The 12 Best CRO Tools in 2026 (And What Each One Actually Costs)

Most CRO tool roundups have two problems. They list products that no longer exist, and they skip the prices.

This one does neither. Every price below came off the live pricing page in August 2026. Where a company no longer publishes a price, that is written down too, because it tells you something about what buying from them will be like.

One note before the list. The single biggest change in this category over the last two years is not a feature. It is that the serious testing platforms stopped showing prices at all. Four of the twelve tools here will not tell you what they cost until you talk to someone.

The tools

Ordered by what it costs to actually do CRO with them, cheapest first, with the four that publish no price at the end. Not by preference, and not by who we are. Where a tool's entry plan is cheaper than its testing plan, the testing price is the one used, because that is the price that matters here.

1. Microsoft Clarity - free

Heatmaps and session recordings, free with no paid tier and no usage limit. Their own site says "free forever" and reports 2M+ sites using it. There is no pricing page because there is nothing to price.

If you have never looked at a heatmap of your own site, start here. It costs nothing and it will change what you think is worth testing.

Does not: build or test pages. Best for: anyone, immediately.

2. Google Analytics 4 - free

Not a CRO tool in the testing sense, but it is where most teams find the pages worth optimizing in the first place. Worth naming because the honest answer to "which tool do I start with" is usually the one you already have.

Does not: test anything. Best for: finding which pages deserve attention before you spend.

3. PostHog - free tier, then usage-based

Open-source product analytics with A/B testing, feature flags and session replay in one tool. The free tier is genuinely generous: unlimited tracked users, unlimited team members, and a monthly free volume on each product before usage-based pricing starts.

The only tool on this list that will run a test for you at no cost. More engineering-flavoured than the rest, which is a fit question rather than a quality one.

Best for: teams where the same people ship the product and optimize the funnel.

4. Crazy Egg - testing from $29

Heatmaps, scroll maps, session recordings and testing, added to a site you already have via a snippet. Four tiers: $29, $99, $249 and $599, billed annually.

The cheapest paid way to test something without rebuilding it.

Does not: build pages. Best for: improving an existing site you do not want to touch structurally.

5. Contentsquare, formerly Hotjar - from $49

Hotjar is now part of Contentsquare and the Hotjar name is being folded into it. Published entry plans are $49 and $99 a month for heatmaps and session replay.

Does not: test. Best for: seeing where attention goes before deciding what to test.

6. Leadpages - testing from $99

We build this, so treat the description as ours and the prices as checkable.

Plans start at $10, but that tier does not include testing, so $99 is the honest number for this list. At $99 you get A/B testing on every paid plan with no traffic caps, so a test never costs you extra visitors. Smart Traffic, which routes each visitor to the variant most likely to convert them, and heatmaps are at $199. Full auto-optimization is at $399.

Best for: teams who want to build the page and test it in the same place, without a specialist.

7. Unbounce - testing from $149

A landing page builder with a strong testing layer. Plans run $29, $99, $149 and $249, plus a custom Concierge tier.

Worth knowing exactly where the testing starts: not on the $99 plan. A/B testing begins on Experiment at $149, which also brings dynamic text replacement. Their AI traffic optimization is on Optimize at $249.

Every plan is capped on traffic: 500 visitors on Starter, 20,000 on Build, 30,000 on Experiment, 50,000 on Optimize.

Best for: teams already committed to Unbounce's editor and ready to pay the testing tier.

8. Instapage - testing from $199

Also a builder with testing above it. Create is $99 and Optimize is $199 monthly, or $79 and $159 billed annually, with a custom Convert tier.

Server-side A/B testing, experimentation history and dynamic text replacement start on Optimize at $199. Heatmaps are on the custom-priced Convert plan only. Caps are 15,000 unique monthly visitors on Create and 30,000 on Optimize.

Best for: ad teams running many page variants per campaign, with budget for the $199 tier.

9. VWO - no published price

One of the most capable testing platforms in the category, and has been for years. As of August 2026 its pricing page lists three plans, Growth, Pro and Enterprise, and every one of them has the same call to action: Schedule a Demo. There is no dollar figure on the page.

There is a free tier you can explore without talking to anyone, which is the fastest way to see the product. Multivariate testing starts on Pro.

Best for: larger teams with a dedicated optimization function and a procurement process.

10. Optimizely - no published price

The enterprise end of experimentation. Deep capability, no public pricing, sales-led from the first conversation.

Best for: enterprises running experimentation as a discipline across multiple products.

11. AB Tasty - no published price

Now part of the same group as VWO following their merger under the Wingify name. Also sales-led, also no published price, and no self-serve free trial.

Best for: mid-market and enterprise teams, usually alongside an agency.

12. FullStory - no published price

Session replay and behavioural analytics at enterprise scale, with a free tier covering 30,000 sessions a month and no published paid pricing.

Does not: test. Best for: product and engineering teams as much as marketing.

The best free CRO tools

You can do a surprising amount before paying anything, and this is the honest starting point for most teams.

