'Book a demo' vs 'Start free': which primary CTA should my landing page use?

There is no single winner in the data — the demo-vs-free-trial swap runs in both directions across 232 detected CTA experiments, where free-trial wording appears in 77 (33%) and demo paths in 25 (11%).[1] The recurring pattern is that product-led companies add a self-serve trial while shifting toward enterprise adds a demo escape hatch. Choose by buyer type, not by a universal best CTA.

Free-trial wording appears in 33% of detected CTA experiments; demo paths in 11% — July 2026.

Lazyweb Research · n=232 · Published 2026-07-07

landing-pageexperimentstrialswebsaas

Finding: the swap goes both ways

Across 232 CTA experiments, wording themes break down as (non-exclusive keyword incidence):[1]

CTA themeCountShare of 232
Free-trial wording7733%
Price embedded in button5323%
Demo path2511%
Waitlist1<1%

Demo-vs-free-trial swaps show up in both directions in the qualitative rows, so the data supports segmentation, not a fixed answer.

Breakdown: two directions, two rationales

Detected CTA swaps illustrate each direction (observed change + inferred rationale, not measured lift):[2]

  • Toward self-serve — Cycle (web): primary nav CTA 'Book a demo' changed to 'Start free trial.' Inferred rationale: moving from sales-led to product-led so self-serve teams experience value immediately.
  • Toward a sales path — Calendly (web): added dual hero CTAs 'Sign up for free' (primary) + 'Get a demo' (secondary) where control had no in-hero CTA. Inferred rationale: the shift toward team/enterprise buyers means a segment needs a sales path a PLG-only control never surfaced.

How to apply

Match the CTA to who converts. If you sell bottom-up to self-serve teams, the pattern favors a free-trial primary (Cycle's direction). If you are moving upmarket, the pattern favors pairing self-serve with a demo secondary rather than replacing it (Calendly's direction). All of these are observed changes with inferred rationale — validate with your own test before rolling out.

The numbers

StatComputed from
Free-trial wording 77 (33%), price-in-CTA 53 (23%), demo path 25 (11%), waitlist 1 of 232 CTA experimentscta_change_themes
Cycle and Calendly CTA swap examplesqualitative[cycle, calendly]
Methodology. Universe: 232 detected CTA experiments, July 2026. Method: non-exclusive keyword incidence plus named-company qualitative rows. Caveat: demo-vs-trial swaps run both directions; all rows are observed changes with inferred rationale, not measured lift.

Sources & citations

  1. [1] Lazyweb Research analysis of 232 CTA experiments (detected-experiment corpus), July 2026. Keyword incidence over summary+learning text; non-exclusive lower bounds.
  2. [2] Lazyweb Research analysis of 225 named-company CTA experiments (detected-experiment corpus), July 2026. Observed before/after CTA changes with model-inferred rationale; not measured A/B lift.

Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.

Related questions

Explore the underlying screens, flows, and A/B tests inside Lazyweb. More research