A/B Testing Tools Pricing Compared (2026)

Most enterprise A/B testing tools do not publish a fixed pricing table: VWO, Optimizely, AB Tasty, and Kameleoon’s core platforms all quote by contract, based on tested traffic, activated modules, and agreement length. Convert.com is a partial exception, with public figures for its entry plans. GrowthBook, PostHog, and Statsig follow a different logic entirely: usage-based or open-source pricing that starts far more transparently than an enterprise sales cycle. This guide does not try to rebuild a price table that does not exist. Instead, it teaches the real dimensions that determine cost in each of these categories and how to request a quote without wasting a sales cycle. For the full tool-by-tool landscape beyond pricing, see the neutral guide to CRO tools compared.

Quick overview

Tool	Billing model	Public pricing?	Trial or evaluation
VWO	Contract by tested traffic (MTU) and modules	No: plan names are public, no dollar figure	Self-serve free option, duration not published
Optimizely	Modular contract by product line	No: pricing page routes to a quote request	Quote on request
AB Tasty	Fully custom proposal, no fixed plan	No	Guided proof of concept, not a standard trial
Kameleoon	Custom contract for the core platform; newer PBX line has a published starting price	Partial: PBX Starter lists a monthly price on Kameleoon’s plans page; core platform stays quote-only	30-day free trial (PBX), plus demo/proof of concept for the core platform
Convert.com	Published tiers by monthly tested users	Yes, for entry plans; Enterprise on request	Free trial (confirm current length on site)
GrowthBook	Free open-source core (self-hosted) plus paid Cloud	Core: yes, MIT-style license; Cloud: partially public	Free Cloud Starter tier (up to 3 users)
PostHog	Self-hosted core plus hosted free tier	Free tier allowance published; usage above it is metered	1 million events per month at no cost
Statsig	Hosted, usage-metered by events	Free and paid tiers both published on pricing page	Free Developer tier (2 million events per month)

The rest of this guide breaks down each row, starting with the dimensions every vendor uses to build a final quote, whether they publish it or not.

The dimensions that make up A/B testing tool pricing

Before asking “how much does VWO cost” or “how much does Kameleoon cost,” it helps to understand that none of these companies sells a single product at a single price. The final number is a combination of at least four variables, and ignoring any one of them is the most common reason two quotes end up not actually being comparable.

Tested traffic (MTU, or monthly tested users) is the heaviest-weighted variable: it counts the unique visitors who actually enter an active experiment in a given month, not a site’s total traffic. A site with 2 million monthly visitors but only one test running on 5 percent of that traffic pays for a much smaller MTU fraction than the total suggests, and it is common for a vendor to ask for that specific number before quoting.

Domains and projects determine whether a contract covers a single site or several properties (different brands, different countries, separate apps and web) on the same account. Agencies and multi-brand operations typically pay an add-on per extra domain, or need a specific tier that already bundles several.

Included modules is the difference between contracting only the A/B testing engine or a full suite with personalization, heatmaps, session replay, and feature flags. The enterprise suites (VWO, Optimizely, AB Tasty, Kameleoon) bundle these modules differently, and turning on more of them almost always raises the price even when tested traffic does not change.

Self-hosted versus managed SaaS is the dimension that separates GrowthBook, PostHog, and Statsig from the rest: the open-source core of the first two can be installed on your own servers with no license fee, trading software cost for infrastructure and engineering cost. For the other side of that choice, see the guide to genuinely free A/B testing tools, which details the real limits of running a self-hosted tool without paid support.

VWO, Optimizely, and Kameleoon: enterprise contracts without a stable public price list

VWO, Optimizely, and Kameleoon’s core platforms follow the same pattern: contract-based quoting tied to MTU and modules, without a public price table stable enough to cite here with confidence. That was already true in our VWO vs Optimizely comparison, and it still describes the landscape accurately as of July 2026 (Kameleoon has one partial exception, covered below).

Checking VWO’s official pricing page directly: plan names (Growth, Pro, Enterprise) are listed alongside their respective features, but no dollar figure is shown publicly. The page offers a self-serve “explore free” path, but without stating a duration or usage limit right there, so treat it as an invitation to try the product, not a confirmed permanent free plan. Optimizely follows the same logic: its pricing page lists no values and routes every visitor to a quote-request page, where the final price varies by traffic, module, and contracted product line.