Free forever, or close to it: Google Analytics 4 for finding the problem pages. Microsoft Clarity for heatmaps and session recordings. FullStory's free tier covers 30,000 sessions a month. PostHog has a free tier that includes testing.

Free to try, then paid: VWO has a free option you can explore without a sales conversation, which is the fastest way to see a serious testing platform without starting a procurement cycle.

The honest limit of free: free tools will show you what is happening and often where. What they will not do is let you build a variant of a page and split traffic to it. That is where paid starts, and it starts at $99 a month.

The best CRO tools for e-commerce

E-commerce has a different shape to lead generation: the conversion event is a purchase, the funnel has a cart and a checkout, and most of the leaks are in the last two steps.

For the store itself, Baymard's checkout research is the reference most serious e-commerce teams work from, and Crazy Egg or Clarity will show you where people abandon.

For the pages that feed the store - the ad landing pages, the campaign pages, the offer pages - you want something that builds and tests pages independently of your store platform, so you are not waiting on a theme developer to run an experiment.

What to avoid: buying an enterprise experimentation platform before you have the traffic to reach significance on a test. Below roughly 1,000 conversions a month, most tests will not call a winner in a useful timeframe, and you are better served fixing the obvious checkout leaks first.

The AI tools for conversion rate optimization

Nearly every tool here now claims AI. It is worth separating what the AI actually does, because there are two very different jobs hiding under one word.

AI that generates. Writes your page, your variants, your ad copy. Common, useful, and increasingly table stakes. Instapage, Unbounce and Leadpages all do a version of this.

AI that decides. Watches what converts and routes traffic accordingly, rather than waiting for you to read a report. Unbounce's AI traffic optimization on their $249 plan, Leadpages' Smart Traffic at $199, VWO's Wandz layer, and Fibr's whole product are in this category.

The second one is the harder engineering problem and the one worth paying for. When a tool says AI, the question to ask is whether it makes the thing or decides between things.

What no AI tool does yet: tell you what to test. That still comes from looking at where people drop off.

How big does your test need to be?

Before you ship an experiment, work out how much traffic it needs to reach a result you can trust. A test called too early is worse than no test, because it feels like evidence.

What this list tells you

Four of the twelve will not show you a price. That is not a criticism of the products, which are good. It is information about the buying process: expect a call, a scoping conversation and a quote based on your traffic.

The rest publish, and the pattern is consistent once you sort by it. Watching behaviour is free to $49. Testing without building starts at $29. Testing the pages you also build starts at $99 with us, $149 at Unbounce and $199 at Instapage - and at the platforms that do not publish, wherever the quote lands.

If you are assembling a stack this quarter, the useful exercise is not comparing feature lists. It is working out how much you will pay before you are allowed to run your first test.

Frequently asked questions

What are CRO tools? Software that helps you find out why visitors do not convert and then improve it. The category covers three jobs: seeing what people do (heatmaps and session replay), testing changes (A/B and multivariate testing), and building the pages you test.

How much do CRO tools cost? Published prices in August 2026 run from $29 a month for behaviour tools to $249 a month for the top self-serve testing tiers. The enterprise testing platforms, VWO, Optimizely and AB Tasty, do not publish pricing.

What happened to Google Optimize? Google retired it in September 2023. Any list that still recommends it has not been checked since then.

What is the cheapest way to start A/B testing? If you need to build and test pages, testing starts at $99 a month with Leadpages, $149 with Unbounce and $199 with Instapage. If you already have a site and only need to test it, snippet tools start around $29.

Do I need a CRO specialist? For enterprise platforms, effectively yes. For the self-serve tools, no. Start with one test on your highest-traffic page and change one thing.

What are the best free CRO tools? Google Analytics 4 for finding problem pages, Microsoft Clarity for heatmaps and session recordings, and FullStory's free tier for session replay up to 30,000 sessions a month. VWO offers a free option for exploring a full testing platform. Free tools show you what is happening; building and testing page variants starts at $99 a month.

What are the best CRO tools for e-commerce? Behaviour tools like Crazy Egg or Clarity for finding checkout leaks, plus a page platform that builds and tests campaign pages independently of your store theme. Below roughly 1,000 conversions a month, fix obvious checkout problems before buying an experimentation platform, because tests will not reach significance quickly enough to be useful.

What are the best AI tools for conversion rate optimization? Separate AI that generates from AI that decides. Generation is common: Instapage, Unbounce and Leadpages all produce pages and variants. Deciding is rarer and more valuable: Leadpages' Smart Traffic at $199, Unbounce's AI traffic optimization at $249, VWO's Wandz layer, and Fibr's platform all route traffic based on what is converting.

Which CRO tools offer real-time optimization? The ones that route traffic automatically rather than reporting after the fact: Smart Traffic on Leadpages, AI traffic optimization on Unbounce's Optimize plan, and the agentic layers from VWO and Fibr.

Which tools support multivariate testing? Multivariate testing, which tests several variables at once rather than one, is generally an upper-tier feature. VWO includes it from its Pro plan. Most landing page platforms offer straightforward A/B testing rather than full multivariate.