Optimizely Cost: What the Plans Page Publishes

Optimizely's pricing page carries no prices. Fetched twice on 1 September 2026, once over plain HTTP and once through a headless browser to rule out client side rendering, the page at optimizely.com/plans returned zero currency figures by either route. What it returns instead is a form and one sentence: "Every Optimizely plan is individually packaged. Tell us a bit about your digital needs, and we'll create a plan together."

That is the answer to the cost question, and it is worth taking seriously rather than treating as a gap to be filled by whatever number a roundup happens to carry.

Below is what the vendor's own pages do disclose, where the figures circulating on other sites come from, what the alternatives with published ladders charge, and the test that decides whether any of these numbers matter to a B2B team.

Why there is no rate card, read from the page itself

The reason is structural and the plans page shows it. Optimizely is not one product with tiers. The page presents three headline products, Agentic CMS, Agentic Experimentation and an Agent Platform, and then a block titled "Even more tools for marketers" listing seven more: Content Marketing Platform, Analytics, Personalization, Commerce, Asset Management, Data Platform and Feature Management.

Ten separately marketed products, typically bought in combination, do not reduce to a ladder. A price would have to name which products, at what traffic, on which deployment, for how many users.

That list is also the short answer to what Optimizely is used for, worth stating plainly because the question is usually asked apart from the price and usually answered by the vendor's own implementation partners.

The experimentation product is the one adjacent to most B2B marketing work, and its own bullet list on that page names "A/B, multi-page, feature and server-side testing", a "Proprietary Stats Engine for results you can trust", "Real-time personalization and audience targeting" and "Warehouse-native analytics to measure impact". Those are capability claims from the vendor rather than measurements, and they describe an enterprise experimentation platform rather than a testing widget.

The customers page points the same way. Its own filter list runs across B2B services, financial services, healthcare, insurance, manufacturing, education and charities, and its featured stories are a national mint, a Gulf retail bank and a European furniture group consolidating seventeen brands off a legacy CMS. This is a platform sold to organisations with a digital estate, and the shape of the buyer explains the shape of the pricing.

Where the numbers on other pages come from

Search the cost question and the results are mostly published by companies that sell experimentation software, plus one procurement marketplace. The marketplace is the only one of them publishing a figure with a stated method behind it, and it is worth quoting precisely rather than paraphrasing.

Vendr's Optimizely marketplace listing states that the "Median buyer pays $81,447 per year", with a low of $26,460 and a high of $261,075, and describes that as "Based on data from 126 purchases, with buyers saving 13% on average". The same page's summary card reports "$81,447 Avg Contract Value" and "100 Deals handled".

Two things about that figure deserve to travel with it. It is a contract value from a procurement intermediary's own anonymised dataset rather than a rate Optimizely puts on a page, and the same listing says "Published list pricing is rarely the final number" while disagreeing with itself on the sample: 126 purchases in the body against 100 deals in the card. Neither number is unreasonable, and the gap between the low and the high is a factor of ten, which is the more useful signal. A spread that wide says the figure is a distribution across ten products and many deployment sizes rather than a price for a thing.

Vendr's own description of the pricing drivers matches what the vendor's pages imply: product selection, traffic and usage volume, deployment model, and feature tier.

Quoted onlyOptimizely, VWO

• Optimizely publishes a form and no figures
• Ten separately marketed products
• VWO's pricing page rendered no currency figures either
• Quotas per monthly active user instead of rates

Published ladderStatsig, GrowthBook

• Statsig meters events, with an allowance inside the paid tier
• Then a published overage rate per thousand events
• GrowthBook meters seats, with a user ceiling on the tier
• Free tiers on both, with unlimited experiments

Usage firstPostHog

• A monthly free allowance on every product
• 1M analytics events and 5K session recordings free
• Experiments billed with feature flags
• Pay as you go above the allowance

What the alternatives publish instead

What the alternatives publish instead is the comparison worth making, because a published ladder is what lets a team budget without a sales cycle. A feature-by-feature table separates these products badly; what each vendor is willing to put a number on separates them well.

Statsig publishes a ladder. Its Developer tier is free with 2 million events a month, unlimited flag and config checks, 50,000 session replays a month, unlimited seats and one year of analytics retention. Its Pro card is $150 a month, described on that page as "5M events included, then $0.05 per 1K events", plus 100,000 session replays a month. Enterprise is quoted, and is where warehouse native deployment, SSO and role based access sit.

