SaaS conversion rate optimization (SaaS CRO) is the practice of testing changes to your website, sign-up flow, pricing page and trial so that more visitors become paying, retained customers.
The goal is not more sign-ups but more revenue per visitor: in our database of 205 SaaS A/B tests, the changes that moved revenue most often sat on the pricing page, the trial offer and the first session in the product, not the homepage.
This page walks through what those 205 tests show, how B2B SaaS differs, and where to start on your own site.
Every claim below links to a test you can read in full on our SaaS hub.
SaaS CRO: what 205 real tests show
Ecommerce CRO usually ends at the order. SaaS CRO has a longer chain: visit, sign-up or demo request, activation, then the first payment and renewal.
A change can lift one step and hurt a later one, which is the single most important thing to understand before you run a test.
Where the 205 tests sit in the funnel
Here is how the SaaS tests in LiftArchive break down by page type.
It reflects where companies chose to test and publish, not where the biggest opportunities necessarily are.
| Page type | Tests | What usually gets tested |
|---|---|---|
| Landing page | 51 | Paid-traffic pages, form length, navigation removal |
| Homepage | 30 | Headline, primary CTA, sign-up form placement |
| Pricing page | 26 | Plan packaging, highlighted plan, annual vs monthly display |
| Other (product, tour, content pages) | 23 | Product page copy, screenshots, link text |
| Signup page | 18 | Form fields, reassurance copy, logos, testimonials |
| Checkout | 13 | Payment options, cart, upsells |
| Free trial | 11 | Trial length and limits, trial offer pages |
| Onboarding | 8 | Checklists, guided setup, first-run experience |
| Site-wide | 8 | Navigation, redesigns, logo |
| Demo request | 6 | Demo forms, pop-ups, exit removal |
| Email and ads | 11 | Subject lines, upgrade emails, ad copy |
Of the 205 tests, 172 are recorded as winners, 23 as losses and 10 as flat.
Treat that ratio with care. Companies publish wins far more often than losses, so a database of published tests overstates how often ideas work.
The internal numbers from large testing programs are much less flattering.
In a 2013 paper, Ron Kohavi and colleagues at Microsoft reported that only about one third of ideas tested at Microsoft improved the metrics they were designed to improve.
Plan your testing roadmap on the assumption that most of your tests will not win.
Measure the step that pays, not the step that is easy to count
The clearest lesson from the SaaS tests is that a lift in sign-ups can hide a drop in revenue. Three tests show it from different angles.
- Server Density replaced per-server pricing with packaged tiers. Free sign-ups fell from 6.92% to 5.19%, but revenue rose 114%, from $394 to $833 in the test period. Read the Server Density pricing test.
- Patrick McKenzie redesigned the Bingo Card Creator site in 2008. Weekday trial downloads rose from about 100 to 125-150, but sales fell from about two a day to slightly under one. See the 2008 redesign result.
- A 2012 full redesign of the same product lifted trial sign-ups on AdWords pages from 20.9% to 23.1%, while purchases stayed flat at 50 vs 46. See the 2012 redesign result.
The fix is simple to state and harder to do: pick a primary metric as close to revenue as your traffic allows, such as paid conversions or revenue per visitor, and track sign-ups as a secondary metric.
If you can only measure sign-ups, at least check trial-to-paid for the test cohort before you roll a winner out.
The pricing page and the offer move revenue most
The biggest revenue changes in our SaaS data came from what was being sold and how, not from button colors.
- Acuity Scheduling offered a 14-day Pro trial instead of a free-forever plan and recorded a 268% rise in paid sign-ups.
- Patrick McKenzie cut the Bingo Card Creator free trial from 15 cards to 8. Paid conversion rose from 1.04% to 1.89%. Read the trial limit test.
- Powtoon showed "10 GB" of storage instead of "unlimited" on the pricing page and saw +27.9% revenue (test published by VWO).
- CodaLift LLC redesigned the Planoly pricing page with a highlighted middle plan. Revenue rose 38% and average order value 24%. See the Planoly pricing test.
Not every pricing change works. Powtoon's new monthly vs annual price display lost, and Conversionista's test moving Fortum's price details into a modal cut click-through by 9% before a later iteration won +9%.
See all SaaS pricing page tests for the full set.
MarketingExperimentsTrials, free plans and how you frame them
How you offer the product before payment is one of the biggest levers a SaaS company has.
The tests do not say "always use a trial" or "always use freemium"; they say the framing is worth testing.
- Acuity Scheduling's switch from a free-forever plan to a 14-day Pro trial is the strongest result in the set, covered above.
- Emaze kept its free plan but added a 30-day premium trial next to it, which brought more paid upgrades with a low cancellation rate (MarketingSherpa gave no exact figure).
- GetResponse added a "Free Trial" button next to "Buy Now" on its homepage and saw +158.6% free accounts.
