B2B SaaS Conversion Rate Benchmarks: Why 3.8% and 1.1% Are Both Correct

Unbounce says the median SaaS landing page converts at 3.8%. First Page Sage says B2B SaaS converts at 1.1%. Both numbers are published, both are widely quoted as the conversion rate benchmark for B2B SaaS, and they differ by a factor of more than 3. Neither is wrong. They divide different things by different things, and almost nobody who repeats them says so.

The short answer. There is no single conversion rate benchmark for B2B SaaS. Unbounce's 2024 median for SaaS landing pages is 3.8%, measured per page. First Page Sage puts B2B SaaS visitor-to-lead at 1.1%, measured across client sites. Use the one whose numerator, denominator and traffic source match your own funnel, and state which one you used.

Why conversion rate benchmarks disagree by a factor of three

A conversion rate is a fraction. The top of the fraction is whatever action the publisher decided to count. The bottom is whichever population of visitors the publisher decided to count it against. Change either and the number moves, even when nothing about buyer behaviour has changed.

The benchmarks in this post differ on four separate axes. The action counted ranges from a click to the next page, through a form fill, to a first paid month. The denominator ranges from visitors to a single landing page to unique visitors to a whole website. The statistic is sometimes a median, sometimes an average, and sometimes unstated. The sample is always one company's customers or clients.

Any one of those four differences is enough to make two figures incomparable. Most published pairs differ on at least three. That is why a practitioner can find a credible-looking B2B SaaS conversion figure anywhere between well under 1% and well over 10%, and why picking one without reading the methodology is closer to choosing a target than looking one up.

The same problem shows up wherever a single headline number travels without its sample. It is the core of the methodology gap in AI ROI surveys, and the fix in both cases is the same: put the definition next to the number every time it is repeated.

The B2B conversion rate benchmark table: dataset, date and definition

This is the table the topic asked for. Each row records what the publisher actually wrote on the page we opened, not what aggregator posts say about it. Where a field is blank on the source page, the table says so rather than filling it in.

All rows checked against the publisher's own page in October 2026. "Not stated" means the page does not say. It does not mean the publisher lacks the information.

Unbounce's SaaS conversion rate is a median of landing pages, not of businesses

Unbounce's 2024 report states that the median conversion rate for SaaS landing pages is 3.8%, 42% lower than the 6.6% baseline for all industries. Within SaaS, hardware pages convert best at 4.1% and data and infrastructure pages lowest at 3.3%.

The unit of analysis is the page. A company running 40 landing pages contributes 40 observations, and a company running 1 contributes 1. The median tells you about a typical page built on Unbounce, which is a narrower thing than a typical SaaS business.

What counts as a conversion on an Unbounce page

The methodology page says the data counts conversion events and does not weigh their value. It does not write out the formula. Unbounce's own SaaS pages distinguish click-through pages, where the conversion is a click onward, from form-fill pages, where a visitor submits details.

That distinction matters. A click to a pricing page and a completed demo request are both conversions in a landing page tool, and they are nowhere near the same commercial event. The 2021 edition found the median SaaS form page converted just 1 in 50 paid search visitors, against 1 in 20 from social and 1 in 5 from email. That is a far lower bar than the 3.8% headline suggests if your page asks for a form.

The exclusions and the channel spread behind the median

The 2024 methodology excludes any industry, subcategory or conversion type with fewer than 400 pages. It also drops pages with fewer than 50 visitors or no conversions. Industries were assigned by a machine learning model that sorted pages into topics.

Dropping zero-conversion pages lifts the median. Every page that converted nobody is removed before the middle value is taken, and those pages exist in every real account. The report is honest that it did this. Few of the posts quoting 3.8% repeat it.

Channel moves the figure further than industry does. Unbounce reports SaaS medians of 16.9% for email traffic and 0.3% for display, a ratio of more than 50 to 1 inside a single industry. A SaaS page fed by a nurtured email list and one fed by display ads are not comparable to each other, let alone to a single blended benchmark.

