Stock photos of people on a subscription ZIP page lost to the control
Test run byMarketingExperimentsResearch publisher
Commentary byLiftArchiveOur analysis, not theirsSwapping a simple newspaper image for stock photos of people reading lowered subscriptions on a home delivery sign-up page; both people-photo versions lost to the control.
MarketingExperiments (MECLABS) ran the test for a newspaper selling home delivery subscriptions.
The two treatments, showing people of different ethnicities, performed the same as each other.
- Subscription rate: the control beat both people-photo treatments. Exact figures were only shown in an image and are not shared here.
- Treatment vs treatment: no difference between the two photos, where the ethnicity of the people was the only variable.
- Negative result: adding people to the image did not help this sign-up step.
The change
Treatments swap the newspaper image for people reading, with headline and CTA layout changes.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
The ZIP code entry page, the first step of home delivery sign-up, showed a simple image of a newspaper on a welcome mat.
Variant
Two treatments used stock photos of people reading the paper: one showed an African American man, the other an older Caucasian couple. The headline and call-to-action layout also changed.
Why it likely lost
Our reading of the result, not the practitioner's claim.
- The newspaper on a doormat shows the product and the promise, delivery to your door, in one picture.
- Stock photos of people may feel generic and pull attention away from the ZIP code field.
- The headline and call-to-action layout also changed, so some of the loss may come from those edits.
Caveats: The image was not the only change; the headline and call-to-action layout also differed, so the loss can't be pinned on the photo alone. Numbers were shown only in an image, and sample size was not shared.
Who ran it
MarketingExperiments ran this test for a newspaper selling home delivery subscriptions and published the results (MarketingExperiments). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on MarketingExperiments →
How to run this test on your site
- Hypothesis
- If we replace our product-focused image on the first sign-up step with photos of people using the product, then sign-ups will change because the image affects how clearly visitors see the offer.
- Primary metric
- Completed subscriptions from visitors who reach the first step
- Guardrails
- ZIP code (first step) completion rate
- Time to complete the step
- Who should test it
- Subscription and delivery businesses with a short, multi-step sign-up. Keep everything but the image the same so you can tell what caused the change. If your current image doesn't show the product or the service, the result may differ.
Questions
Do photos of people increase conversions?
Not always. In this test, stock photos of people reading lost to a simple image of the product on a doorstep.
Does the ethnicity of people in marketing images affect sign-ups?
In this test it made no difference. The two people-photo versions performed the same, though both lost to the control.
What image works best on a sign-up step?
One that shows what the customer gets usually makes sense to test first.
Here, a paper on a welcome mat said 'home delivery' clearly.