GrowthBook prices per seat. Its Starter tier is free for up to three users on one project, with unlimited feature flags, unlimited experiments and unlimited traffic. GrowthBook's Pro card is $40 per seat per month for up to fifty users and three projects, adding a visual editor, multi arm bandits, safe rollouts and a power calculator. Enterprise is quoted, with the page offering "Pricing tailored to your scale and licensing requirements". Both cloud and self hosted deployments are offered.

PostHog leads with the free allowance. Its pricing page opens with "97% of companies use PostHog for free" and publishes a monthly allowance per product that renews whether or not a card is on file: 1 million analytics events, 5,000 session recordings, 1 million feature flag requests, 1,500 survey responses, with experiments billed alongside feature flags. Above that it is pay as you go, and the page states that usage stops at the free tier limits without a card, so an account cannot be charged by surprise.

VWO answers the same way Optimizely does. Its pricing page returned no currency figures to the fetch made on 1 September 2026, and that page varies quotas per monthly active user instead of carrying rates, an approach set out alongside the other meters in landing page optimization tools.

Before the pricing call

• ✓Yes: You can name which of the ten products you would actually deploy
• ✓Yes: You have your monthly page views and conversions to hand, counted rather than estimated
• ✓Yes: You know whether a published ladder elsewhere would meet the same requirement
• ✓Yes: You have decided what a test needs to be conclusive at your volume
• ✕No: You are treating a third-party contract-value figure as a budget number
• ✕No: You are comparing an enterprise DXP with a single-purpose testing tool on price alone

The test that decides whether the price matters

There is a question that comes before the quote, and for most B2B teams it settles the purchase without a conversation.

An experimentation platform converts traffic into a decision, and it needs a sample to do it. A campaign landing page receiving a few hundred visitors a month, converting at single or low double digit numbers, cannot separate a real effect from noise however good the platform is. The arithmetic behind that threshold is worked through in landing page optimization tools, and the same sample logic applied to email is in cold email A/B testing benchmarks.

Two supporting reads make the point concrete. What the published conversion benchmarks actually counted, and why none of them describes an outbound fed page, is in average landing page conversion rate. And the reason a rate computed over a few dozen records is dominated by chance is in conversion rate.

Below that threshold the useful spend is on observation and on asking real buyers, both of which work at low sample, and on the flow of visitors rather than on the landing page itself. Whether a campaign page or the demand reaching it is the constraint is the prior question, worked through in landing page lead generation. Above it, an experimentation platform becomes the correct centre of the stack, and then the question of whether a quoted enterprise suite or a published ladder fits is a real one.

1. Step 1Count the sample

   Monthly visitors and monthly conversions on the pages you would test, counted rather than estimated

2. Step 2Name the products

   Which of the ten Optimizely markets you would deploy, since the price is per product

3. Step 3Price the published ladders first

   Statsig, GrowthBook and PostHog all publish figures you can budget against without a call

4. Step 4Take the quote last

   Arrive with volume, product list and an alternative that has a number on it

What we are not saying

Nothing above is a claim that Optimizely is overpriced, and nothing here is a quote. The vendor puts no list price on the page that would carry one, saying instead that "Every Optimizely plan is individually packaged", so no independent page can tell you what yours would be, and any article naming a single figure is repeating a third party's contract data or another article.

The honest position on a quote only vendor has three parts. Say what the vendor's own pages disclose, which here is a product list and a packaging statement. Name any external figure as somebody else's measurement, with its method attached. And be clear about what would have to be true for the purchase to make sense at all, which for an experimentation platform is a traffic volume rather than a budget.

That last part is where most of the value sits, and it is also the part a vendor comparison page has no incentive to write.

The short version

Optimizely publishes no prices on the page that would carry them. That plans page, read twice on 1 September 2026, returns a form and the line "Every Optimizely plan is individually packaged", and the reason is that it markets ten products that are bought in combination.

The one external figure with a stated method is Vendr's marketplace listing, whose line is "Median buyer pays $81,447 per year", against a range of $26,460 to $261,075, on a page that names 126 purchases in one place and 100 deals in another. Treat it as a distribution across products and deployment sizes rather than as a price.

Alternatives carrying a rate card on the page include Statsig at $150 a month with 5 million events, GrowthBook at $40 per seat per month, and PostHog with a monthly free allowance and pay as you go above it. VWO, like Optimizely, shows none on the page we fetched.

Before any of that matters, count the conversions your pages produce in a month. Below the sample a test needs, the platform choice is not the decision in front of you. If the shortage is qualified traffic rather than tooling, see what one campaign produces against your own market.

Vendor statements and figures above were read from each vendor's own pages on 1 September 2026, except the contract value figures, which are Vendr's own marketplace data and are attributed to it. Verify current terms with the vendor before relying on them.