How Product Image Size Impacts Value Perception [Original Research]

How Product Image Size Impacts Value Perception [Original Research]

What makes a good product page?

Well, there are tons of elements that come together to make a successful product page. These include price, product image, product copy, layout, etc.

One element in particular, product image size, seems to affect the value perception of the product. In this study from ConversionXL Institute (part one of three of a full eCommerce product page study, the others to come soon), we look at product page design, and in particular, how you can increase the value perception of your particular product.

Results summary

  • The average perceived value for the large image of the hard drive (search good) was $13.50 more than the smaller image.
  • The average perceived value of the shirt (experience good) was $1 less when the image was larger. In other words, participants perceived the shirt to be higher quality/more valuable when the product image was smaller and there was more white space on the product page.

keke

How do I apply this research?

  • If you sell a spec-heavy search good (e.g. a hard drive, printer, camera, etc.), test a larger image on your product page.
  • If you sell a design-heavy experience product (e.g. a shirt or some other wearable), test more white space on your page (to minimize visual complexity). In our tests, all 3 website variations with more white space resulted in higher perceived value (due to image size and other web elements).

Full study setup description

This experiment consisted of 3 distinct studies aimed at testing hypotheses regarding the layout presentation of e-commerce product pages, specifically:

  1. how pricing perceptions change with image sizes
  2. how people visually perceive the page with differing image sizes (coming soon)
  3. how the presentation format of the specification/description text affects how people visually perceive the page (coming soon)

Schematic of full 3-part e-commerce study components
Schematic of the 3-part e-commerce study components.

We test these hypotheses across product ‘classes.’ The product classes we use vary according to the commonly used ‘experiencesearch’ product classification, an economic theory developed by Philip Nelson. We use a men’s dress shirt as an ‘experience’ product, an external hard drive as a ‘search’ product, and a pair of over-ear headphones as a hybrid.

Screen Shot 2016-04-25 at 10.15.46 AM

Product type can also be characterized by a ‘design-spec’ range as well. The idea behind this range is that some products’ value are derived from their overall design (the shirt) while other products are defined by their specs (the hard drive). The headphones represent a hybrid product since both design and specs contribute to their value.

Product class examples ranging from shirt to headphones to hard drive

Does Image Size Affect Value Perception?

It seems like such a minor thing, but can increasing the product image increase how much people think it’s worth? That’s what we set out to study. Here’s the background of our research…

Hypothesis: Large product images on product pages result in a higher perceived value compared to smaller product images on the same page. In other words, people will perceive a product to be more expensive/valuable when its display picture is large.

We also wanted to understand if this effect was more pronounced in different product categories. So we tested a product from each end of the spectrum, from “experience goods” (products that are hard to evaluate before consuming/using them) to “search goods” (products with features that are easy to evaluate before consumption).

The study treatments included a dress shirt product page and a hard drive product page. The dress shirt test had unexpected results, opposite our hypothesis, so just to add some validity we replicated the test on two additional websites, varying in design characteristics and price point. The dress shirt represents an experience good while the hard drive represents a search good.

For the dress shirt, product image and brand logo remained consistent across all three treatments and acted as control variables. The size of the product image acted as the treatment variable and varied across the three websites. Any prices and indications of price were removed from the pages.

Each treatment (n = 8; 6 shirt & 2 hard drive) were sent to around 300 people (specific numbers below). We asked two pricing questions:

  1. At what price is this product a bargain?
  2. At what price is this product too expensive to consider?

These questions are a modified version of the Van Westendorp Price Sensitivity Meter.

For all surveys, our audience was limited to a United States general population, and participants were limited to only one treatment, resulting in 2,982 total participants. Here are the numbers per treatment:

Participants for each format treatment
Participants for each format treatment

Shirt (a) Variation

Variation one: This website was the original source, except we either enlarged or reduced the shirt image size for the treatment variations. The original price of this shirt was $145.

Original shirt variation

Shirt (a) Statistics

Shirt (a) Results: What’s the Takeaway?

  • The smaller image actually produced a higher perceived value, significantly so for the ‘bargain’ question in a t-test (p-value = 0.012) and nearly so for the ‘too expensive’ question (p-value = 0.117).
    • It’s likely that with additional responses, the ‘too expensive’ question would’ve produced significant results as well. As of now for this ‘too expensive’ question, there’s 94.1% certainty that the smaller shirt image has a higher perceived value (smaller shirt mean > larger shirt, p-value = 0.0587).
  • Not only does this make us reject our hypothesis, it flips it. We now have evidence that a smaller image can not just change but increase value perception (at least for a men’s dress shirt, a clear ‘experience’ product).
  • An alternative hypothesis that we now want to test is whether the effects of simplicity and more whitespace increase value perception more than a large product image can.
    • This is why we tested the two additional shirt variations below, both had ‘busier’ displays.

Shirt (b) Variation

Variation two: Here we’ve modified a Macy’s product page which is generally a busier design (e.g. core colors, text, elements, etc.) than variation one. The price of the shirt originally shown on this page was $52.50.

shirt2

Shirt (b) Statisics

Shirt (b) Results Summary

  • There were no significant results between image sizes for either question. However, the pattern of lower value perception for a larger product image exists.
  • Although not significant at the 0.05 level, there’s a 78.5% chance that the means between the small and large product images are different for the ‘bargain’ question.

