Testimonial A/B Tests: Do Testimonials Increase Conversions?

The best testimonial examples in A/B tests are short, specific quotes from customers who look like the visitor, placed right next to the form or button where the visitor decides.

In our database, a single student quote above a brochure form lifted submissions 50% for Hotel Institute Montreux, and three short testimonials lifted purchases 34% for WikiJob.

But testimonials can also lose, especially when they push the action further down the page or show up for visitors who had already decided.

This hub covers SaaS and lead-gen tests tagged testimonials, reviews or social proof.

The full grid of matching tests is below.

This page is the analysis: what won, what lost, and how to test testimonials on your own site.

For ecommerce star ratings and product reviews, see review A/B tests.

For logos, user counts and other proof types, see social proof A/B tests.

What customer testimonial A/B tests show

Most testimonial tests in our database won. That sounds better than it is: many are bundled redesigns where testimonials were one change among several, and teams publish winners more often than losers.

The cleaner evidence comes from the handful of tests that changed only the testimonials or only their position.

Treat the big percentages with care. Mida's test was about 200 sign-ups on its own product, and the KlientBoost and WikiJob write-ups give no sample size or significance.

They are good evidence that testimonials can matter, not a forecast of what yours will do.

Pattern 1: placement beats presence

The strongest pattern is position. WikiJob's case study notes that an earlier test with testimonials lower on the page had no effect; moving three of them near the top produced the 34% lift.

KlientBoost's client nearly doubled its conversion rate just by moving existing testimonials higher, and IMD's alumni videos did the same kind of work once they moved up.

Testimonials that sit at the bottom of a long page are mostly seen by people who already scrolled through everything.

The ones that move results sit next to the form, the price or the main button, where doubt peaks.

Pattern 2: specific and relevant beats generic praise

Buildium's winning variant showed each customer's portfolio size (unit count) beside the quote, so a property manager with 50 units could find someone like them.

Hotel Institute Montreux used a student, not a parent or an industry partner, because students were the ones requesting brochures.

"Great product, highly recommend" tells a visitor nothing they can check; "We cut our month-end close from five days to two" does.

Pattern 3: they don't always help

The losses are as useful as the wins.

Dallas Theological Seminary's donation page lost 28% when NextAfter put value copy and endorsement quotes ahead of the ask: visitors reaching that page had already decided to give, and the extra material slowed them down.

On a Shopify product page, a block of three testimonials placed beside the add-to-cart area cut conversion 9.2% in Clean Commit's test (not significant, but down on both devices).

Voices.com gained 33% more sign-ups by removing a featured-talent section from its homepage, and a 28% retest confirmed it.

Proof that crowds the page or pulls attention off the main action can cost more than it adds.

Pattern 4: the details of the proof block matter

WorkZone kept its testimonials and only turned the customer logos beside them from full color to black and white.

Demo request submissions rose 34% over 22 days at 99% significance.

The likely reason is visual: bright logos competed with the form for attention.

How a testimonial looks can matter as much as what it says.

Pattern 5: testimonials stack with other proof

Testimonials are one kind of proof among several, and the tests suggest they add to the others rather than replace them.

In a three-arm KlientBoost test, adding general social proof took a lead page from 33.22% to 39.95%, and adding testimonials on top of it reached 43.52%.

Planoly's sign-up page paired a testimonial with achievement badges and a single sign-up button, and CodaLift reported 17% more paid clients, though that was a bundled change.

Format matters less than fit. IMD used alumni video testimonials, Hotel Institute Montreux used a text quote, and both won once the proof sat near the form.

Video costs more to produce, so test a strong text quote in the same spot first; if it wins, a video version is the follow-up test.

Should you build a testimonials page?

A dedicated testimonials page (often called "customer stories" or "wall of love") is a common B2B pattern.

Our database has no A/B test of adding or removing a standalone testimonials page, so we can't tell you it lifts conversions.

What the tests above do show is that proof works hardest on the page where the visitor is deciding, not one click away from it.

That suggests a sensible split:

• Testimonials page: a full archive for prospects doing research, sales teams who need a link to send, and searchers looking for reviews of your company. Organize it by industry, company size or use case so visitors can find someone like them.
• Decision pages: pull the two or three most relevant quotes from that archive onto the sign-up page, pricing page, demo form and key landing pages. This is where the tested lifts came from.

If you do test a testimonials page, measure it on what it is for: assisted demo requests or trial starts from visitors who viewed it, not its own bounce rate.

Expect few visitors to see it, which makes a clean A/B test slow.

Stay within the rules

Testimonials are regulated advertising in the US. The FTC's rule on consumer reviews and testimonials, effective October 21, 2024, bans fake testimonials (including AI-generated ones), undisclosed testimonials from insiders, and buying fake social media indicators, and it lets the FTC seek civil penalties.

Under the FTC's Endorsement Guides, a testimonial must reflect the customer's real experience, and if it features unusually good results you need proof they are typical or must state what people can generally expect.

Never test a made-up quote as a "placeholder".

How to test testimonials on your own pages

Most sites already have testimonials somewhere. The tested opportunities are usually about moving, choosing and trimming them, not writing new ones.

1. Find the decision point. Look for the page with the highest drop-off before a sign-up, demo request or lead form. That is where Hotel Institute Montreux, Mida and WikiJob put their quotes.
2. Test position first. If testimonials exist below the fold, move the best two or three next to the form or main button. It is the cheapest test and the pattern with the most support in our data.
3. Match the quote to the visitor. Pick customers with the same role, company size or problem as the page's audience. Add a name, job title, company and photo when the customer agrees to it; anonymous quotes are easy to dismiss.
4. Prefer concrete outcomes. Quotes that name a result, a timeframe or an objection overcome ("I thought setup would take weeks; it took an afternoon") give visitors something to believe. Make sure any results claim is typical or properly qualified.
5. Keep it short and quiet. WikiJob's winners were three short, sober quotes. WorkZone's lift came from making the proof less loud. Don't let a carousel or bright logos pull attention off the form.
6. Watch for motivated audiences. On pages reached by people who have already decided, such as a donation page, checkout or a returning-user log-in, test removing proof as well as adding it.

Pick one primary metric (form completions, trial starts or purchases), estimate the traffic you need before launch, and run full weeks.

Many published testimonial wins come from small samples, so check statistical significance rather than stopping at the first good day.

For more tested ideas across every page element, browse the elements index or search all A/B tests by vertical and page type.