How Does Product Copy Format Affect User Engagement? [Original Research] Institute Revenue optimization services Blog Search Start 7-day trial for $1 Subscriptions Minidegrees Sprints Online courses Free courses Resources

How Does Product Copy Format Affect User Engagement? [Original Research]

How Does Product Copy Format Affect User Engagement? [Original Research]

Product page copy is something that, generally, gets ignored.

Not just the content, but how it is formatted.

In this 3 part study from CXL Institute on eCommerce product pages (part 1 here, and part 2 here), we wanted to explore how elements of a product page affect users’ visual and value perceptions.

This experiment looks at how users view a page and read product text descriptions when the text format changes.

Results summary

  • Text format changes how people read text descriptions.
  • ‘Techy’ product classes (i.e. spec-driven things like hard drives) are influenced the most by text format.
  • Generally, people read bulleted text for a longer amount of time than paragraph text. Though this pattern depends on the product class.

How do I apply this research?

  • If you sell ‘tech’ type products (e.g. electronics like printers, hard drives, routers, etc.), then text formatting matters a lot. Test it.
  • If you sell ‘design’ type products, test a smaller image size. Longer copy does get more attention, but format matters less.

Full study setup description

This experiment was part of a 3-part series exploring the layout of e-commerce product web pages, specifically:

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


We test these hypotheses across product ‘classes.’

Product classes vary according to the commonly used ‘experiencesearch’ product classification 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.

This ‘class’ range can also be characterized on a ‘design-spec’ classification, where the shirt represents a design product, the hard drive a ‘spec’ product, and the headphones again represent the hybrid.

Product xamples ranging from shirt to headphones to hard drive

Research Cheat Sheet

Get a Roundup of Our Most Insightful Original Behavioral Science Research
  • This field is for validation purposes and should be left unchanged.

Full Study – Visual Perceptions & Text Formatting

The formats we tested included:

  1. Paragraph format.
  2. Bulleted format.
  3. No specifications at all.

Note, we retained a small amount of descriptive text that was consistently formatted across treatments and within product types. These variations are illustrated, with the shirt product, in the figure below.

Text format variations of no specification text (left), text in paragraph form (middle), and in bullet form (right). Primary area of interest indicated by red box.
Text format variations of no specifications at all (left), text in paragraph format (middle), and text in bullet format (right). Primary area of interest indicated by red box.

This study was conducted using eye-tracking.

For each combination of product type (n = 3; dress shirt, headphones, hard drive) and specification text format (n = 3; paragraph format, bulleted format, and no specifications at all), we had at least 42 people view the page with the task of assessing the product with the intention to purchase or assess value.

Participants for each format treatment
Participants for each format treatment.

Here are aggregated heatmaps for our treatments, which are interesting to look at, but mainly provide us with a visualization of all the variations tested. What we really are interested in are the stats behind how much time people spent reading the text area, but we settle for how much attention was paid to the text area.

Aggregate heatmap for each treatment
Aggregate heatmap for each treatment.

Here’s a gif of the same thing.

Aggregated heatmap gif for each treatement
Aggregate heatmap gifs for each treatment.

Our primary statistic of interest here is the average fixation time on our primary area of interest, which, in this case, is the text in the right column.

So sticking to product categories, let’s see if format means anything. Here are the raw summary stats:

Summary data for time fixating on the text area of interest for each product type
Summary data for time fixating on the text area of interest for each product type.

And here is a better way to visualize it with comparative ANOVA within product classes.

Histograms of average time fixating on descriptive text for each product type. Connective arrows indicate a significant post hoc Tukey test result, with p-value indicated.
Histograms of average time fixating on descriptive text for each product type. Connective arrows indicate a significant post-hoc Tukey test result, with p-value indicated.

These results suggest that text formatting is important for ‘search’ products or products that are heavily ‘spec’-driven.

The most telling result is, for the hard drive, the mean time for fixation on the paragraph vs. the bullet format is significantly different (p = 0.0199).

The differences between no extra text and paragraph and/or bullet text are not too surprising. Even though we normalized the fixation time to the amount of text, just the fact that there is less to read surely created less of a visual anchor for viewers in general.

