Skip to main

New Research: Do Pictures of People Increase Facebook Engagement?

Authors: Mark Kelley Mark Kelley
Posted Under: Social Media
Listen to this Article

badge-guest-post-FLATTERMarketers often spend hours selecting and producing visual content to post on Facebook brand pages. Creatives, strategists, and managers can go round-and-around debating which images work and which don’t for a brand. Sometimes they debate over whether or not the brand should show people in brand images, and everyone has their differing opinions.

At Taggs, we decided to bring data to help settle the debate – Do people pictured in brand images help or hurt Facebook engagement?

Our Hypothesis

Brand images without people would be associated with greater engagement than those images including people. The hypothesis is based on our anecdotal observations that when a Facebook user encounters a brand image of a product, lifestyle, or landscape without a person shown, the user is better able to project themselves into the image and therefore more likely to like, share, or comment on the image.

How Does the Presence of People in Brand Images Relate to Engagement?

We used Taggs’ visual content marketing software to index 3,656 brand images published on Facebook since the start of the year. We collected these images from 14 leading Facebook consumer brands in retail and restaurant sectors.

We classified each image as having a person, not having a person, or showing only a part of a person, such as a hand holding a product but without showing a person’s face. We found that over half of the images (54%) published by brands did not include people, and only 41% included a person.

We calculated engagement for each image as a percent of fans so that we could aggregate information across brands. We compared engagement among categories and found the following:

A surprise to us! We had hypothesized that images without a person would have the highest engagement. We were surprised to find that images showing only a part of the body, such as a hand holding a product, had the highest engagement of the three categories. We dove a little deeper into the data and broke out our analysis by post likes, shares, and comments to shed more light on these results.

We found that images showing partial body part earned 29% more likes than images with a person and 10% more likes than images without a person. Likes are, of course, the most abundant and lightest engagement, and this is the only metric for which we found partial body shots beating images without people. It’s really interesting that most of these “hand holding product” shots are really casual images, clearly snapped with a phone camera.

We found that images without a person earn 124% more shares than images with people and 15% more shares than images showing partial body.

We also found that images without a person earn 104% more comments than images with people and 59% more comments than images showing partial body.

Breakdown of People Frequency of all Brands in the Study

Here’s the breakdown of the brand in our analysis, showing the frequency with which they post images with or without people.  Take a close look. Any implications of brand strategy jumping out?

The national restaurants, Pizza Hut, Dunkin’ Donuts, Starbucks, and Subway, all together jump out at me. They all tend to rely heavily on images without people (which are usually promotional images of their food and drink products).

Dunkin’ Donuts images, in particular, only show people about 7% of the time, which is the lowest of any brand here. Also, 11% of DD images are of a body part only, almost always a hand holding a drink or snack product. The overwhelming majority of DD images, 4 out 5 in fact, don’t include people.

Interesting Findings about Engagement and People

Most of the brands we analyzed conformed to our original expectations that images without people in them would outperform images with people in terms of Facebook engagement, especially in the retail sector. Here are some notable findings.

Three brands though showed images with people being more engaging than images without people: Subway, Victoria’s Secret, and Abercrombie & Fitch.

We indexed 130 Subway Facebook images and found that 28% showed a person. We also found though that Subway Facebook images with people achieve 16% greater engagement than Subway images without people.

Subway’s images of people often show one or more of their many spokespeople or customers. These images perform well for likes and comments, but they underperform in earning shares.

For Victoria’s secret, we found that images with people (most of whom are scantily clad Angels models), engage the audience at an incredibly high rate, more than 2:1 over images in which people don’t appear.

Though much smaller than the boost seen with Victoria’s Secret, it’s noteworthy that Abercrombie & Fitch also had highest engagement for images with people, about 9% higher. Again, Abercrombie & Fitch gets engagement help from a squad of super attractive models, especially male models.

Key Takeaways for Visual Content Marketing on Facebook

Here are some of the key findings that will help marketers in creating a solid visual content digital marketing strategy for Facebook:

1. Across all brands we saw that images without people outperformed images with people by about 17%.

In retail, we see some even larger differences in engagement between images with and without people, ranging from +41% for Old Navy to +113% at Kohl’s. These findings suggest that users prefer to see pictures of retail products without people, making it easier for them to visualize wearing or having an advertised product.

2. Causal images that show partial body like hands and feet are associated with higher Facebook likes. However, images without people or body parts entirely earned more shares. If you’re social strategy prioritizes earning shares, keeping people out of the images may improve your likelihood of earning shares on Facebook.

3. This takeaway may seem contradictory to 1 and 2. However, if your retail brand has a unique brand asset comprised of people, such as the notable models at Victoria’s Secret and A&F, then the images of people may indeed help boost engagement.

I think the overarching takeaway is that brands need visual content strategies that take into account unique brand identity, objectives, and audience.

Now it’s time for your thoughts. How do you see people influencing Facebook engagement for your brand images?

The Research Methods Behind the Madness

Using the Taggs’ visual content marketing software, we analyzes 3,656 Facebook image posts from fourteen different brand pages. We only analyzed images published to the Timeline from January 1 to June 31 2013.

The brands (and the number of images indexed) are Abercrombie & Fitch (358), American Eagle (238), Dunkin’ Donuts (262), Forever21 (406), H&M (111), Kohl’s (412), Macy’s (253), Old Navy (245), Pizza Hut (163), Starbucks (39), Subway (130), Target (97), Victoria’s Secret (299), and Wal-Mart (643). The brands selected were based on three factors: that they are in the top 100 FB brands, have a strong US national brand presence, and are a retail store or restaurant brand with physical locations (as opposed to brand like KitKat).

For each image post, we collected the Facebook likes, shares, and comments and classified the image posts has having a person, nor having a person, or showing only a body part such as a hand holding a product but not showing a person’s face.

In calculating engagement as a percent of fans, we used the formula (post likes + shares + comments) / page likes * 1000 (a multiplier to make decimals more manageable in visualizations).

Blog Banner Image

Get our best tips in your inbox! Join the smartest marketers who receive our ON newsletter.

Join the Social Pros LinkedIn Community

Join a community of real social pros doing social media on LinkedIn. Receive all the inspiration and ideas straight to your feed and add your thoughts to the conversation.

Follow Social Pros on LinkedIn

Subscribe to Social Pros Podcast

b2b influencer

Ready to Get Started?

I am looking for...