The report also mixes statistics. It says some benchmarks are medians and others averages, depending on available data. Its SaaS device figures are labelled averages while its headline is a median. I would not quote any Unbounce figure without checking which of the two it is.

First Page Sage's B2B SaaS benchmarks: one agency's clients, four different cuts

First Page Sage is an SEO and marketing agency. It publishes benchmarks drawn from its own client data, which is a legitimate source and a specific one. Its pages are among the most cited conversion benchmarks in B2B, partly because they give industry rows that few other publishers attempt.

The agency's own caveat is worth repeating. Its funnel page says the benchmarks assume "a high level of competence from the team or agency conducting the activity". In plain terms, these are the results of clients who hired an agency and stayed long enough to be measured.

Two reports, one 1.1%, two definitions

The landing page report covers 80+ clients from 2021 to 2025, each in an SEO campaign of at least 6 months and an average of 2.4 years. A conversion is a form fill, appointment booking or demo signup that produces a marketing qualified lead (MQL). Micro-conversions are excluded. B2B SaaS comes out at 1.1%.

The separate industry report uses client data from January 2022 to August 2025 and defines the rate as conversion actions divided by total unique website visitors. B2B SaaS is again 1.1%. Same figure, different window, different wording of the numerator, and no client count on the page.

Both pages carry 2026 in their titles. Neither contains 2026 data. That is a common practice in benchmark publishing and it is still worth noticing, because "2026 benchmark" in a slide implies a recency the data does not have.

The SaaS funnel page contradicts the headline 1.1%

The agency's B2B SaaS funnel benchmarks draw on 50+ SaaS clients over "the last decade", mostly with revenue between $10M and $100M. Here a lead is any non-spam visitor who submitted contact information, a looser bar than an MQL.

On that definition, visitor-to-lead runs from 0.7% for paid search to 2.2% for LinkedIn. By company size it falls from 2.3% for businesses under $10M revenue to 0.7% for those above $1B. Most of its SaaS sub-vertical rows sit above 1.1%.

So the same publisher reports a B2B SaaS rate of 1.1% on one page and channel rates up to twice that on another. The difference is the sample and the definition of a lead, not a change in the market. If you benchmark against First Page Sage, name the page.

The free trial numbers count a single paid month

The free trial benchmarks cover 86 SaaS companies, 71% of them B2B, from Q1 2022 to Q3 2025. For organic traffic, opt-in trials convert to paid at 18.2% and opt-out trials, which take a card up front, at 48.8%.

The page defines a conversion as "even a single paid month". An opt-out trial that bills a forgotten card once and churns counts the same as a customer who stays three years. The page itself tells you to read it next to churn and lifetime value, which is the right instruction and the one most quoting posts leave out.

Multiplying the stages tells a different story from the trial-to-paid headline. The arithmetic below is ours, not First Page Sage's, and it assumes the visitor and trial stages describe the same companies, which the page does not confirm.

Per 1,000 organic visitors, the derived figures are roughly 15 paying customers from an opt-in trial, 12 from an opt-out trial and 3 or 4 from freemium. That ordering is the opposite of what the trial-to-paid figures alone suggest. The broader argument about how free users behave once agents enter the funnel is made in the piece on product-led growth when the user is an agent.

Baymard's cart abandonment average is an ecommerce benchmark in SaaS decks

Baymard Institute is a usability research firm, and its cart abandonment list is one of the most cited statistics in conversion optimisation. It turns up in SaaS checkout and pricing page discussions often enough to belong in this comparison. It does not belong in your SaaS benchmark.

How the 70% figure is built, and why it shows two values

The list page, last updated 22 September 2025, reports an average of 70.22% across 50 studies. Baymard's checkout usability research page still describes a list of 49 statistics averaging 70.19%. The gap is trivial in size. It shows that even a careful publisher's figure exists in two versions at once.