Shirt (c) Variation

Variation three: The last variation is a modified version of an Amtify shirt product page. The price of the shirt originally shown on this page was $29.90.

shirt3

Shirt (c) Statistics

shirt_c_stats

Shirt (c) Results Summary

  • Again, there were no significant results among the image sizes for either question. However, the pattern of reduced value perception for a larger product image does exist.
  • However, this one is closer to significance than the Macy’s variation. There’s an 83% chance that the ‘bargain’ question resulted in different means between the image sizes, and a 91.5% chance that the small image mean is greater than the large image.
  • This result shows a stronger pattern compared to the Macy’s variation. This product page is visually less complex than the Macy’s page, which is a theory on why this result is stronger than the Macy’s variation but weaker than variation one’s simple design.

So, what about search goods?

Hard Drive Variation

We modified the product page here.

harddrive

Hard Drive Statistics

Remember, our hypothesis was that the larger image would translate to a higher perceived value and that the ‘experience’ type of product (shirt) would show a larger difference among image sizes compared with the ‘search’ type of product (hard drive).

harddrive_stats

Hard Drive Results Summary

  • This ‘search’ or ‘spec’ product category confirms our hypothesis that a larger product image will increase value perception. And it did so with pretty large numbers, especially compared to the shirt variations.
    • When asked “at what price is this hard drive a bargain?” respondents perceived the product page containing the larger image to be 11$ more valuable than the hard drive product page with the smaller image.
    • When asked “at what price is this hard drive too expensive?” respondents were willing to pay $15 more for the hard drive with the larger product image.
  • The difference between perceived value in the ‘bargain’ question was significant (p-value = 0.039), and the difference in the ‘expensive’ question was nearly so (p-value = 0.053). Pretty strong results here.

Conclusion

What we found was surprising.

We thought that larger images for experience goods (like shirts) would be viewed as more expensive/valuable. However, it was the opposite. People viewed experience goods as less valuable (generally) when larger images were used.

However, it was the converse for search goods (like hard drives). Larger images made people perceive the products as being more expensive/valuable.

Note: this is a 3-part series on e-commerce products. Make sure to check out study 2 and study 3 as they come out in the next week or two.


cxli-logo

Join the Conversation Add Your Comment

  1. I’d be interested to see if these results were the same for women; whilst I understand women buy hard drives, I’m guessing more males purchase them than females.
    Women tend to be better at filtering visual distraction so the results could be very different.

    Reply
  2. Ben Labay

    Good point Gideon, our survey didn’t ask questions about gender…but we should have! We might go back and revisit the survey to send it out again with this in mind, so stay tuned for updates.

    Reply
  3. I also wonder if the shirt website looks ugly with the larger image. From the pictures, I think the image looks disproportional to what you would normally expect and that may explain why the price perception difference. With regard to the hard drive, I feel the design of the webpage itself is not good and therefore, a larger image doesn’t make the page look as ugly. But also compare the small image of the hard disk to the page – there is so much of white space around the small hard disk image that it makes the webpage look weird. To really compare apples with apples, it might have better to use the same webpage to sell both items – that would have reduced one of the variables.

    Reply
    1. Ben Labay

      Hey Ajit, all good points. What provides me some level of certainty on the ‘pattern’ with regards to the shirt was it was seen across all sites…and generally we do come across sites with varying levels of ‘space’ around product images. Cheers, and thanks for the comments!

  4. I would be fascinated to know what effect Amazon’s zoom in feature would have on the perception of pricing. (Presumably the same results and the reason for Amazon being more tech orientated than fashion?)

    I would also like to know the effect on light boxes? In theory they too should be optimised price perception in mind?

    One quick question: where on your scale would you place digital goods such as software?

    Reply
    1. Ben Labay

      Hey Chris, good questions….have to think of how to implement those variations. I would tend to place software in the search good category, as most of the time you’re provided with the info needed or a trial to evaluate the product before purchase. But some digital goods, like video games, are squarely an experience good, as they have to be purchased in order to fully evaluate. That make sense? There’s also a third class, credence that we didn’t get into. This type of good can’t even be evaluated well after purchase…think vitamins, pet food, an oil change…things you just have to have faith in that they are working. Thanks again for the comment! Cheers, Ben

  5. I’d be interested to know if this holds up for other products…I also don’t think the hard drive was a good choice when people already know that you want a hard drive with more ‘space’ so bigger picture = bigger hard drive = must have more space inside…

    As for the shirt, the fabric quality is a very important factor, and from the looks of it the bigger picture shows off it’s crappy fabric.

    Reply
    1. Ben Labay

      Hi Kalki, stay tuned for a follow-up study with other products that also tests across gender (from other feedback). Good point on the hard drive, though most people pay more for smaller form factors…so maybe a wash? As far as the shirt fabric quality, that’s not very easy to determine from the pic….thus the ‘experience’ good type. Also, this shirt is from a high-quality retailer and normally sells for $145, I’d think it was very nice fabric. Thanks for the comment! Cheers, Ben

  6. Interesting study, Ben! Love that you’re investigating other variables besides a general “which image size performs better.”

