40 A/B Test Ideas for SaaS Landing Pages (Ranked by Likely Impact)

How should you rank landing page test ideas?

Rank landing page test ideas by two things: how much the change could plausibly move conversion, and how much traffic you'd need to detect that move. Those two are linked. A change that could shift conversion by a third is detectable with a few thousand visitors per variant; one that could shift it by a few percent needs tens of thousands, as our guide to A/B test sample size shows.

So the list below is ordered in four tiers, from changes to the offer and message down to visual details. The ranking is our judgement of typical impact from experience with SaaS pages, not measured averages, and your research should always override it. This list is part of our complete guide to A/B testing landing pages.

Which AB test ideas should you try first?

Try offer and path ideas first, because they change the decision the visitor is making, not just how it's presented. These are tier 1: the ideas with the largest plausible effect and the smallest traffic requirement.

1. Free trial vs demo request as the primary action.
2. Free plan vs time-limited trial, if your pricing allows both.
3. Card required vs no card for the trial.
4. A specific CTA label vs a generic one. In an Optimizely test reported by Convertica, "Watch a Demo" converted at 1.59% against 0.91% for "Get Started".
5. Instant scheduling after the form vs a "we'll be in touch" message.
6. Interactive demo or tour as the primary action vs sign-up, for complex products.
7. Showing prices on the page vs linking to a pricing page.
8. A lower-commitment first step (template, calculator, sandbox) vs a direct sign-up.
9. Single-step vs two-step form, where step one asks only for email.
10. Removing the secondary CTA so the page has one action.

What message and headline tests are worth running?

Message tests are worth running when visitors arrive but don't understand fast enough what the product does or who it's for. These tier 2 ideas change the first screen, where most visitors decide whether to keep reading. Our landing page headline formulas give you strong variants to test.

1. Outcome headline vs category headline ("Close the books in two days" vs "Accounting automation").
2. Naming the target user in the headline vs a general audience.
3. Naming the workaround you replace (spreadsheets, agencies, a legacy tool) vs not.
4. Pain-led vs gain-led framing of the same value proposition.
5. A specific number in the subhead (only if it's true) vs a qualitative claim.
6. Product UI in the hero vs an illustration or abstract visual.
7. A short loop of the product working vs a static screenshot.
8. Plain-language description vs technical description, especially for AI and devtools.
9. Message matched to the ad or keyword vs the generic homepage message.
10. Leading with the biggest objection answered ("No migration needed") vs leading with the benefit.

For AI products in particular, "magic promise, no proof" is the pattern commentators keep naming, so testing a concrete outcome plus visible proof against it is a strong tier 2 idea. Our guide to AI landing page design covers that pattern.

Which page structure and proof tests come next?

Structure and proof tests come next, once the offer and message are settled. These tier 3 ideas change the order, length and evidence on the page. Their effects are usually smaller than tier 1 and 2, so they need more traffic.

1. Long page with objection handling vs short page.
2. Moving pricing above or below the features.
3. A customer quote next to the CTA vs a logo strip.
4. One detailed customer story section vs several short testimonials.
5. An FAQ section answering sales objections vs none.
6. A comparison table against the main alternative vs no comparison.
7. Use-case sections by role vs one generic feature list.
8. A security or compliance summary on the page vs a footer link, for B2B.
9. Sticky CTA bar on mobile vs CTAs only in sections.
10. Founder note or team section vs none, for early-stage products.

Which small details are worth testing at all?

Small details are worth testing only on pages with a lot of traffic, after the bigger tiers are exhausted. These tier 4 ideas usually produce small effects, which need very large samples to detect reliably. On most startup pages, just ship the better version.

1. Button color or contrast, if the current button is genuinely hard to see.
2. Button size and placement within a section.
3. Number of form fields beyond the first two or three.
4. Field labels and placeholder text.
5. Image swaps between similar product screenshots.
6. Icon style in feature lists.
7. Micro-copy under the button ("No card required", "Cancel anytime").
8. Navigation shown vs hidden on a paid landing page.
9. Exit-intent or scroll-triggered offers.
10. Testimonial photo vs no photo.

How do you turn an idea into a test?

Turn an idea into a test by tying it to something you observed and writing it as a hypothesis before you build anything. An idea from a list is a starting point; the observation is what makes it your test and not someone else's.

1. 1

   Find the evidence

   Check analytics, recordings, support tickets and sales notes for where visitors hesitate or misunderstand.

2. 2

   Pick the matching idea

   Choose the highest-tier idea from this list that addresses that problem.

3. 3

   Write the hypothesis

   Use our A/B test hypothesis template: because we observed X, we believe changing Y will cause Z.

4. 4

   Check the traffic

   Confirm the sample size is reachable in a few weeks for the lift this tier can plausibly produce.

5. 5

   Build, QA and run it properly

   Avoid the analysis traps in A/B testing mistakes that produce false winners.

For variant inspiration, our SaaS landing page examples and the live hero rewrite in our landing page teardown show what strong versions of tier 1 and 2 changes look like on real pages.