Proving it works
Find out whether testimonials actually move your conversion rate
Adding social proof feels like an obvious win, until someone asks for the evidence. A fair test needs a hypothesis written down before you start and a rule for when to stop. Build both below, then run them with the analytics you already have.
- A one-page test plan written from your details
- A results log and a decision note for afterwards
- Separate plans for split tests and low-traffic pages
The problem, solved
Why testimonial tests end in an argument, and how StoryProof keeps them clean
Situation 1 of 4
The widget shipped alongside three other changes
When the proof goes live in the same release as a new headline and a price tweak, any lift belongs to all of them and none of them. Nobody can say afterwards what the testimonials did.
How StoryProof solves it
A StoryProof widget is a two-line HTML embed, so it can go into the test variant on its own, separate from every other release on the page.
Situation 2 of 4
The quotes on test speak to the wrong buyer
A result from generic praise says little about whether relevant proof would help. A flat outcome may only mean the visitor on that page never saw themselves in the quote.
How StoryProof solves it
Each widget shows only the approved testimonials you pick for it, so the variant can carry the handful written by customers like that page’s visitors.
Situation 3 of 4
We cannot agree on which format to test
A carousel, a single spotlight quote and a badge are different treatments. Arguing about it in a meeting delays the test, and guessing wastes a round of traffic.
How StoryProof solves it
StoryProof’s widget layouts run from walls and carousels to compact badges, so a follow-up round can pit two formats against each other with the same quotes inside.
Situation 4 of 4
Every new set of quotes means another deploy
After a flat result the next step is usually better testimonials. If each swap needs a developer and a release, the second round never gets scheduled.
How StoryProof solves it
The widget loads its testimonials from StoryProof each time the page opens, so the next round’s quotes go in by approving them, with no republish of your page.
Test plan builder
Write your test plan before you change anything
Fill in the page, the metric and who owns the call. Pick the kind of test your traffic allows. Then copy the plan somewhere your team will see it.
- Test plan
Use when your testing tool can show two versions to random halves of visitors.
TEST PLAN: testimonials on [the pricing page] Hypothesis: showing customer testimonials [under the plan cards] will increase [trial signups per visitor]. Version A: [the pricing page] as it is today. Version B: the same page with a [Quote Spotlight] widget [under the plan cards]. Nothing else changes. Primary metric: [trial signups per visitor]. Decided in advance; other metrics are for context only. Guardrail: watch that nothing downstream gets worse, such as refunds or support requests. How long: run for whole weeks so every weekday is counted equally, and until your testing tool reports the sample size it needs. Do not stop early because one version is ahead. Decision owner: [owner name].
- Results log
RESULTS: testimonials on [the pricing page] Dates run: Visitors in each version or period: [trial signups per visitor] in each: Difference, with the confidence your tool reports: Anything unusual during the test: Result: better / no clear difference / worse
- Decision note
DECISION: testimonials on [the pricing page] We tested a [Quote Spotlight] widget [under the plan cards] against [trial signups per visitor]. Outcome: What we are doing next: What we would test next: Signed off by [owner name].
The plan is written in your browser. Nothing is sent.
A fair test versus a story you tell yourself
Easy to fool yourself
- Changing the headline and adding testimonials in the same week
- Checking every morning and stopping when it looks good
- Choosing the winning metric after seeing the numbers
- Comparing a holiday week with a normal one
Worth trusting
- One change, everything else held still
- A stopping rule written down before day one
- One primary metric chosen in advance
- Whole weeks compared with whole weeks
In StoryProof
Run it with StoryProof and your analytics
- 1
Pick the testimonials that fit the page
Choose the few approved testimonials that speak to that page’s visitors and add only those to one widget.
- 2
Choose a layout that suits the placement
Every layout has a live preview on the widgets page, so you can match the format to the space before the test starts.
- 3
Add the embed code to version B only
Paste the widget’s two-line embed code into the variant your testing tool serves, or into the page at the start of your after period.
- 4
Read the result in your own tools
Your conversion metric lives in your analytics or testing tool. StoryProof’s analytics dashboard, included on every plan, is useful context alongside it.
What happens next: Record the result in the log whichever way it goes. A test that shows no difference is still worth knowing, and points you at a better placement or better testimonials.
Before you start
Honest limits
No promised uplift
Testimonials help on some pages and do nothing on others. That is why you test.
The split happens in your tool
This plan assumes your own testing or analytics tool decides who sees which version and counts conversions.
Low traffic means weak evidence
A before-and-after comparison can suggest a direction. It cannot rule out everything else that changed.
One page is not every page
A result on your pricing page tells you about that page and those visitors. Run a fresh test before assuming the same widget helps a signup form or a landing page.
FAQ
Questions people ask
Short answers to what usually comes up next.
How long should a testimonial test run?
In whole weeks, and until your testing tool says it has enough visitors to call it. The right length depends on your traffic, so there is no single number.
What should I measure?
The action the page exists for: a signup, a booking, a purchase. Pick one before you start and treat the rest as context.
What if the result is no difference?
Try a different placement or more relevant testimonials before concluding proof does not help. Then test again.
Can I test two widget layouts against each other?
Yes, as long as the testimonials inside them are the same. If you change the layout and the quotes at once, you will not know which one made the difference.
Does a slow-loading widget skew the result?
It can, if version B loads noticeably slower than version A. Check the page speed of both versions before the test starts so a delay does not end up counted as the effect of the testimonials.
Still have a question? Ask the StoryProof team or book a demo.
Build the widget for version B
Choose the testimonials, pick a layout and get the embed code for your test.
Last updated 2026-11-04.