- Going changed "Sign up for free" to "Trial for free" and recorded +104% trial starts, measured month over month rather than in a split test.
The wording can also backfire. Encyclopaedia Britannica tested "Become a subscriber" against "Start your free trial" and both variants lost, with click-through down as much as 20%.
Our SaaS free trial tests collect the rest.
Activation is the most under-tested stage
Only 8 of the 205 SaaS tests are on onboarding, yet they include some of the largest effects on paid conversion.
That gap is an opportunity: activation sits right before payment, so lifts there tend to reach revenue.
- VWO added a "Get Started" checklist for free-trial users and recorded +127% free trial to paid conversions.
- AURUM added a floating onboarding checklist inside its Astrea trial. Activation rose 187%, with key setup events up 21% and 25%. Read the AURUM checklist test.
- Citrix added an in-product guided walkthrough to the GoToWebinar free trial and doubled conversions.
There are more examples in our SaaS onboarding tests.
If you run a self-serve product, our product-led growth page covers the free-plan and trial side in more depth.
Sign-up friction: fewer steps usually wins, but not always
Cutting steps and fields is the most repeated pattern in the data, and it holds up well.
- MarketingExperiments removed one page from Encyclopaedia Britannica's registration flow. Free trials rose 117%, paid subscriptions 61% and revenue 130%. Read the Britannica test.
- Kareo Marketing cut fields on the DoctorBase sign-up form and recorded +30% new doctor sign-ups.
- Teamleader added "No credit card required" to its trial sign-up page. Conversion rose from 48.7% to 54.8% in the Netherlands and from 33% to 38.6% in Belgium. See the Teamleader test.
The counterexamples matter. A three-step free trial sign-up on RegOnline's homepage produced 24.5% fewer completed leads.
A madlibs-style sign-up form on Bingo Card Creator converted at 21.73% against 27.55% for the standard form, about 22% worse.
Novel form formats that worked for one company are worth testing, not copying.
Friction is not always bad either. MarketingExperiments added an email field to a free trial call to action and the trial sign-up rate rose from 0.68% to 1.57%, +133%.
Asking for a small commitment early can filter for intent. Browse SaaS sign-up page tests for more.
Proof near the decision point
Social proof works best in SaaS when it sits next to the form or plan choice and matches the visitor.
- VWO showed country-specific customer logos on its trial and demo pages and recorded +376% leads.
- Buildium used a size-focused headline and testimonials that named each customer's unit count for +22% sign-ups.
- PayPanther built a longer pricing and sign-up page with FAQs and testimonials and saw +372.6% sign-ups.
Proof is not automatically good. AssessmentDay removed the FAQ section from a landing page for +62% sales, the opposite of PayPanther.
Context decides, which is why a test beats a rule.
B2B SaaS conversion rate optimization
B2B SaaS adds a sales team, longer buying cycles and several people in each decision.
That changes what you can test and how you measure it.
Low traffic is the main constraint
Most B2B SaaS sites have far less traffic than a mid-sized online store, and the conversions that matter (demos, qualified opportunities, closed deals) are rarer still.
A test that needs thousands of conversions per variant will not finish on a page that gets 40 demo requests a month.
Here is a worked example using the standard sample size formula for comparing two conversion rates at 95% confidence and 80% power:
n per variant = (1.96 + 0.84)² × [p1(1 − p1) + p2(1 − p2)] / (p2 − p1)²
Say your demo page gets 2,000 visitors a month and converts at 3%.
You want to detect a 20% relative lift, from 3.0% to 3.6%.
| Lift to detect | p1 | p2 | Visitors per variant | Months at 2,000 visitors/month |
|---|---|---|---|---|
| 20% relative | 3.0% | 3.6% | about 13,900 | about 14 |
| 30% relative | 3.0% | 3.9% | about 6,450 | about 6.5 |
Halving the lift you want to detect roughly quadruples the sample you need, because the difference is squared in the formula.
Revenue metrics need even larger samples: Kohavi and colleagues note in their practical guide to controlled experiments that conversion rate typically has lower variability than revenue, so tests on revenue take longer to detect the same relative change.
This formula is the wrong choice when visitors are not independent (for example, several people from one account), when you check results daily and stop at the first significant reading, or when conversions are so rare that the normal approximation breaks down; in those cases use a sequential or exact method, or test a higher-traffic step.
Before you start a test, run the numbers in our A/B test sample size calculator.
If the test would take more than a couple of months, you have three practical options:
- Test bigger changes, which produce larger effects that need fewer visitors to detect.
- Test on a higher-traffic step, such as clicks to the demo page, and confirm the downstream effect in your CRM.
- Make the change without a test when the evidence from other companies is strong and the risk is low, then watch the trend.
Demo request pages: remove exits, add proof
For sales-led products, the demo request page is the B2B equivalent of checkout.