Kameleoon’s core enterprise experimentation and personalization platform sits in the same enterprise-contract category. According to its own plan comparison page, capabilities such as CUPED variance reduction and deeper personalization are gated to higher tiers rather than priced on a public monthly table, a pattern we cover in more depth in the AB Tasty vs Kameleoon comparison. One exception worth flagging: Kameleoon’s newer PBX (AI prompt-based experimentation) line does list a starting price directly on the company’s own plans page (checked July 2026), currently around 495 per month for a Starter tier capped at 50,000 tested visitors, plus a 30-day free trial with no credit card required. That published figure covers PBX specifically, not the broader enterprise contract for Kameleoon’s full experimentation and personalization suite, which remains custom-quoted. A recent industry event reinforces why any specific number circulating for these vendors deserves extra scrutiny: in January 2026, VWO (the brand behind the company historically named Wingify) and AB Tasty announced merging into a single digital experience optimization platform, backed by Everstone Capital (already a VWO investor), according to TechCrunch’s reporting and the two companies’ joint announcement. Market consolidation like this is exactly the kind of event that turns a “frozen” price table (the kind sometimes found in third-party comparison blogs) obsolete overnight, even while a brand keeps operating under its existing name for the time being.

None of this makes these tools impossible to budget for, only that the accurate number exists after a sales conversation with your real traffic on the table, not inside a generic table.

AB Tasty and Convert.com: a real transparency exception

Not every enterprise tool treats pricing the same way. AB Tasty confirms, on its own pricing page, a fully custom-proposal posture: the language used there is that a customer’s business does not fit a standard model, so pricing does not either. There is no self-serve checkout, and instead of a standard trial the company offers a guided proof of concept. In other words, even with the announced VWO merger, AB Tasty’s commercial model remained, as of this article’s check, individual quoting based on traffic, domains, modules, and implementation scope, with no public table.

Convert.com breaks that pattern. Checked directly on its official pricing page in July 2026, Convert.com publishes real monthly figures for two entry plans, each tied to a range of monthly tested users, with a discount for annual billing, and the Enterprise option remaining quote-only for larger volumes. It is the clearest exception inside the enterprise category: a buyer can build an initial estimate without booking a call first, something none of the other three enterprise suites (VWO, Optimizely, AB Tasty) or Kameleoon offers today. Our Convert vs VWO comparison covers this same transparency gap from the feature side.

Even so, treat any specific figure, including the ones Convert.com publishes, as a snapshot of the moment it was checked, not a constant. SaaS pricing pages change without advance notice, plans get renamed, and tested-user ranges get adjusted often; the safe habit is always reconfirming directly on the vendor’s official page before budgeting a contract.

GrowthBook, PostHog, and Statsig: usage-based and open-source alternatives

GrowthBook, PostHog, and Statsig belong to a different market category than the four enterprise suites above, and that changes the pricing logic at the root, not just the final number. All three either start from an open-source core or price primarily by measurable usage rather than a negotiated contract.

GrowthBook’s core is open-source under an MIT-style license: the software itself can be inspected, modified, and self-hosted with no license fee, trading that cost for your own server, database, and engineering time. According to GrowthBook’s own pricing page, its managed Cloud tier adds a Starter plan that is free for up to three users, and a paid Cloud Pro plan at 40 dollars per seat per month for up to 50 users, with CUPED and sequential testing gated to the Cloud Pro tier and above (Bayesian reporting itself is a core statistical option available on every tier, including the free Starter plan and the open-source self-hosted core). We cover this self-hosted-versus-managed tradeoff in depth in GrowthBook vs Statsig.

PostHog runs on a different metering unit: according to PostHog’s pricing page, its hosted free tier includes 1 million events per month at no cost for product analytics, with no credit card required to sign up. Experiments are billed together with feature flags on their own separate meter, which carries its own free allowance of 1 million requests per month, not the analytics-events allowance. A self-host option also exists for teams that want to keep data on their own infrastructure instead.

Statsig prices by metered events rather than seats or tested traffic. According to Statsig’s own pricing page, its free Developer tier includes 2 million metered events per month with unlimited seats, 50,000 session replays per month, and a full year of analytics retention, all hosted with no self-hosting required. Its Pro tier is listed at 150 dollars per month for 5 million included events (additional events billed at 0.05 dollars per 1,000), still with unlimited seats, and an Enterprise tier priced by volume and contract term for larger organizations.

Running the self-hosted core is never literally free: it only moves the cost from software to operations (server, database, and the engineering time to deploy, patch, and handle an incident when something breaks). Teams without spare engineering capacity tend to feel that hidden cost sooner than expected, a point covered in more detail in the guide to free A/B testing tools.

Cost dimension versus what each category typically charges

The table below does not list dollar or real figures (most of the category simply does not publish those stably), but it summarizes what each category tends to do with each cost dimension, which is more useful for comparing quotes than a single, possibly stale number.

