Until this week, our Professional plan cost $599 a month. For that, you got one active experiment. One. You could build as many drafts as you liked, but only a single test could be running on your site at any moment. We are not going to dress that up: charging $7,188 a year for one concurrent A/B test was not defensible, and we are the ones who set the number.
As of today, Professional is $49 a month. Business is $149. Enterprise is $1,499. That is a 92% cut at the entry point, and it is permanent. There is no countdown timer on this page and no promotional banner, because this is not a sale. It is a correction.
What we got wrong
We priced against Optimizely instead of pricing against the value we actually deliver. When your competitor charges $50,000 a year and you charge $7,188, the deck slide practically writes itself: 86% cheaper than the enterprise incumbent. It is a comfortable story to tell internally, and it is the wrong question entirely.
The right question is not "what is the largest number that still looks small next to Optimizely?" It is "what does this product do for a business, and what share of that is fair to ask for?" Judged that way, one active experiment for $599 a month fails immediately. A marketing agency running tests for three clients would have needed our $999 Business plan and still only gotten three concurrent experiments. A solo founder validating a landing page would have paid $599 to run a single test that took two weeks to reach significance.
“A price anchored to a competitor is a price anchored to someone else's cost structure, someone else's customers, and someone else's problems.”
David S., Founder
The new numbers
- Free: $0. Build one site and one experiment, edit variants in the visual editor, and open a preview URL to see it live on your own site. Launching to real visitors requires a paid plan.
- Professional: $49/mo, or $39/mo billed annually ($468/year). One active experiment, 7-day data retention, full per-variant analytics with statistical significance, multi-page funnel goals, Slack notifications, and no Segmently branding.
- Business: $149/mo, or $119/mo billed annually ($1,428/year). Three active experiments, 30-day retention, advanced analytics with revenue-per-variant tracking and CSV export, experiment scheduling, GA4 and Mixpanel sync, webhooks, and live chat.
- Enterprise: $1,499/mo, or $1,199/mo billed annually ($14,388/year). Nine active experiments, unlimited sites and seats, 1-year retention, custom integrations, REST API, and a dedicated account manager.
Every one of those prices is published. There is no sales call to find out the number, no quote form, and no per-visitor metering that quietly reprices you after a good traffic month.
The rule we price by now
We introduced a constraint that governs every pricing decision from here: a subscription must always cost less than assembling the same capacity out of credits. If the workaround is cheaper than the plan, the plan is mispriced. It is that simple, and it was not true of our old pricing.
Credits are our pay-as-you-go mechanism. One extra active experiment costs 60 credits a month, which works out to roughly $79 at our smallest pack. Under the old pricing, a Professional subscriber paid $599 for a plan whose headline capacity could be bought outright for $79. That is backwards, and customers noticed before we did.
- Professional: $49/mo for one active experiment. The same slot bought with credits costs about $79/mo. The plan wins, and it also includes analytics, funnel goals and Slack.
- Business: $149/mo for three active experiments. Those three slots bought with credits cost about $237/mo. The plan wins by $88 before you count retention, scheduling or integrations.
- Enterprise: $1,499/mo. This one is deliberately not sold on slot arithmetic. If all you need is a few more experiments, credits are genuinely cheaper and we would rather say so than pretend otherwise. Enterprise exists for unlimited sites and seats, year-long retention, and custom integration work.
We now enforce that rule in our test suite. If someone on our team sets a subscription price above the credit-equivalent cost, the build fails. Pricing discipline that lives only in a document is pricing discipline that erodes.
What this means if you run a small business
Take a store doing $2 million a year at a 2% conversion rate. A single successful test that lifts conversion to 2.5% adds roughly $500,000 in annual revenue without a dollar of additional ad spend. Against that, $468 a year is a rounding error. But the old $7,188 was not a rounding error for a business at $200,000 in revenue, and those are exactly the businesses that have been locked out of experimentation entirely.
That exclusion is the actual problem in this market. Amazon runs thousands of experiments a year. Booking.com runs tens of thousands concurrently. The compounding advantage of a testing culture is enormous, and it has been available almost exclusively to companies that could absorb five-figure annual software bills. A $49 plan does not close that gap on its own, but it removes the excuse that the tooling is unaffordable.
The trade-off we made, stated plainly
We are not going to bury this. At the same time as the price cut, our Free plan became preview-only. Previously a free account could run one live experiment against real traffic. It now lets you build a site and an experiment, edit variants visually, and open a preview URL on your own site, but launching to real visitors requires a paid plan or a credit-purchased slot.
That is a genuine reduction, and it is the trade that funds the other side of the ledger. We would rather have a $49 entry point that a small business can actually afford than an indefinite free tier subsidised by enterprise pricing that no small business would ever reach. If you are on the Free plan today and were running an experiment, you will see a notice in your dashboard explaining exactly what changed. Your sites, experiments, variants and results are all intact.
Three things we fixed while we were in here
Rebuilding the pricing meant reading every line of billing code, and we found problems we should have caught sooner. We are listing them because you deserve to know what was wrong, not just what is new.
- 1Our smallest credit pack contained 50 credits while an experiment slot cost 60. You could buy the Starter pack and unlock nothing. Starter now contains 60 credits at the same $79 price, sized to cover exactly one slot.
- 2Add-on slots renewed at ten times the price they were sold for. A slot bought for 60 credits was billed at 600 credits on its next monthly renewal, a leftover from an older rate card that was never updated. Purchase and renewal now read from the same number, and a test enforces it.
- 3Our pricing page carried a "limited time offer, 50% off" banner attached to prices that never changed. A permanent discount is not a discount, it is a fake anchor. It is gone, along with every strikethrough "was" price on the site. What you see is the price.
How this compares
For context, using each vendor's published pricing: Optimizely starts around $50,000 a year and runs past $120,000 for larger deployments. VWO begins at roughly $199 a month for 10,000 monthly tracked users, and climbs as your traffic does. Convert starts at $299 a month billed annually. AB Tasty does not publish pricing at all; you book a call.
Segmently Professional at $468 a year is under 1% of Optimizely's starting price, about 80% below VWO's entry plan on an annual basis, and roughly an eighth of Convert's. And because we charge flat rates rather than metering monthly tracked users, a successful campaign that triples your traffic does not triple your bill.
What happens next
The new prices are live now. Annual billing saves 20% on every paid tier. Nobody needs to take any action: there is no migration, no proration to reconcile, and no grandfathering to opt into.
If we have this wrong, we would like to hear it. Pricing is a hypothesis like any other, and we run a company built on testing hypotheses rather than defending them. The difference between this and our last pricing page is that this time we did the arithmetic from the customer's side of the table first.