Email traffic converts SaaS landing pages at a median 16.9%, paid search at 4.1%, paid social at 2.9% and display at 0.3%, according to the SaaS section of Unbounce's Conversion Benchmark Report. The 3.8% SaaS median everyone quotes is the blend of those four numbers, so a paid-media team should never be judged against it.
This page is the SaaS-only cut: channel and copy data inside the category, the form-field studies read together, the speed and message-match tests, and what trial and demo asks do to the denominator. Cross-industry medians are in "Landing Page Conversion Rates by Industry 2026"; the funnel after the form is in "Demo Request Conversion Rates 2026".
The channel cut inside SaaS
Unbounce's report covers more than 41,000 landing pages, 464 million visitors and 57 million conversions. Inside SaaS, every paid channel sits further below the all-industry figure than the median suggests.
| Channel | SaaS | All industries |
|---|---|---|
| 16.9% | 19.3% | |
| Paid search, all engines | 4.1% | 10.9% |
| Google search ads | 5.1% | 11.3% |
| Paid social, all networks | 2.9% | 12.0% |
| Instagram ads | 9.2% | 17.9% |
| Facebook ads | 3.5% | 13.0% |
| Display ads | 0.3% | Not reported |
| Category median | 3.8% | 6.6% |
Source: Unbounce Conversion Benchmark Report, SaaS page and overview, 2024
Email is the only channel where SaaS nears the rest of the market. Paid search converts at under half the all-industry rate, so a 5% Google Ads page is on the median, not under it. Display converts one visitor in roughly 300. The SaaS sub-categories, hardware at 4.1% and data and infrastructure at 3.3%, are in the industry post.
Device and copy: the two on-page findings for SaaS
Device. Mobile carries 79% of SaaS landing page visits, and SaaS is one of two industries where mobile does not convert worse: a 6.4% average on mobile against 6.2% on desktop, where all industries together have desktop 8% ahead. Those are averages, not medians, so both sit above 3.8%.
Reading level. This is the largest single effect in the SaaS data. Pages at a 5th to 7th grade reading level convert at 12.9%; professional-level pages convert at 2.1%. The best-converting SaaS pages run 250 to 725 words with 50 to 140 words of three or more syllables. Across all industries, difficult words correlate with conversion at minus 24.3%.
Form length: three datasets, one curve and one correction
The demo-request post gave the field studies one row each. Here is the full NP Digital curve, because its shape matters more than any single point.
| Fields | Conversion rate |
|---|---|
| 1 | 18.2% |
| 2 | 13.0% |
| 3 | 11.5% |
| 4 | 9.9% |
| 5 | 8.1% |
| 7 | 7.6% |
| 8 | 5.8% |
| 9 or more | 4.2% |
Source: NP Digital, 404 landing pages with form conversions, December 2024. No figure is published for six fields.
The steepest loss is between one field and two: 5.2 points. From three to seven the curve is nearly flat, then it falls again. NP Digital's second study, 833 top-of-funnel pages (November 2024), starts higher at 28.2% with one field and 20.5% with five, then drops to 8.6% at seven and 5.1% at nine or more. The shape agrees: the first two fields are expensive, three to five are cheap, and the seventh starts a second cliff.
The correction from HubSpot. HubSpot's analysis of more than 40,000 landing pages found conversion falls as fields are added, but less steeply than its author expected. Field type was decisive: extra single-line fields made very little difference, multiple textareas had a "powerful depressing effect", and multiple dropdowns were associated with lower rates.
Multi-step forms. Formstack is the most-quoted source for multi-step lifts, and three different figures circulate under its name. None links to a primary report we could open, so none appears here.
Form completion once someone starts: Zuko's software row
Zuko's form analytics benchmarks track 93 million form sessions, with conversion defined as view-to-completion. Its software row is 557,487 sessions, 0.6% of a finance-heavy dataset: directional only.
| Form type or industry | Sessions | Desktop | Mobile | Overall |
|---|---|---|---|---|
| Software (industry) | 557,487 | 53.5% | 38.6% | 50.6% |
| Contact | 321,937 | 51.2% | 39.4% | 47.4% |
| Enquiry | 4,797,087 | 47.7% | 37.7% | 42.7% |
| Registration | 26,582,277 | 63.0% | 57.2% | 60.7% |
| All forms | 93,022,997 | 54.5% | 47.5% | 51.7% |
Source: Zuko form type and industry benchmarking pages, 2026
Software has the widest device gap in the table, 14.9 points against 7 for all forms; next to Unbounce's 79% mobile share, the mobile form is the likelier leak.
