Providing the right kind of product images is essential to users’ ability to understand key visual details and make a purchase decision about a particular product.
During Baymard’s large-scale usability testing across desktop and mobile, we observed that, for certain products, simple “Cut Out” images of the product against a while background are simply not enough for users to get a sense of their physical qualities.
Specifically, products designed to be worn — apparel; accessories such as bags, jewelry, or watches; and cosmetics — require the context of a human model in order to get the truest sense of the product.
Without this context, users must estimate the suitability of a product based on other less helpful images and product information, resulting in lower confidence and lower likelihood to move forward with purchasing.
In this article we’ll discuss the test findings from our large-scale UX research related to “Human Model” images. In particular, we’ll discuss:
When shopping in-store, users have the luxury of physically trying on apparel or sampling makeup to better understand the product’s qualities and if the product will work for them — which obviously isn’t possible when shopping online.
When trying on apparel in-store, for instance, users can better judge the product’s length, tightness, and overall fit, which are difficult to gauge from online images.
It can therefore be impossible for users to gain the same level of understanding when considering products on an e-commerce site, without which they may lack confidence in purchasing the product.
To best approximate the information users would gain from trying on a product in-store, sites can provide images of the product being worn on a human model. During mobile and desktop testing, having “Human Model” images was observed to be key for three major product types: apparel, accessories, and cosmetics.
Some products require “In Scale” images for users to get a sense of the product’s size and scale. For example, a stand mixer might be shown on a kitchen counter next to a stack of cupcakes, helping users imagine how large the mixer is in real life.
More specifically, for products designed to be worn, such as apparel and bags, the appropriate basis for comparison and understanding the product’s scale is naturally the human body.
When such products are displayed only “Cut Out”, without a human model for reference, users cannot accurately determine critical product attributes, such as the product’s fit and length — how large is this backpack, how fitted is the shirt, or does the skirt fall above or below the knee?
“Without “Human Model” images, it’s left up to the user’s imagination to try to picture how a product might look when worn.”
Likewise, without the context provided by a model, finer details of proportion and quality can be nearly impossible to estimate — how high is the rise of these jeans, or how does the fabric of this coat drape?
Images of the product isolated from being worn fail to provide the context necessary to determine how the product will look when worn — the product’s primary use case.
Some sites may provide some explanation of these product details within the product description. However, most users rely on product images for their first impression of the product and may disregard it based solely on details they can glean from the product images.
What’s more, many users neglect to read product descriptions in detail, while some overlook them altogether; if key product details are only provided in text form, many users will miss out on this product information and thereby have a more limited understanding of the product.
Product videos are also a poor substitute, as the majority of users don’t engage with this kind of content on product pages.
Displaying model images can also help provide additional context about the product’s primary customer — for instance, if a sweater is meant for children, teens, or adults, or if a tee shirt is meant for women specifically or is unisex.
Without a “Human Model” image, users must rely on peripheral details for this information, such as breadcrumb trails or the product size selector, and might misinterpret whom the product is designed for, wasting time considering an inappropriate product.
Combined with other rich product images, “Human Model” images can furnish a more comprehensive picture of the product and help users understand its visual qualities, which help them decide if the product is suitable for them.
When looking for “Human Model” images, users strive for a general understanding of qualities of length, scale, and fit.
When the apparel depicted on models is styled — for example, tucked in, rolled up, or pinned — these critical product attributes can be obscured, making it difficult to get a true overall sense of the product. Ideally, at least one basic “unstyled” image of the product being worn on a model should be provided to help users gauge product attributes such as length and fit.
Introducing elements of “styling” does, however, create opportunities for promoting coordinating products by showing them as part of a complete outfit. Such images can also thereby promote “supplementary” cross-sells, inspiring users by showing coordinating styles they might wear for different occasions, such as going to work or a formal party.
Finally, when it comes to using the human body as the frame of reference for apparel and other products, some sites may occasionally rely on artificial “human models” in lieu of a real person.
