Social Media Research

New Research: Do Pictures of People Increase Facebook Engagement?

badge guest post FLATTER New Research: Do Pictures of People Increase Facebook Engagement?Marketers 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.

 New Research: Do Pictures of People Increase Facebook Engagement?

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:

 New Research: Do Pictures of People Increase Facebook Engagement?

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.

 New Research: Do Pictures of People Increase Facebook Engagement?

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.  New Research: Do Pictures of People Increase Facebook Engagement?

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

 New Research: Do Pictures of People Increase Facebook Engagement?

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?

 New Research: Do Pictures of People Increase Facebook Engagement?

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.

 New Research: Do Pictures of People Increase Facebook Engagement?

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.

 New Research: Do Pictures of People Increase Facebook Engagement?

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 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).

Related
  • http://MerlinUWard.com/ Merlin U Ward

    Great research and very insightful! One question on your methodology. When calculating the percent, did you use the Page Likes at the time of the image posts (historical page likes) or the Page Likes on July 31?

    • TAGGS

      Sharp observation, Merlin! We used the page likes as of July 31 for each brand in the study.

  • http://www.buildandbalance.com/ Michael_N

    This was a fascinating post. I can’t recall ever reading one on this topic. Thanks for shredding new light on how to use images on Facebook. My mind is already humming with ideas like showing a graduation certificate alone without a graduate holding it as I’ve done in the past.

    • TAGGS

      Michael, Glad to hear that the article got the wheels turning!

  • http://msileanespeaks.com/ Ileane

    HI Mark,
    Very interesting. I’d be curious if this data is only reflective of Facebook engagement or if you’d see similar results from Pinterest. Thanks for sharing the study.

    • TAGGS

      Ileane, our early research with other social platforms indicates that it’s not advisable to apply findings of images and user engagement from one platform to another. We’re beginning to see differences in content and engagement across platforms on topics related to the use of people, brand logos, and CTAs, though honestly it’s too early in our work to tell exactly those differences.

      • http://msileanespeaks.com/ Ileane

        That makes sense. For one thing, even the optimal image size for each platform is different which makes our job even more difficult. The other thing is, just when we get it all sorted out – they change it. :)

        Have you done any research in terms of engagement of images vs. video on Facebook? That one is a tough nut to crack as well. I heard that uploading video directly to Facebook was better than posting a YouTube video but I’m not too sure about that. If you have any insights to share I’d really appreciate it. Thanks so much!

        • TAGGS

          Hi Ileane, we’ve not done research images vs video on Facebook. I’ll keep in mind for future research though.

          • http://msileanespeaks.com/ Ileane

            Just to be clear, what I meant was – any difference in video uploaded directly to Facebook, versus sharing YouTube videos.

            Thanks!!

    • http://www.sierratierra.com/ Lisa Kalner Williams

      Hi Ileane,

      I know that Curalate recently did research on pins with “thing” v. “thing with human.” They found that brand images without faces get 23% more repins that images with faces. Don’t know about partial body parts like mentioned here with Facebook though.

      Here’s the link to the Curalate study/infographic: http://blog.curalate.com/the-perfect-pinterest-pin-how-to-optimize-your-images-for-maximum-engagement-on-pinterest/

      • http://msileanespeaks.com/ Ileane

        Thanks so much Lisa. Great article.

  • Billy

    I would like to see this study done for non-profit or causes. It’s understandable, yet sad, that all this research is about getting people to buy, buy, buy, not necessary give, share, change. They can be related, but takeaways like “visualize wearing or having an advertised product,” aren’t helpful.

    My work has shown consistently that real-life snapshots of faces — not just persons, but more specificially close ups of faces — far out perform anything when it comes to getting people to click for a cause.

    • http://www.convinceandconvert.com/ jaybaer

      Good idea on follow up research project.

    • TAGGS

      Billy, great idea for further analysis. I’m glad to hear though you’re already tracking how different visual content engages your audience.

  • http://msherrwhenonline.com Ms. Herr / @MsHerr

    I agree with the ponderings about how this would play out across other industries. I expect that images without people would perform well in auto and technology. I’m curious what this would look like for Red Bull, where athletes are as much assets as Victoria’s Secret models. Red Bull has drawn such tight alignment between the brand and action-heavy sports, it would be hard to illustrate that action without having people in the photos.

    As for media sites, where customers are not buying products but consuming content, I expect that images should align with the type of content covered.

    • TAGGS

      Ms. Herr, We’ve got a lot more work to do to tease out the differences across all industries, but you’ve given me some good ideas for future research. Thank you!

      • http://msherrwhenonline.com Ms. Herr / @MsHerr

        Happy I could help triggers some ideas! I work for RightThisMinute (http://www.rightthisminute.com); we’re a TV show and media organization tapping people into web videos. In many ways, we enable voyeurism. Our fans and followers can spectate, so my gut would tell me that having people in the picture is important.

  • Oremo Ochillo

    I agree with this post as it mirrors some of my own social media experiences. Especially when working with restaurants. We have found that when you just show pictures of the food item, it makes it easier for people to envision themselves eating the food so they are much more likely to comment about their thoughts and share the content.

    • TAGGS

      Thanks for sharing, Oremo. We’ve met many marketers that have similar experiences as you do about food images.

  • http://theryancox.com/ Ryan Cox

    This is super-bad-ass research Mark. I found it to be extremely interesting and for retail store or restaurant brands — research like this should be your gospel. I’d love to see the same context of research for Twitter, Google+, and Pinterest. Additionally, I’m interested to expand and explore the findings from this research for service-based businesses.

    • TAGGS

      Ryan, I’m honored for the super-bad-ass rating. More research to come for Facebook and other platforms. Hope it’s all just as interesting. Cheers!

  • Kohinoor Devroy

    Interesting to know how a part of a human body is visually more appealing than a full body in communication & engagement

    • TAGGS

      Thanks for the comment, Kohinoor. Glad you found it interesting!

  • Graciousstore

    I agree with the hypothesis that people tend to be more engaged morewith brands that do not have images of people on their brands

  • michael bian

    Absolutely brilliant and simple post. Thanks for sharing it.