The tests here point the same way: fewer distractions and more proof near the form.
- POSist removed header and footer links from its demo page for +7% conversions, and a narrower demo form with logos and testimonials brought +20% demo sign-ups.
- WorkZone switched to black-and-white customer logos on its demo page and saw +34% form submissions.
- Apexure built Flare a short Book a Demo page with a headline, logos and a form, and recorded +65% conversion.
More examples are on our SaaS demo request tests page.
Track the lead through to revenue
B2B tests are easy to win on lead count and lose on lead quality.
The best published B2B tests report a metric past the form.
Omnisend featured custom plans on its pricing page and recorded +22.97% demo requests, but also a lead-to-customer rate that rose from 2.94% to 11.54% and +16.69% new business MRR. Read the Omnisend test.
To do the same, pass the test variant into your CRM as a field on the lead, then compare opportunity and close rates by variant once enough deals have closed.
Small samples here are normal; treat the result as a direction, not a verdict.
CTA wording matters more in B2B than you would expect
B2B buyers hesitate at anything that sounds like a sales call.
Several tests found big swings from a few words.
- Veeam Software changed "Request a quote" to "Request pricing" for +161.66% click-through.
- Greenlane Search Marketing changed Corcentric's homepage CTA from "Get a Demo" to "Free Demo" for +99.42% clicks.
- MECLABS Institute tested "Speak to an Expert" plus a demo CTA against "Contact Sales" for VolunteerMatch and recorded +291% clicks.
Click lifts on CTAs are the easiest wins to get and the hardest to carry to revenue.
Use them, but measure the next step too.
Paid traffic landing pages
KlientBoost published 20 SaaS landing page programs in our database, most combining new landing pages with campaign changes.
The results are large, such as +407% conversion rate on a high-volume Yesware page (+48% account-wide), but they are program results rather than single controlled tests.
Read them as evidence that dedicated, focused pages for paid traffic beat sending ads to the homepage, not as the lift you should expect from one change.
Our SaaS landing page tests list every one.
SaaS website optimization: where to start
If you are starting a SaaS CRO program from scratch, this order follows where the evidence in our database is strongest and where changes reach revenue fastest.
1. Instrument the whole funnel
You need conversion rates for each step: visit to sign-up or demo, sign-up to activated, activated to paid.
Without them you cannot tell which step is weakest or check that a winning test helped revenue.
Our SaaS conversion rate benchmarks give a rough comparison point, though your own baseline matters more.
2. Fix the offer and pricing page
Test trial length and limits, plan packaging, a highlighted plan and how you describe usage limits.
These changes touch every paying customer. Server Density, Acuity Scheduling and Planoly above are the models.
3. Shorten sign-up and add reassurance
Remove fields and steps you do not need at sign-up.
Add reassurance next to the form: Teamleader's customer logos lifted sign-up conversion from 36% to 40.3%, and Mida added testimonials to its sign-up page for +56.94% sign-up completion.
4. Build an activation path
Add a checklist or guided setup that gets new users to the first useful result.
The VWO, AURUM and Citrix tests are the strongest evidence in this database that it pays.
5. Then work on the homepage and landing pages
Homepage tests are popular because the page gets traffic, and some results are strong.
Tom's Planner moved a sign-up form above the fold on its homepage for +44% sign-ups, and Hubstaff's homepage redesign brought +49% visitor-to-trial conversion.
But the homepage is also where results most often stop at the click.
Pair each homepage test with a downstream metric. See SaaS homepage tests.
Prioritize with evidence, not opinions
With limited traffic, every test slot is valuable.
Score ideas on expected impact, confidence and effort; our guide to test prioritization covers the common frameworks.
Use similar tests from the full test database to set your confidence score, and lower it for ideas that only ever won on clicks.
Treat huge lifts with suspicion
Many SaaS results in our database are three-digit percentages.
Some are real, especially where a change removed a real barrier.
But large lifts on small samples are often noise.
The Microsoft team behind the 2013 paper put it bluntly: most amazing results turn out to be false when reviewed carefully, and need replicating with high statistical power.
One test in our set shows the right instinct.
A tech supplier's value proposition rewrite showed a +445% lift that the team did not trust, so it is recorded as flat.
When you see a giant win, rerun it or check it on a later step before you build on it.
Our statistical significance guide explains what a significant result does and does not tell you.
When SaaS CRO is the wrong tool
Testing is not always the right move. If your product does not keep users, conversion work will fill a leaky bucket faster; fix retention first.
If you have only a few hundred visitors a month, you will learn more from talking to ten trial users than from an A/B test that never reaches a result.
And if a redesign is forced by a rebrand or platform change, test it to catch damage, not to prove a lift.
If you want outside help, our directory of SaaS CRO agencies lists practitioners with published SaaS tests.