Message match: the one controlled test worth citing
Message match, making the headline repeat the ad or query that brought the visitor, is widely recommended and rarely measured. The cleanest published test is Unbounce's write-up of a ConversionLab experiment for Campaign Monitor: dynamic text replacement swapped the headline verb to match the visitor's Google search. Over 77 days from October 2017 to January 2018, on 1,274 visitors across two variants, the matched version lifted trial sign-ups 31.4%, with more than 100 conversions per variant.
That is one small test run by an agency with a stake in the result, but it is the only one we found with a stated duration and traffic count.
Page speed: three studies, one direction
| Study | Sample | Finding |
|---|---|---|
| Google and SOASTA, 2017 | Model trained on bounce and conversion data | 53% of mobile visitors leave a page taking over three seconds; bounce probability rises 113% from one second to seven |
| Deloitte and 55 for Google, 2019 to 2020 | 37 brand sites, 30 million sessions, 30 days | A 0.1-second gain lifted lead-generation progression to the form-submission page 21.6%, retail product-to-basket 9.1%, travel bookings 10% |
| Portent, 2022 | 100 million+ page views, 20 sites, 14 B2B | B2B conversion near 40% at one second, 34% at two, 29% at three; a one-second site converts 3x a five-second site |
Sources: AMP Project blog reporting Google and SOASTA; web.dev case study of Milliseconds Make Millions; Portent
The Deloitte figure is the one for a SaaS team, because lead generation is the nearest vertical to a demo or trial page. The study required four metrics to improve on every page in the journey, so 21.6% is for a whole-funnel fix. Portent's B2B rates use those sites' own conversion definitions, so read the multipliers. We build marketing sites in Webflow and put load time in the page brief for this reason.
Free trial versus demo: the CTA decides the denominator
What a page asks for moves its conversion rate more than the layout does. Kyle Poyar's January 2026 analysis of 200 B2B products with ChartMogul and ProductLed, reported by Userpilot, gives each model per 1,000 visitors.
| Model | Free sign-ups | Paying customers within six months |
|---|---|---|
| Freemium to premium | 90 | About 5 |
| Ungated freemium | 70 | 5.6 |
| Standard free trial, no card | 45 | 3.6 |
| Credit-card-required trial | 35 | 10.5 |
Source: ChartMogul and Growth Unhinged, 200 B2B products, January 2026, as reported by Userpilot
The median trial-to-paid rate is 8%, and the distribution is bimodal: 20% of products convert under 2.5% of trials and 23% convert over 25%. The card-required page has the lowest sign-up rate and produces three times the customers from the same traffic: judge a page on customers per visitor, not conversion rate.
Demo pages sit lower. Saas Hero's aggregation of third-party figures puts demo-request pages at a 1.5% to 4% median with a 4.8% top quartile and trial pages at 8%; both are vendor compilations, and the trial figure conflicts with ChartMogul's 4.5%.
Video and social proof: thinner evidence than the advice
Video. Unbounce examined roughly 35,000 customer pages in each of three windows from late 2020 to early 2021. On form pages, embedded video made little difference and background video did slightly better. Click-through pages with video converted worse, and video pages had a lower conversion distribution across every channel and both devices. The much-shared Vidyard case of video doubling conversion is a single 2011 test.
Social proof. We found no controlled study of logos or testimonials on B2B landing pages that we could verify. What exists is buyer research: TrustRadius's 2026 B2B Buying Disconnect report, from 1,862 technology buyers, finds that 74% use reviews to inform the decision and that 83% shortlist three or fewer products. That argues for a review score near the form; it is not a lift figure.
How to use these numbers on a SaaS page
Set one target per channel, not per page: nurture email against 16.9%, Google Ads against 5.1%, paid social against 2.9%. Test the headline first, where reading level and message match carry the largest measured effects. Cut the form to the fields that qualify, replace any textarea with a single-line field, and measure mobile form completion separately. Fix speed before adding video, because the speed studies show a lift and the video data shows none. Our landing page service builds and tests pages against this set of numbers.