    Would it have been possible to compare an experience description vs search description for the same product? For example, when researchers investigate utilitarian vs. hedonic goods, they’ll often use the same product (e.g., blender), while varying the types of benefits, such as utilitarian (e.g., cheap price, easy-to-use) or hedonic (e.g., make delicious cocktails). With the same product, it’s a little more valid control.

    I agree with one of the other comments. I think the results were due to processing fluency. With shirts, size is neither beneficial nor harmful. With hard drives, however, size IS beneficial. People want LARGER storage. Even though the physical size is irrelevant to storage size, it doesn’t matter. Due to the underpinnings of fluency, people will misattribute the physical size to the metaphorical size of storage.

    If that explanation is true, it might be interesting to test image sizes in domains where size — metaphorical or physical — is beneficial. I have a hunch that larger product images will perform better. In cases where size is negative (e.g., travel sized products, a medicinal pill you have to swallow, etc.), smaller images would probably perform better.

    Reply
  7. Ben Labay

    Yo Nick, all good comments. We’ll have to test the description on the same product idea…maybe across experience-search-credence good types? Good idea, but it is definitely a separate question from what we were asking.

    You mean perceptual fluency? Though I haven’t thought about these concepts in a while. And I agree with you that size of hard drive is important, but that was the hypothesis, we wanted to get data behind it. Though I’ll also say that people tend to want and pay more for smaller hard drives….though maybe that doesn’t jive with perception (which is interesting to think about).

    To me, the key is the shirt finding generally, and the DATA for the hard drive finding, meaning that it is intuitive but also good to have some data behind it to start up conversations like these and get other test ideas!

    You’ll soon see more and more studies come out like this, and like all experiments, result takeaway extrapolation outside of the specific study design is dangerous. What’s good is the advancement in thinking around the idea, and new hypotheses and test ideas generated.

    BTW, your 125 UI tweak post gave me some test ideas recently! Send over any more ideas if you have em! Cheers, Ben

    Reply
  8. This is interesting and all but how did the image sizes affect ROI? What people say and what people actually do are two very different things.

    Reply
    1. Ben Labay

      Hi Michael, this is additional data that would inform testing to figure out what people actually do. There’s only so many tests people are able to run, better to have tests designed with data-backed theories and hypotheses. This experimentation is meant to help refine and improve hypotheses. To get at what you refer to, we’d need datapoints across hundreds of websites all testing image size within and among product types….we working on it! Thanks for the comment! Cheers, Ben

  9. Electronics and apparel are good examples. How about for the home / organization type solutions like garment racks/folding drying tables/cloth closets/shoe racks/hangers in the under $50 price range. Does image size impact things?

    Reply
    1. Ben Labay

      Yo Tim, we’re getting enough feedback on this study that we’re planning to do a follow-up with different products, we’ll keep your request in mind. Thanks for the comment! Cheers, Ben

  10. Just a thought, but I wonder if the percieved value of the hard drive was affected by the fact that it was a hard drive. When shopping for hard drives, your goal is usually to get the most space for the least amount of dollars. So your mindset tends to be “I want something as big as possible.” Would making the hard drive you physically larger on screen satisfy this at some subconscious level? Might be interesting to try this study with a search product where ‘size’ (of any type) is not the defining factor :)

    Reply
    1. Ben Labay

      Hi Richard, same comment as Kalki’s above maybe. Good point on the hard drive, though it’s even more complex than that as people tend to want large amounts of storage on a hard drive, not a large form factor. That said, we’re expanding this study idea on more products, stay tuned! Thanks for the comment, Cheers, Ben

  11. Sorry if this has been mentioned before, but…

    I’m wondering why you chose a shirt with a pattern which some people (myself included) would consider to be an unattractive shirt — magnify a pattern which a person thinks is unattractive and it likely becomes more unattractive.

    Would a plain white shirt have had the same results?

    Reply
    1. Ben Labay

      Hey Stephen, it was the same shirt for both variations, so the effect of the shirt pattern was controlled for. Thanks for sharing your thoughts on this, Cheers, Ben

    2. My point was that a white shirt would not be any uglier when enlarged, but a shirt with a pattern you didn’t like would be uglier enlarged. Would that not affect the results?

    3. Ben Labay

      So your argument is that a white shirt is more universally liked than a simple pattern shirt? I’d think that people not liking the pattern in the shirt would be a minority…it’s a pretty vanilla pattern, and as mentioned, controlled for. Then the effect you’re talking about is that there’s an INCREASE in ‘not-liking’ due to image size….likely a minority within the minority. And considering the number of people survyed (600+), I’d think that it wouldn’t be a confounding variable. The white shirt would have the same potential issue as some people don’t like white shirts right? Maybe less so, but nevertheless I don’t think this is a big issue here. Might be worth testing though! ‘Experience’-type products will always have this problem I think.

Leave a Reply

Your email address will not be published. Required fields are marked *