the 14 days. No email, no retention call, no notice period. |
| What if we cancel late? | A 30 day money back guarantee covers your first paid month. Beyond that, cancelling stops future charges and is not prorated: you keep paid access to the end of the period. |
| Where is the trial not available? | Annual billing and Enterprise. Both are ineligible, so start monthly and switch afterwards if annual suits you better. |
| Who takes the payment? | Paddle, as merchant of record. Prices exclude tax and applicable tax is handled at checkout. |
Starter and Pro, the two plans an evaluation sits on
Business and Enterprise exist and are on the pricing page. They are left out here because an evaluation almost never needs them: the decision you are making is whether the platform holds up on one experiment, and both plans below run real A/B tests on day one.
| Starter | Pro | |
|---|---|---|
| Monthly | $99 / mo | $249 / mo |
| Annual | $79 / mo, billed $948 a year | $199 / mo, billed $2,388 a year |
| Tracked users a month | 25,000 | 100,000 |
| Active experiments | 5 | Unlimited |
| Sites and seats | 2 sites, 2 seats | 5 sites, 3 seats |
| Included on both | A/B and split URL testing, the Monaco code editor, page and URL targeting, behavioral and audience targeting, goals with two system defaults, preview and QA links, frequentist statistics, segment filters, revenue metrics with outlier control, scheduling, clone, CSV export and change history. | |
| Pro adds | Not included | Geo targeting, activation triggers, sequential and Bayesian statistics, experiment heatmaps, time to decision, global JS, variation libraries, and the per variant speed guardrail. |
A tracked user is one unique visitor bucketed into at least one experiment during the calendar month, counted the first time they are bucketed and counted once per organization rather than once per site. A visitor bucketed into a control arm that changes nothing still counts.
The 30 minute assisted setup
Optional, free, and the fastest way to find out whether this is going to work on your stack. Half an hour, on your own site, with the person who built the product rather than a sales engineer reading from a deck.
What the half hour covers
- Installing the loader. A few lines in the head of your page, and a check that the runtime is arriving.
- Rendering your variation. We open a preview link on your real page and confirm the change looks right before any visitor is bucketed.
- Checking the instrumentation. That the right visitors qualify, that exposure fires only after consent, and that the goal fires on the action you actually defined.
- Answering the awkward questions. Consent wiring, single page app routing, and anything about your selectors that looks fragile.
What it is not
- Not a demo of someone else's account. If we are not on your site looking at your page, the call is not worth your time or ours.
- Not a qualification call. There is no budget question and no procurement step before it.
- Not a commitment. It carries no obligation, and it does not shorten or start the trial clock.
- Not required. The docs walk through the same steps unattended if you would rather not talk to anybody.
To book it, say so in the same email you send the experiment in. There is no form and no scheduling link to chase.
Send us a redacted experiment
You get back three plain lists: the parts that carry across as they are, the parts that have to be re authored by hand, and the parts ABTestly cannot reproduce at all. The third list is the one that matters, because it is the one that tells you not to switch. If you would rather see what that looks like before sending anything, one experiment built end to end shows the whole thing on a worked example, and switch one experiment sets out the longer version of this offer.
Where ABTestly may not be the right fit
This section is here because a page that only argues one way is a brochure. Each of the following is a real reason to stay where you are.
- Your program depends on a visual editor used by people who do not write code. Variations here are JavaScript and CSS you review in a code editor. That is a deliberate choice, and it excludes teams who need a point and click workflow.
- You need multivariate testing, personalization campaigns, feature flags, or server side experiments. ABTestly does A/B and split URL tests with multiple variants. It does none of those four.
- You need session recordings or surveys. Experiment heatmaps are included from Pro, and the Signals add on brings heatmaps to Starter. Recordings and surveys are not part of this product.
- You are not testing yet. There is nothing to evaluate. Start with the A/B test calculator and the methodology guides instead.
- Your renewal is close. If you are on an Optimizely contract, notice periods can be long enough that giving notice today matters more than evaluating anything. Check your renewal date before you start.
What you can and cannot learn in 14 days
What 14 days does show
- How long the same change takes to build here.
- Whether the variation code says what it does, in a diff you can read.
- Whether exposures and goals fire where you expect, only after consent.
- What the results panel tells you, and whether it says plainly that it is too early to decide.
- What it costs at your tracked user volume, against your renewal quote.
What 14 days does not show
- A winner. Whether an experiment reaches a verdict depends on your traffic and effect size, not on the trial length. Put your own numbers in the calculator first.
- A head to head conversion rate race. Run the same change in two tools at once and the numbers stop meaning anything: some visitors land in both, some in neither, and two scripts fight over the same elements.
- An exact numbers match against your current tool. Even done properly, two tools will not agree to the decimal. Identity, attribution, consent handling and goal definitions all move the figures for legitimate reasons.
- Your history. Past results do not import from any vendor. Export yours before you cancel anything.
Common questions
Is a card required to start the trial?
Yes. A card is required at signup. Nothing is charged today, and your plan's monthly price is charged on day 15 unless you cancel. You can cancel from your billing page at any point during the trial. The trial runs on monthly billing only, so Enterprise and annual billing are not eligible.
What happens if we decide not to switch?
Cancel before day 15 and nothing is charged. If you have already been billed, there is a 30 day money back guarantee on your first paid month. Either way the variation code we write is yours to keep, and you can paste it into your own repository whether or not you stay.
Do you need access to our current testing tool?
No. We never ask for a login to Convert, VWO, Optimizely, AB Tasty or PostHog, and you should not hand one to any vendor pitching you. Send the structure of the experiment with the URLs and the numbers stripped out if you prefer. The structure is what we read.
What does the 30 minute assisted setup cover?
One call, half an hour, on your own site. We install the loader with you, confirm the variation renders on your real page through a preview link, and check that exposures and goals fire where you expect them to. It is free, it carries no obligation, and it is booked by email rather than through a form.
Which plan should we evaluate on?
Starter at 99 dollars a month if the experiment is a straightforward variation on page load and you are under 25,000 tracked users a month. Pro at 249 dollars a month if you need geo targeting, activation triggers, sequential or Bayesian statistics, experiment heatmaps or the speed guardrail. You can start on either and move later.
Checked on 7 September 2026. Prices and product behaviour change, so verify against the source before you rely on a number. Plans, trial terms and the tracked user definition: abtestly.com/pricing. Refund terms: abtestly.com/refund. Signals add on: abtestly.com/signals. Install and consent: docs.abtestly.com. ABTestly prices exclude tax and are billed via Paddle as merchant of record.
ABTestly is an independent A/B testing platform for CRO developers and technical teams. Convert, VWO, Optimizely, AB Tasty and PostHog are trademarks of their respective owners. ABTestly is not affiliated with, endorsed by or sponsored by any of them.