Our client lampenlicht.nl operates in a crowded marketplace with thousands of lamps to choose from. Their product pages were already rich with images, reviews, descriptions, and technical specifications—yet something wasn’t clicking.

In our quest to boost Add to Cart rates and revenue per visitor, we dove deep into the data. Hotjar heatmaps and Google Analytics revealed intriguing patterns: visitors were obsessively clicking the “Show More” link in product specs, and filtering primarily by Fitting, Material, Style, and Shape on category pages.

Armed with these insights, we launched four different A/B tests:

Test 1: Highlighted specs matching the filters visitors had just used

Test 2: Auto-expanded specifications—no more clicking “Show More”

Test 3: Swapped the specs section with product descriptions

Test 4: Added a specs link above the image carousel for quick access

Plot-Twist – Surprisingly all 4 tests showed significantly lower add to cart and revenue per visitor. None beat the control.

We then returned to the drawing board and sought context for the data through User Research.

We then returned to the drawing board and sought context for the data through User Research. Actual user interviews where we asked the following questions:

Which attributes are important when choosing a lamp?

Feedback:

• For prospects, searching for products in a specific style seems to be an important entry point.
• Good combination of detail photos, photos from different angles, mood/lifestyle shots, and an image showing dimensions.
• Fairly general specifications like: cord length for pendant lamps, dimmable options, type of light, whether fixtures are included or not.
• Not important: technical information like light temperature or materials – visitors struggle to understand these characteristics.
• The ‘In Stock’ label and accompanying text “ordered before 8:00 PM, delivered tomorrow” is a very powerful purchase driver. Users notice this. (Note: besides functional information, this also serves as a USP).

Is the product information you find understandable?

Feedback:

• Additional clarification using ‘i’ tooltips is needed for certain terms like IP rating.
• Users indicate that some terms mean little to them, e.g., light temperature.

Can you easily find the product information that matters to you?

Feedback:

• To the right of the image, visitors want to see a list of key features. Specifically about the product. Currently, general USPs are shown here which aren’t relevant to the specific product being viewed. (Note: This is only a solution for desktop, not mobile).
• Users struggle to find the specifications.
• When specifications are found, users often click the manual download link instead of the “show all specifications” link.

Following the feedback from user testing, we took another look at the data in analytics:

• Which filters are frequently used and align with the feedback from user tests:
• Which filters describe general functionalities?
• Which filters describe important product attributes that can’t be directly derived from an image (so not things like color, material, or shape)?

The following filters/product attributes stood out:

• Type of light (e.g., indirect light, ambient light, reading lamp) at 6.88%
• Fitting (e.g., E27, E17, G7, etc.) at 7.67%
• With/without included bulb at 0.10%
• Dimmable at 0.70%
• Suitable light source (which type of bulb is suitable: LED, Halogen, Incandescent) at 6.5%

This revealed a clear discrepancy between analytics and the user research results. Although some filters might not have been used frequently on the category listing page (compared to other filters), these product attributes turned out to be quite important on the product page itself. But now we have everything we need to make an effective hypothesis.

If one crucial product feature or USP is shown textually below each product image from the second product image onward (in the carousel), the number of transactions will increase, since buyer information needs will be better served in a location that is already frequently viewed.

A script was run where specifications or attributes that mattered to prospective buyers (as sourced from User Research and Google Analytics) were formulated into statements and placed on the carousel images from the second image onwards.

In order to make sure we had a clean data set we set another activation condition for the experiment. Which was that the experiment could only start when the visitors started browsing through the product images and the 2nd carousel image entered the browser’s viewport. This was the cue for Convert Experiences to serve the variant or the control.

Once visitors saw the USPs below the images, they could not stop scrolling through them, as they became even more informative.

It became clear that image carousels without further context (attribute and spec data) are very similar to a comic strip with empty speech bubbles. Readers fill them in with their own stories, which may not serve your goal of driving more sales.

With the added USP’s the image carousels becomes a sort of elivator pitch where in X amount of slides the key features of the product is explained.