More on apples
Mobile optimization in ecommerce
The short version is that all the key metrics went up.
We could only take the story so far, though, since we’re dealing with client sales data. We left out “desktop” (non-mobile) stats, for instance, and this raised some obvious questions. The most frequent of these was, “were the increased mobile transactions cannibalizing desktop transactions?”
Here are those numbers again comparing 3 weeks of control (no mobile optimization) against 3 weeks with typical Responsive Design (RWD) patterns, this time with “desktop” figures.
CONVERSION RATE: + 65.71%
TRANSACTIONS: + 112.50%
REVENUE: + 101.25%
CONVERSION RATE: + 407.32%
TRANSACTIONS: + 333.33%
REVENUE: + 591.42%
CONVERSION RATE: + 20.30%
TRANSACTIONS: + 71.11%
REVENUE: + 41.13%
As you can see, non-mobile conversions went up as well. So, it’s hard to say that mobile gains were simply visitors who were previously browsing on mobile and then later bouncing to a desktop device to purchase. Maybe a small portion of it was. But the data doesn’t support the speculation, and nobody wants a browse-only ecommerce store any way. The more likely path would be that visitors who can’t purchase from your mobile store are bouncing somewhere else entirely. Like Amazon.
This raises another question, though. Wouldn’t the sitewide trends suggest that mobile conversion rate would go up as well by a similar rate even left unchanged? That simply isn’t the case with the retail projects we’re familiar with. Optimizations that effectively ignore the small screen prove to have far less impact on mobile conversions.
This is why we got scolded by that one guy for not running an A/B test. For the record, we do a lot of A/B testing. Mobile optimization doesn’t fail the “is it an obvious improvement?” question, though, so the additional effort didn’t occur to us as practical. That said, we’d love to go back in time and reintroduce the test as A/B, if only for self-gratification.
What we can do, though, is look instead at sustained growth. Mobile conversions continue to outpace sitewide gains, and mobile transactions now represent a significant portion of revenue. That’s saying a lot considering we’ve been doing an increasing amount of post-launch marketing with O’Neill, most of which has a far greater impact on non-mobile traffic.
It wouldn’t be fair to compare figures from the extended Holiday sales frenzy, so let’s look at a comparison of the control versus a similar time span in slower, post-Holiday January:
CONVERSION RATE: + 162.50% ⬆ TRANSACTIONS: + 382.61%
REVENUE: + 370.61%
CONVERSION RATE: + 71.70%
TRANSACTIONS: + 177.03%
REVENUE: + 136.06%
You’ll notice we aren’t parsing device types in these figures. We’ve grouped mobile devices, leaving tablets out of this group since our mobile optimizations didn’t target tablets. As a sidenote, the modifications occurred prior to the release of the iPad mini.
You’ll probably also notice the non-mobile increases are pretty significant considering the control already included the new “desktop” design. Again, these gains are due to our active, post-launch marketing including retargeting, SEO, and PPC rather than simple sitewide, organic growth. This is important to note since nearly all of this marketing (everything except email) targets non-mobile traffic. In other words, the additional cost required to make these mobile conversion gains was really just the cost of the RWD optimization, something that was very quickly recouped.
The story still isn’t complete, but hopefully we’ve filled it out a bit. If you know of any similar tests, we’d love to hear about them. We have more in the works ourselves, so maybe this will even get us blogging again.
Posted on February 22, 2013
You like apples?
“From CSS to PHP — Electric Pulp knows what it means.”