Reading Time: 3 minutes

Product recommendations are known to increase ecommerce sales, but how many products should you recommend and what’s the best way to design that interaction? The answer might surprise you.

Watch this episode of The Cro Show, a game show for conversion rate optimization and marketing experimentation fans, and see if you can guess which test variation performed better during a recent Cro Metrics client experiment.

Subscribe to the Cro Metrics newsletter to get future episodes of The Cro Show sent to your inbox.

Full Transcript:

Cristi Alvarez: All right so, for context, this is a shoe and apparel brand, um, and we had done some previous testing with them. And obviously e-commerce, um, we had done some previous testing on their product display pages around a carousel that would have like a shop by style carousel so that people could kind of navigate through and saw success with that when we placed it underneath these product information accordions.

So this right here is a variation. This was the previous control with no shop by style. And we saw about a 5% lift in transactions.

The client, however, didn’t really like the placement of where that was living and wanted to kind of future proof placement of this type of carousel on the PDP or product display page. And so they wanted to see if, um, expanding it out and placing it underneath the actual product images on the page would help to also, like both the future-proof the placement of the content since they don’t really like the placement of it here.

But then we also wanted to see if this would help kind of increase conversions because it’s getting people to the styles that they potentially are looking for. And they have a little bit- a couple more options that they can browse through versus just a single tile or card being shown to them on the control experience. Any questions?

Grant Tilus: This is the same amount of images in the carousel, Christie?

Cristi Alvarez: It is, yes. It’s the same exact amount of images in the carousel. You just have to scroll obviously a bit more on the control whereas on the variation you have about five displayed in front of you.

Grant Tilus: And this was desktop only?

Cristi Alvarez: This one was desktop only correct.

Chris Neumann: I guess I kind of feel like the context is important in fashion, so you sort of know how to wear it, all that stuff. So I like this test going to be one.

Grant Tilus: Well, let’s go with thumbs up If you think V1 one’s a winner, thumbs down. If you think uh, control in the original experiment is the winner.

Cristi Alvarez: All right, You guys ready?

It’s actually the control that won. So we did see an increase in engagement with the test. There was a ton of clicking happening within the actual shop by style carousel here, you know, saw like 194% lift there. We did see an increase in like PDP page views. So people going to the product display pages but then when it came to the down funnel metrics. So add to carts both on PDPs then site-wide cause you can go to collections pages also an add to cart and transactions and revenue saw some dips in those conversions there.

We’re thinking about now like obviously the placement is one thing but potentially what could have happened here too is just putting too many options in front of people. I’m kind of pausing a little of option fatigue there. We could play around with how many styles were actually showing at once as a potential iteration or just the specific content that’s within here, you know, showcasing different content. As we get into more personalization efforts we can start to personalize. What’s actually being shown there based on the type of visitor that’s there.

That’s kind of what we’re thinking as far as follow-ups from here.