The studies run from a 2006 MarketingSherpa figure of 59.80% to a 2025 result of 71.72%. Individual values range from 55.00% to 84.27%. The page does not define abandonment and does not say whether the average is weighted. Each underlying study defined it its own way, across nearly 20 years of changing checkout design.

The reasons data is more useful than the headline. Baymard reports that 42% of US online shoppers have abandoned a cart because they were "just browsing". Among the remaining reasons, 40% cite extra costs, 18% being asked to create an account and 17% a checkout that was too long or complicated. A SaaS self-serve checkout shares some of those failure modes, particularly forced account creation and surprise costs at the final step.

Baymard also claims the average large ecommerce site can gain a 35.26% conversion increase through better checkout design, from benchmarking 344 US and EU sites. That is a ceiling estimate for solvable usability issues on large retail sites. It is not a forecast for a B2B SaaS signup form.

Which conversion benchmarks are comparable, and which are not

Two benchmarks are comparable when they share a numerator, a denominator, a statistic and a roughly similar sample. On that test, almost none of the commonly paired figures qualify. The verdicts below are ours and apply the test strictly.

The practical rule follows from the table. Compare figures within one publisher's report, on one page, using one definition. Across publishers, treat each number as a description of that publisher's customers rather than a fact about SaaS. The same discipline applies when a vendor quotes a channel multiple, as covered in the analysis of the AI referral traffic conversion multiple.

Where this comparison is weakest

This post argues that the common benchmarks are routinely misused. It does not argue that they are useless, and the case for them is stronger than the critique alone suggests.

Agency and vendor data is still real data

Unbounce's dataset is very large, with 57M+ conversions in a single year, and it publishes exclusion rules most competitors do not. First Page Sage gives industry and channel rows that independent research has not produced. A biased sample with a stated method beats an unbiased sample that does not exist.

No neutral, audited benchmark for B2B SaaS conversion is publicly available that we could find. Until one is, vendor and agency data is what the industry has, and the right move is to qualify it rather than discard it.

Our comparison has its own gaps

We worked only from what each publisher states on its own pages. We could not see the underlying data, the per-row sample sizes or any confidence intervals, because none of these pages publish them. Some of our "not stated" entries may be answered in material behind a download form.

We also left out ad platform benchmarks. WordStream and LocaliQ publish Google Ads conversion rates by industry, but their page refused automated access during our research, and the secondary summaries we found disagreed on the campaign count and the headline rate. We would rather omit a benchmark than describe its methodology from a copy. The derived visitor-to-paid figures above are our own multiplication and inherit every assumption noted beside them.

How to build a conversion benchmark you can actually use

The most useful conversion optimisation benchmark is your own funnel measured the same way over time. External figures are for sanity checks. My recommendation is to build the internal one first and only then go looking for a comparison.

Define the numerator as one named event, such as a demo request that reaches your CRM. Define the denominator as unique visitors to the page or site, and write down which. Pick a median or an average and keep it. Then split by traffic source before you compare anything, because the Unbounce channel data shows that source alone can move a SaaS rate by more than 50 times.

Your form design changes the event you are counting. Adding or removing gating moves the conversion rate without changing demand, which is the tradeoff examined in the economics of gated content forms. Record any change of that kind next to the number, or your own trend line stops being comparable to itself.

Finally, be honest about what you cannot see. A growing share of B2B buyers research before they arrive and convert on a later, direct visit, which flatters some pages and starves others. That measurement problem is set out in the piece on attribution and dark discovery, and it affects every benchmark here equally.

Frequently asked questions

Where to start this week

Find the conversion benchmark currently sitting in your planning deck or board slides. Next to it, write the publisher, the data window, the definition of a conversion and whether it is a median or an average. If you cannot fill in all four from the source page, replace the number with your own trailing 90 days.

Then split your own rate by traffic source before your next optimisation review. If email traffic converts several times better than paid, the blended figure is describing your channel mix, not your page. Fix the comparison before you fix the page. Teams publishing their own comparison pages can apply the same rule to the numbers they cite, as set out in the B2B comparison page template.

Related reading