However, users during testing harshly judged images that appeared “cheap” or poor quality, including when products were displayed on mannequins; thus, mannequins or virtually rendered “models” should be considered only as a last resort, with real human models used whenever possible.
The other main product type where human models were critical during testing was cosmetics. Without the reference of a human model, users are less able to visualize how the product will look when worn and may be hesitant to purchase.
The ability to see cosmetics such as lip colors or eyeshadow on a human face helps users have a better understanding of how different product variations look when worn, helping bolster their confidence.
Users are often wary that particular product variations can look very different when worn versus when depicted only by a swatch. With cosmetics in particular, key qualities such as opacity or sheerness, shine, and finish can appear vastly different when worn versus depicted in a swatch or in the packaging.
Without the context of a human model, some users may be skeptical that the color depicted by the swatch will be true to the actual product — especially how the color will look on their particular skin tone.
Furthermore, testing revealed that for cosmetics it’s important to show the cosmetic where it will actually be worn in real life — it’s not enough to simply display different shades of lipstick on an arm, for example.
During testing, we repeatedly observed that such “Product Variation” images were necessary but not sufficient when it came to helping users decide among colors for makeup products.
Once users had used the product variation image of lipstick shades, for example, to compare available colors, they then specifically looked for their color of choice being modeled on a face to get a better sense of how it would look when worn.
Failing to provide an image of the cosmetic where it will actually be worn will leave some users without sufficient visual information to decide if the product will suit them.
Some apparel sites go a step beyond providing only some “Human Model” images and display products on multiple models of different sizes and body types.
Combined with text descriptions to specify the model’s measurements and size he or she is wearing, providing a diversity of models can help users better envision how the product will look on their body by finding a model of a similar size.
While shooting a product on multiple models can be a significant added expense, this can be a major benefit for products where fit is a critical concern.
It can also serve as a unique brand differentiator, as this enhanced offering remains uncommon among apparel sites.
However, it should be noted that no users during testing were deterred from purchasing apparel due to a lack of diversity in model body types, so providing multiple models for apparel and accessory sites should be considered an enhancement and not a necessity.
In contrast, for cosmetics we observed multiple users reject products when they could not find an image depicting a model with a similar skin tone to their own.
With cosmetics, users are not using a model as an “In Scale” comparison, but are rather concerned with the model’s complexion.
To serve as an appropriate basis for comparison, users must have a skin tone that is somewhat similar to the model being shown in order to get a sense of whether or not the color would be suitable for them. One user stated, noting the model did not share her skin tone,
Ideally, each product variation should be viewable on multiple models of diverse skin tones to provide this comparative information.
However, this can present a sizable expense for sites where cosmetics isn’t their only product type, and given the likely abundance of variations for many products. As doing so for all variations of all products is likely infeasible for these sites, consider providing multiple model images for the most popular products.
Alternatively, crowdsourcing user-generated images through user reviews can be considered as a less-expensive way to cultivate a diverse collection of “Human Model” images.
When shopping online, users cannot physically try on or sample products like apparel, accessories, and cosmetics to get a true sense of whether the product will work for them.
Only with the context provided by seeing the product worn on a human model’s body or face can users come close to this in-store experience, gaining a better understanding of the product and feeling more confident in purchasing it.
Without “Human Model” images, it’s left up to the user’s imagination to try to picture how a product might look when worn — which for many users is simply not enough to make them feel confident enough to purchase the product.
(Tip: see 239 examples of product pages showing how many of our benchmark sites use “Human Models” in product images.)
This article presents the research findings from just 1 of the 580+ UX guidelines in Baymard Premium – get full access to learn how to create a “State of the Art” product page user experience.
Join 25,000+ readers and get Baymard’s research articles by RSS feed or
Topics include user experience, web design, and e-commerce
Articles are always delivered ad-free and in their full length
1-click unsubscribe at any time
© 2021 Baymard Institute US: +1 (315) 216-7151 EU: +45 3696 9567 firstname.lastname@example.org