A/B Test - Does Default Product Display Order Matter for Ecommerce?

Challenge

For a product listing page (PLP), users are typically offered multiple options via a dropdown for choosing how they would like to display products to make it easier for them to find the right one. Here is an example of the PLP at Targus, with an arrow showing where the sort order dropdown is located.

When a user clicks that dropdown, they are offered several common sorting options which, at Targus, included:

• Alphabetical, A to Z

• Alphabetical, Z to A

• Lowest Price

• Newest

• Best Selling

• Featured

While some users engage with the dropdown, not all of them do. And if a user can select their desired sorting option, does it matter what the default sort option is and if so, and to what degree does it matter? We ran a test to find out.

Solution

First, we intentionally removed the alphabetical sorting options from our test. We already knew that they had almost no usage and studies show similar behavior on other ecommerce sites and they even recommend against alphabetical sorting as an option altogether.

Secondly, we isolated the test to backpacks, the most popular product category on Targus.com.

Since we had been using “Featured” as the default display order, here is how we designed our test, using an even 25% traffic split to each variant.

• Control (A) – “Featured”

• Test (B) – “Lowest Price”

• Test (C) – “Newest”

• Test (D) – “Best Selling”

We monitored two metrics for this test:

• PLP > PDP Click Rate – Percent of PLP visits that result in a click to a Product Detail Page (PDP)

• PLP Conversion Rate – Percent of PLP visits that result in a purchase

For fun, we also placed some friendly coffee wagers on which option, if any, would emerge as the clear winner in terms of the conversion rate metric. I went with “Lowest Price” as my predicted winner.

Results

Although “Lowest Price” outperformed “Featured”, it fell short of beating out Test (D) – “Best Selling” as the top default sorting option across both metrics we were monitoring.

• +27% improvement in PLP clicks to the PDP when “Best Selling” was the default sort option

• The “Best Selling” test scenario also resulted in a +21% improvement in PLP conversion rate

Takeaways

We had several thoughts on why “Best Selling” won and beat out other options including “Lowest Price”:

• Social Proof – A list of best sellers can act as a form of social validation (“if others like it, I would probably like it too”)

• Narrow price band – Products in the backpack category were all within a narrow price range where average price variance was under $20. For a category where users are more likely to exhibit price sensitivity (i.e. laptops), we may have seen different results.