Now let’s separately consider the amount of time paid to the text area among the product types. Here are the analogous ANOVA results for testing differences in mean fixation time among product classes:

Histogram of ANOVA results for testing differences in mean fixation time among product classes. Connective arrows indicate a significant post hoc Tukey test result, with p-value indicated.
Histogram of ANOVA results for testing differences in mean fixation time among product classes. Connective arrows indicate a significant post-hoc Tukey test result, with p-value indicated.

Our thinking before the experiment was that viewers would pay less attention to text descriptions on ‘experience’ products compared with ‘search’ products. We get mixed support for this.

The shirt experienced the least amount of text-reading, but the headphones, our ‘experience-search’ hybrid product, saw more than the hard drive.

Considering the image size results from Study 2, the shirt is turning out to be a bit of an odd-ball. Viewers aren’t looking at the shirt very long, and they’re not reading about it much either.

Considering the price valuation results from Study 1, suggesting that the page’s general look and feel might provide a value basis, the results from Study 2 and 3 start to make some sense.

The possible implication is, at the extreme end of  “experience” products, people don’t care as much about the look of the product itself. Rather, they use the look and feel of the page (as a whole) as a basis for judgement and attention.

Everyone has a good idea of how a shirt will fit (you likely know your size, generally). So without prior knowledge of a brand, and how it differs in feel and fit to other brands, the only thing a person needs to process with regard to the product image is the shirt pattern.

The look of the ‘shop’ serves as a surrogate for quality and value. Just a hunch to test another time.


This wraps up our 3-part series on eCommerce product page layout (part 1 here and part 2 here).

In this study (part 3), we learned that product specification copy does matter in how people view the page. Particularly, it matters a lot for “technical” products (spec-driven things like hard drives, computers, and head phones). Bullet-point format tended to garner more attention, so it’s worth testing different text formats and layouts.

Be sure to read the whole series to get a better idea of holistically optimizing the layout of your product pages.


Join the Conversation Add Your Comment

  1. This is a great study Ben!

    I’ve been testing a couple different formats for product copy on Hammock Town.

    Some products you’ll see have the description under the ATC button and fixated to the right.

    Other’s have it on the bottom of the image and centered aligned.

    Nothing reached stat sig yet but eye tracking would be nice on here!

  2. Ben Labay

    Yo Luis, thx! What was your hypothesis for the test? I’d be almost certain that if you wanted people to read the text, then moving it up below the CTA and fixed to the right (above fold) would be the way to get more eyeballs on it. Whether this makes people buy is the question…I’d hypothesize it would. We have a similar test coming out that looks at where to list benefits on a product page for best recall rates….but if you’re doing an A/B test here, maybe I can justify a follow-up study. Wanna share your A/B case study? We can maybe publish together with eyetracking results? Cheers, Ben

  3. I am certain different people want different amounts of technical information as well depending on how technical they are themselves so as you want to do on any other landing page, have that extra technical information clearly accessible and nothing in the way for those who do not want it as your likely sweet spot with this.

    1. Ben Labay

      Hi Corey, yup, there’s definitely a balance with the presentation of information, no matter the product type. Thanks for the comment, Cheers, Ben

  4. This is an amazing study Ben, got a lot out of it.
    Few learnings here I will be taking into the online store I am leading.

    1. Ben Labay

      Thanks for the encouragement Manoj, and if you think of any related studies that you’d like to see let us know, we’re always looking for more test ideas. Cheers, Ben

  5. Hi Ben,

    The articles {1,2&3}are definitely precise in terms of methodology and I like how you presented your data. Really, good job!
    Im working on a (slightly) similar study in online marketing where I compare packaging strategies through content-analysis. I hope you don’t mind if I ask which theory/ies in marketing and/or in communication you based your study; if you have theories to recommend, id appreciate it. Thank you.


    1. Ben Labay

      Thanks for the comment Dodge, we try to relay the theories being tested within the studies themselves. Feel free to reach out with specific questions on any of them though.


Leave a Reply

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

Apply for the Conversion Research Sprint

Applications are open for our first live interactive training program. This program is designed to give you advanced-level skills in conversion research. There are limited spots available.

Apply now

We’re running the CXL Sprint program in four other disciplines:

Sign up here to stay updated with upcoming program dates and enrollment info.