Do Photos of People Improve Ecommerce Conversion?

A recent study by Dianne Cyr, professor at Simon Fraser University, found that online stores that used "friendly faces and culturally appropriate colors" were perceived as more appealing and trustworthy than those that didn’t. Findings were based on lab tests with participants from Canada, Japan and Germany, using eye tracking, interviews and questionnaires.

Showing pictures of smiling humans may make give off more happy vibes - but do they result in more site visitors taking action? (E.g. buying something, signing up for email offers, contacting a sales representative).

The answer is yes, no and maybe.

Testing images

Showing pictures of real artists, rather than thumbnails of art pieces, doubled conversion rate for an online art store.

On the flip side, a landing page test for Overnight Printing conducted by Wider Funnel found 18.2% more people clicked the call to action on the version without “shiny happy people.”

Interestingly, Wider Funnel conducted another test on behalf of Cook Travel and found the version without a photo of people resulted in 23.77% more telephone conversions, but the "people version" received 5.25% more conversions online. The "no-people version" was the hands-down winner in terms of revenue, as telephone customers typically have higher order value.

ClearDebt ran a split test between two images with models and found that the image on the right achieved 33% higher conversion.

Perhaps it's because the woman on the left looks bored out of her tree, or the monster looks extra creepy. Or, it could be that the model on the right is looking toward the call to action (it's well known that our eyes tend to follow where the model's eyes are looking).

The ClearDebt test illustrates how simply testing a page with a human image and without isn't enough. It also depends on which images are tested. If Version A (above, left) was tested against a control with no human face, it may have "lost," or Version B may have won against a control. Tested as one-offs, ClearDebt may have concluded that human images do or don't work. A better approach would be to test a control (no humans) against several images with humans. Perhaps Overnight Printing and Cook Travel can still find images that "win."

Tips and cautions for choosing human images to test

1. Understand the demographics, attitudes and preferences of the target market.

Images of people can produce very strong positive or negative reactions. They can persuade or repel. It's important to consider the impact of showing a young face to an older crowd, a male face to females, a female face to males, etc. You can estimate the demographic distribution of your site with a free tool like Microsoft AdLab Demographics Prediction or a paid service like Compete.com. Invest some time in understanding what works and doesn't work for your target market and/or personas.

2. Remember the call to action.

The benefit of human images are they attract attention - they'll capture the eye gaze. But the problem is they can take away from the visibility of your call to action. Examples (from email campaigns):

Ralph Lauren's email features Mayor Michael Bloomberg. The eye is either drawn to the men or to the key. Do you see the call to action?

Both models are looking in different directions, neither at the call to action buttons below. The image is also framed which may stop the eye even more. Unfortunately, it's important for the recipient to take an action on this email, otherwise he or she may continue to receive Direct to Business emails that are not relevant.

Better to choose a model that looks toward the copy or button you want the end viewer to notice.

And don't take your online design inspiration from the print industry. Print ads are designed for recall, not immediate action.

3. Consider using "real" people.

You can expect George Foreman's smiling grill to resonate better with customers than without, George IS the brand, and his face builds instant trust and goodwill with the online store.

VS.

For the majority of online shops that don't have the luxury of a celebrity brand, consider using customers' and staff photos.

EyeBuyDirect's Wall of Frame showcases the site's own community users, which builds a lot of trust while showing the products in use. Seeing images of customers like them encourages site visitors to use the try-before-you-buy tool, which leads to an emotional investment in the site and a better idea of what the frames would look like on one's face.

Crutchfield was the king of human faces on its home page (including pictures of Crutchfield staff and even its CEO), but today's home page includes the photo of an actual customer accompanied by a testimonial. Whether showing helpful staff or satisfied customers, these images support the Crutchfield value proposition of excellent customer service.

4. Know what to measure.

Before testing, understand what it is you hope to gain by adding human faces. Is it a reduced bounce rate on the home page? Is it more pages per visit? Is it more repeat visits? These are quantitative measures of site trust - even if there is no improvement in conversion (remember, it might be your products and offers that are the problem), these metrics indicate higher interest in and loyalty to your site. You've found a branding/trust "win."

Is your goal to maximize conversion on a landing page or email campaign? Measure click throughs, conversion rates, telephone leads, etc. Don't forget to measure revenue and average order value.

And finally, test a control (no human image) against several versions of human images, rather than just one treatment version. This reduces your risk of concluding "images work" or "images don't work" prematurely.