A/B tests on SaaS and software websites beat their control 9% of the time, against 10.1% across all industries, according to Mida's benchmark of tests concluded between August 2023 and August 2026. That puts SaaS eighth of eleven measurable industries, and it is the closest thing to a B2B SaaS win rate that any testing platform publishes.
This page collects the numbers a B2B SaaS team needs before it budgets for testing: how often tests win, how much traffic one needs at SaaS conversion rates, how large programmes are, and what the tools and agencies cost. Every figure is dated and the vendor sources are labelled as such.
There is no B2B SaaS win rate, only a SaaS one
No experimentation platform publishes a win rate for B2B SaaS on its own. Mida comes closest: it classifies every tested domain by industry from the site itself, not from what the customer said at sign-up, and publishes win rates for each. Its methodology page notes that self-reported and verified industries disagree 65% of the time, which is a reason to distrust any benchmark built on self-reported industry.
| Source | Dataset | What counts as a win | Win rate |
|---|---|---|---|
| Mida, SaaS and software | 500+ variants, Aug 2023 to Aug 2026 | Beat control on the primary goal at 95% chance to beat or p ≤ 0.05 | 9% |
| Mida, all industries | Same period, 11 industries | Same | 10.1% |
| Optimizely | 173,000 experiments, all industries | Significant improvement on the primary metric | 12% |
| VWO (Wingify) | 193,000 experiments, 38,000 sites, 2024 to 2026 | Not defined on the public page | CTA copy about 10%; click-through tests about 10% |
| Microsoft, as cited by Mida | Microsoft's own platform | Idea improved the metric it was designed to improve | About one third |
Source: Mida A/B testing benchmarks and methodology, 2026; Optimizely; Wingify experimentation benchmark 2026
The figures agree more than they first appear to. Somewhere around one test in ten produces a significant win on a typical site, and the Microsoft figure comes from a dedicated experimentation team testing at a scale no B2B SaaS marketing site reaches. Our A/B testing statistics post covers the cross-industry figures, including lift distributions, in more depth; this page stays with what changes for a SaaS site.
What wins on SaaS sites, and what does not
Mida breaks the SaaS figure down by the kind of change tested.
| Change type | Tests | Win rate |
|---|---|---|
| Styling | 100+ | 13% |
| Split URL (whole-page redesign) | 100+ | 12% |
| Body copy | 33 | 9% |
| Headline | 100+ | 6% |
| CTA copy | 50+ | 2% |
Source: Mida, A/B testing benchmarks for SaaS and software, updated 2026-08-23
The order is the reverse of what most testing backlogs contain. Button wording and headline swaps are the cheapest tests to build and the least likely to win: 2% and 6%. Whole-page alternatives and visual changes, which take a designer and a developer, win at least twice as often as headlines. On Mida's headline data for SaaS, 85% of headline tests showed no measurable difference and 9% lost outright.
Two further cuts matter for a B2B team. First, the goal. Tests measured against a proxy goal, such as a click or a scroll, won 13% of the time; tests measured against a business outcome won 9%. A programme that reports click-through wins will look better than one that reports pipeline, and will not be better. Second, the page. Mida's SaaS headline tests ran mostly on homepages (166), with few on landing pages (26) and lead capture pages (16), and 54% measured sign-up or registration. Very little of the published data comes from demo or pricing pages, which are where B2B SaaS revenue is decided.
How much traffic a B2B SaaS test needs
This is where B2B SaaS differs most from ecommerce. HockeyStack's December 2025 study of 31 million visitors across 80 B2B SaaS companies found a whole-site conversion rate of 1.1% per unique visitor, a pricing page conversion rate of 3.8% and a demo page rate of 5.5%. Our B2B SaaS website conversion rate statistics post explains where those definitions come from.
We calculated the visitors each variation needs at those three baselines, using the standard two-proportion formula at 95% confidence (two-sided) and 80% power.
| Baseline conversion rate | 10% lift | 20% lift | 30% lift | 50% lift |
|---|---|---|---|---|
| 1.1% (whole-site lead rate) | 148,111 | 38,769 | 18,003 | 7,037 |
| 3.8% (pricing page) | 41,644 | 10,885 | 5,047 | 1,967 |
| 5.5% (demo page) | 28,237 | 7,373 | 3,416 | 1,328 |
Source: Octo calculation, two-proportion test, alpha 0.05 two-sided, power 0.80; baselines from HockeyStack, 2025
Now apply HockeyStack's traffic shares. Pricing pages took 16.5% of site traffic in its sample and demo pages 0.91%. A B2B SaaS site with 30,000 unique visitors a month sends about 4,950 to the pricing page and about 270 to the demo page.
| Page and goal | Monthly visitors to the page | Lift to detect | Visitors needed, both variations | Time to finish |
|---|---|---|---|---|
| Whole site, any lead form | 30,000 | 20% | 77,538 | About 2.6 months |
| Pricing page, form submission | 4,950 | 20% | 21,770 | About 4.4 months |
| Pricing page, form submission | 4,950 | 30% | 10,094 | About 2 months |
| Demo page, form submission | 270 | 30% | 6,832 | About 25 months |
Source: Octo calculation from the table above and HockeyStack traffic shares, 2025
The demo page cannot be A/B tested on a site of that size. That matches what Convert sees on its own platform: about one experiment in ten runs on fewer than 1,000 visitors, the most common size is 10,000 to 50,000 visitors (37% of experiments), and about 9% reach 100,000. A typical B2B SaaS site has enough traffic for a small number of tests a year on its two or three busiest pages, aimed at lifts of 20% or more.
