Reading Time: 3 minutes

Do product upsell offers from direct-to-consumer ecommerce retailers increase revenue as intended or can they actually end up hurting sales? 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 sent to your inbox.

Full Transcript:

Kim Young: This test is for a ecommerce client that deals with direct-to-consumer sales of printers and sewing products and supplies. They also deal with B2B, but this particular test is on their direct-to-consumer side.

In this test we were trying to see if we could coax users to purchase printer supplies along with their new printer purchase.

And so in V0, when a user adds something to the cart, they just get this mini cart up here that summarizes you know the product, the price and they can either continue to shop from here or proceed to their cart.

In our variation, what we’ve done is suppressed the mini cart and pop this modal that offers a selection of related products to the particular item that they’ve added to their cart. So it can be toner, printer sheets, any kind of accessory that could possibly go with the printer.

And functionality of the modal, all items were taken to account whether or not they were in stock. If an item was added to the cart from the modal that was indicated within the modal. And if an item was out of stock, or if they attempted to add an item to cart and it had been out of stock but the site hadn’t quite refreshed yet, then we displayed this error message here.

Any questions, comments, feedback?

Cristi Alvarez: Kim, did this test run just on desktop or did it also run on mobile as well?

Kim Young: Great question, desktop only.

Jamie Brown: Hey Kim, so is the what’s the primary goal here? Like higher AOV?

Kim Young: Higher AOV, yep.

Cristi Alvarez: I’m going to go ahead and say that V1 won. Because it kind of lays out the the upsell options pretty neatly, so I think I’m going to go with V1.

Leila Meidam: I agree especially since items are associated with the product in cart more clearly than some upsells I’ve seen.

Cara Binsfield: Kim, for the, you said, the main goal was AOV, but were you looking at total revenue and transactions as well?

Kim Young: Yeah

Cara Binsfield: So it had to be increased AOV, but in terms of really increasing revenue, right?

Kim Young: Correct and I believe the primary metric on this test was actually cart visits. But ultimately, order value was what we were trying to accomplish.

Cara Binsfield: All right, I’m going to controversial and say that the test variation lost because I’ve seen some upsells in the cart be distracting and reduce total transactions by enough that even with the increased AOV revenues down, nice.

Cristi Alvarez: Nice. Alright, let’s go ahead and vote. So if you think it was the control you’re going to do a thumbs down. If you think the variation won, you’ll do a thumbs up. So go ahead and cast your vote. Alrighty Kim, go ahead and do the big reveal.

Kim Young: Alright, so this was a big loser. Alright, Cara! Big loser we found that cart visits were down, not at stat sig, but we actually ended up pausing this one early because it was such a big loser up front. Average order value was down 3%. Revenue per visitor down more than 10%. And orders, as a whole, down 18%. So, and actually, added recommended product, nobody really added anything to the cart from the modal. So the conclusion we drew here was because users were already adding some really high priced, high ticket items to their cart, that showing those upsells, such as toner, with high values, sort of increased price sensitivity and distracted users.

Paul Swinand: So this would be like if Gillette gave you some info on how much they’re going to charge it for the blades once they give you the razor for free or something?

Kim Young: Yeah. Yes. Yeah so we we hypothesize that once users actually saw how much their future investment would be if they bought particular that particular printer, it sort of gave them pause and they they ended up not converting.