Dimension	Enterprise suites (VWO, Optimizely, AB Tasty, Kameleoon)	Convert.com	Usage-based / open-source (GrowthBook, PostHog, Statsig)
Tested traffic (MTU)	Quoted by contract-agreed range, negotiable by volume	Published price by monthly tested user range on entry plans	No charge on self-hosted; hosted tiers meter events or MTU above a free allowance
Domains and projects	Usually included in contract, negotiable count	Set by the contracted plan	Unlimited on self-hosted; hosted tiers may limit by workspace or seat
Modules	A/B testing, personalization, heatmaps, and feature flags sold as a bundle	Mostly A/B testing and personalization	A/B testing and feature flags; Statsig and PostHog also bundle product analytics
Self-hosted vs managed	Always fully managed SaaS	Always fully managed SaaS	Real choice: free self-hosted (infrastructure cost) or paid managed cloud
Contract length	Annual or multi-year common, discounted by volume and commitment	Monthly or annual, no long-term commitment required	No contract on self-hosted; hosted tiers are typically monthly

How to request a quote efficiently

Since none of the enterprise suites publish a reliable table, the work of discovering the real price falls on the buyer, on a call or contact form. Arriving prepared with the right answers cuts down the number of rounds before you see a real number:

• How many monthly tested users (MTU) does your real traffic actually generate? Not a site’s total traffic, the volume that actually enters an active experiment. If you cannot estimate that yet, the calculator below helps translate “I want to detect an X percent improvement” into a real number of visitors per test, which is the starting point for estimating expected monthly MTU.
• How many domains or projects will share the same account? An agency running 8 different brands needs to say so on the first call, since it changes the tier or add-on quoted.
• Which modules beyond pure A/B testing do you actually need? Personalization, heatmaps, session replay, and feature flags cost more once activated, so list only what your operation will genuinely use, not what “might be useful someday.”
• What contract length does the vendor require for the quoted price? Monthly almost always costs more per unit than annual, and some vendors only unlock their most competitive price with a multi-year commitment; ask explicitly about both before comparing proposals.
• What happens if you exceed the contracted volume mid-period? Automatic overage billing, service degradation, or just a renegotiation notice varies by vendor, and that clause matters as much as the base price when comparing risk.

Current conversion rate%Minimum detectable effectrelative (%)absolute (pp)Confidence909599%Power8090%Visitors per week (total)Testtwo-sidedone-sided

-Visitors per variation

-Total (2 variations)

-Estimated duration

Two-proportion normal approximation, 2 variations (50/50). Tweak the inputs and watch it update live.

Running the calculator above with your real traffic, baseline conversion rate, and the minimum effect you want to detect gives you the sample size per test, which combined with how many tests you plan to run per month becomes the basis of your MTU estimate. For the full walkthrough of reaching that number, see the guide to what A/B testing is and how sample size works.

Common mistakes when comparing pricing across tools

Mistake	Why it backfires
Bringing a number seen in a third-party comparison blog into a sales call	Aggregator tables go stale fast, and most do not distinguish MTU from a site’s total traffic
Comparing VWO or Optimizely’s quote directly against Convert.com’s published entry price	These are different contract categories; Convert’s entry-level figure does not predict an enterprise volume contract’s cost
Treating a self-hosted GrowthBook or PostHog deployment as “free” without counting infrastructure	The software carries no license fee, but the server, database, and engineering time still cost something
Requesting a quote without knowing your own MTU	The vendor will ask anyway; showing up without that answer only adds a round of back-and-forth
Assuming a brand’s current pricing will stay stable in a consolidating market	Mergers like VWO and AB Tasty’s in 2026 change bundles and price positioning without advance notice

Automate This in Donnu

You just saw how much work it takes to get an honest number out of an enterprise A/B testing quote: estimating MTU, listing the modules you actually need, negotiating contract length, and reading the fine print on overage. If your operation does not need the depth (or the enterprise contract) of VWO, Optimizely, AB Tasty, or Kameleoon, and you would rather skip the opaque quoting process altogether, it is worth considering Donnu as an additional, neutral option: a lighter, more predictable starting price with no call required to learn it, built for smaller marketing or agency operations that need honest statistical rigor without a full suite’s complexity.

Start a free 14-day trial and see the real price for your traffic volume before deciding. To understand the statistical foundation behind any A/B testing tool first, see what A/B testing is and the complete guide to CRO tools compared.

Read also: VWO vs Optimizely: the full decision criteria | AB Tasty vs Kameleoon: testing and personalization compared | GrowthBook vs Statsig: open-source versus managed testing

Leia em português: preço de ferramentas de teste A/B.

References

• VWO. Pricing & Plans. vwo.com/pricing, checked July 2026.
• Optimizely. Plans and pricing for Optimizely products. optimizely.com/plans, checked July 2026.
• AB Tasty. Pricing. abtasty.com/pricing, checked July 2026.
• Kameleoon. Plans, including PBX pricing. kameleoon.com/plans, checked July 2026.
• Convert.com. Convert Experiences Pricing. convert.com/pricing, checked July 2026.
• GrowthBook. Predictable pricing, free tiers, enterprise plans. growthbook.io/pricing.
• PostHog. Pricing, including the free tier allowances. posthog.com/pricing.
• Statsig. Pricing. statsig.com/pricing.
• TechCrunch. Everstone combines Wingify and AB Tasty for $100M+ digital experience optimization platform. techcrunch.com.