How big B2B SaaS testing programmes are
There is no SaaS cut of programme size, so the figures below cover all industries. Technology and SaaS companies made up 23% of respondents to Speero's 2025 Experimentation Maturity Programme Report, the second largest group after retail and ecommerce at 27%, as cited by Convert.
| Measure | Figure | Source |
|---|---|---|
| Experiments a year, median company | 34 | Optimizely |
| Experiments a year, top 10% | About 200 | Optimizely |
| Experiments a year, top 3% | More than 500 | Optimizely |
| Companies with no dedicated experimentation staff | 10% | Speero 2025, cited by Convert |
| Companies with exactly one person running the programme | 25% | Speero 2025, cited by Convert |
| Companies with no QA process before launch | 52% | Speero 2025, cited by Convert |
| Companies with no clear prioritisation framework | 58% | Speero 2025, cited by Convert |
| Companies at strategic or transformative maturity | 54%, up from 35% in 2021 | Speero 2025, cited by Convert |
Sources: Optimizely, top 10 takeaways from 173,000 experiments; Convert, A/B testing statistics, updated 2026-05-24
The median of 34 experiments a year is roughly three a month. Set that against the traffic table: a B2B SaaS site of 30,000 monthly visitors could finish perhaps five or six well-powered page tests a year. The rest of a 34-test programme on a site that size would be underpowered, or would run in the product, where sign-up and activation volumes are higher.
Convert's 2025 agency report points the same way. Of 237 experimentation agencies, 60% say experimentation is moving from marketing towards product teams, and 65% report clients bringing more testing in-house. For SaaS, the in-product experiments in onboarding and upgrade flows are where the volume is.
Adoption is still thin overall. BuiltWith data cited by Convert finds an experimentation tool on about 0.2% of websites, but on 32% of the top 10,000 sites and about 11.5% of the top million.
What B2B SaaS testing costs
Budgets split into three lines: the tool, the people who design and build tests, and, for many SaaS teams, an agency.
| Item | Price | Source and basis |
|---|---|---|
| Convert Growth plan | $299 a month billed annually ($3,588 a year), or $399 monthly; 100,000 tested users a month | Convert pricing page, read 2026-10-28 |
| Convert Pro plan | $420 a month billed annually ($5,040 a year), or $599 monthly | Convert pricing page |
| VWO (Wingify) Web Experimentation | No public price; Growth, Pro and Enterprise plans by demo | Wingify pricing page |
| Optimizely, median contract | $81,447 a year; range $26,460 to $261,075 across 126 purchases | Vendr, updated February 2026 |
| CRO agency retainer, small agency | $2,000 to $5,000 a month | Aureate Labs, an ecommerce agency, 2026 |
| CRO agency retainer, mid-tier | $6,000 to $15,000 a month | Aureate Labs |
| CRO audit, one-off | $2,000 to $5,000 | Aureate Labs |
Sources: the publishers named in column three
The tool is rarely the large line. A B2B SaaS site under 100,000 monthly visitors fits in Convert's entry tier, while Optimizely's median contract is more than twenty times the cost of Convert's entry plan. The agency figures are vendor-sourced, from an agency that sells CRO, and they are the only published ranges we could find; treat them as a quote range, not a survey.
The people line is harder to price. Convert's agency report finds that about 60% of agencies manage two or fewer tests per person per month, and nearly a third run ten or fewer experiments a month in total. A test that changes layout, which on Mida's data wins more often than copy, needs design and development time on top of the tool.
A five-page test order for a B2B SaaS site
The order below follows the traffic arithmetic, not the backlog. Each page is ranked by whether it can support a powered test, with the goal measured as close to pipeline as the volume allows.
1. Pricing page. It carries 16.5% of traffic and converts 3.8% in HockeyStack's data, so it is the only high-intent page most B2B SaaS sites can test properly. Test structural changes, such as plan layout or a published price against "talk to sales", aimed at a 30% lift, and measure submissions and then pipeline. HockeyStack found transparent pricing converted 2.8% against 4.6% for non-transparent, but turned 17.5% of submissions into pipeline against 10.3%.
2. Homepage. It feeds every other page and has the most traffic. Make the goal the share of visitors who reach pricing or the product pages, not a click on the hero button: a proxy goal, but one directly upstream of revenue. Test whole-page alternatives rather than headlines, which win 6% of the time on SaaS sites.
3. Paid landing pages. Traffic is bought, so it can be split evenly and scaled to finish a test. SaaS landing pages convert at a 3.8% median on Unbounce's data. Our B2B SaaS landing page statistics post covers form length and message match, the two things most worth testing there.
4. Sign-up or trial flow. For product-led SaaS, sign-up volume is higher than demo volume and the test can run in the product. This is where the programme volume of the median company is realistic.
5. Demo request page. At under 1% of traffic it cannot support a classic A/B test on most sites. Change it on evidence from session recordings, form analytics and sales feedback, and compare before and after. Our B2B SaaS demo request statistics post has the form benchmarks to compare against.
What this means for a B2B SaaS testing budget
Plan for about one win in ten tests, a handful of powered page tests a year on a site of 30,000 monthly visitors, and lifts that are real but smaller than the recorded winners suggest. That favours spending on fewer, larger tests on the pricing page and homepage, measured on pipeline where volume allows, over a long list of copy tweaks. It also means the tool line can stay small until traffic or product testing justifies more. We build B2B SaaS marketing sites in Webflow, and we would set up the tracking for these goals at build time so the first test can start on launch day. Our marketing website service explains how we scope that work.
