M-Commerce: Terrible UX

2017: Mobile UX Still Bad, but Slightly Better

Adobe Analytics' data from "Cyber Monday" in 2017 is based on an analysis of one trillion visits to over 4,500 ecommerce sites. Desktop computers now only account for slightly more than half of visits to ecommerce sites, but the sites got dramatically less revenue from visitors on mobile devices than they got from visitors coming from desktops. Thus desktop still generates 2/3 of the revenue, which is what matters to business.

The data is as follows:

The ratio between revenue and visits is the most important indicator of UX quality: it indicates how good the user experience on that platform was at converting visits into dollars. (This is a slightly different metric than the traditional conversion rate, which measures how good the design is at converting visitors into shoppers. Looking at revenue per visit takes into account how many people bought something, but also how much they bought.)

We see that tablet visitors bought almost as much as desktop visitors, though slightly less. This is a confirmation of our finding from previous user research that tablet usability is pretty good.

However, the average mobile visitor bought less than half of the spending by average desktop visitors. The ratio between sales-per-visitor for desktop vs. mobile is 2.11, meaning that each desktop visitor is worth 111% more than a mobile visitor.

There is no real reason to believe that desktop visitors are richer than mobile visitors. Indeed, today a desktop computer is cheaper than the newest generation of smartphones. Thus, I believe that the lower sales on mobile are caused by a worse user experience than the one offered up to shoppers visiting from desktop computers.

How to fix this problem: a first step would be for websites to follow the design guidelines for mobile user experience in the design shown to mobile visitors of an ecommerce website. This would be true for all websites, of course, because there is no reason to believe that non-selling sites are any better than ecommerce sites — we just don't have as explicit a way of quantifying the business value of different designs for those sites.

Ecommerce sites should also be more meticulous in following the UX guidelines for ecommerce design. Doing so would improve sales on both desktop and mobile, but probably generate a bigger lift on mobile. My reason for this belief is that mobile devices are inherently a weaker user-interface platform than desktop computers in most aspects of usability, particularly in terms of screen size and text input. (Mobile has its advantages, of course, such as a built-in camera and location awareness.) The weaker the platform, the higher the usability impact of getting the UX absolutely right.

As an example of this latter point, consider the display of a product list. On a typical desktop screen, the user might be able to see 6 products at a glance (above the fold), whereas a mobile-phone user might only see 3 products. (The actual numbers obviously depending on the specific design and the size of the two screens.) Thus, it's dramatically more important for the site to get the prioritization of the product listing right on mobile than on desktop. Similarly, good and clear product photos are even more crucial on mobile where they will appear as small thumbnails — whereas a desktop page design should allow for bigger and more detailed photos.

4-Year Trend: Desktop UX Superiority Dropping, but Still Big

Here's a list of the superiority of desktop user experience vs. mobile user experience during a 4-year period:

(I'm not analyzing the tablet data, since the conclusion from all 4 years is the same: tablet UX is almost as good as desktop UX. The small differences between the years are uninteresting, relative to this big-picture conclusion.)

From the 4-year table, we can conclude three things:

• Mobile UX continues to improve, with the superiority of desktop computers getting smaller every year.
• The big improvement happened from 2015 to 2016; improvements both before and after have been smaller.
• Mobile is still far behind desktop.

2016 About the Same as 2017

The comparable data from 2016 (also from Adobe) is as follows:

In 2016, the ratio between revenue and visits was 143% better for desktop visitors than for mobile visitors.

Comparing 2017 with 2016, we see that the mobile user experience has improved a little during the year, changing the superiority of desktop vs mobile from 143% to 111%. This is a nice improvement, though not as much as seen in previous years.

2015: Mobile UX Slightly Improved

IBM's Watson Trend Report for "Black Friday" 2015 (November 27) shows a slight increase in the performance of mobile ecommerce sites compared with the miserable statistics from 2014:

• Mobile (smartphones): 44.7% of traffic and 20.6% of sales = ratio of sales-to-visits of 0.46
• Tablets: 12.5% of traffic and 15.5% of sales = ratio of sales-to-visits of 1.24
• Desktop: 42.7% of traffic and 63.8% of sales = ratio of sales-to-visits of 1.49

One interesting finding is the huge drop in tablet use from 2014 to 2015, possibly caused by bigger screens on today's mobile phones.

The ratio between desktop and mobile in terms of value-per-visitor was 3.24 in 2015, compared to 4.72 in 2014. The desktop user experience was still vastly better than the mobile user experience in 2015, but the difference was not as stupendously huge as in 2014.

At the end of 2015, the data shows that if a website can make a dollar from a mobile visitor, then that site will on average make $3.24 from a desktop visitor. If these sites were able to improve their mobile user experience to the extent that mobile equaled desktop, then their mobile revenue would more than triple. It's not a trivial amount of money these sites currently leave on the table, considering that mobile now accounts for 45% of their traffic.

(There's also a small potential for gain from improving the tablet user experience, and we do have UX guidelines for tablet design. However, since the gap between tablet and desktop is small and since tablet traffic furthermore is relatively small, there's no doubt where the investment should be focused for smaller sites that can't do everything: start by redesigning your mobile user experience to follow more of the guidelines for good mobile design.)

2014: Mobile UX Was Terrible

Traffic data collected by IBM shows the following allocation of traffic and sales  to online shopping sites during Christmas 2014:

• Mobile (smartphones): 40.6% of traffic and 15.9% of sales = ratio of sales-to-visits of 0.39
• Tablets: 15.9% of traffic and 18.4% of sales = ratio of sales-to-visits of 1.16
• Desktop: 42.9% of traffic and 65.2% of sales = ratio of sales-to-visits of 1.52

Comparing these numbers shows that ecommerce sites have 288% higher conversion rates on desktop than on smartphones. (Desktop has 31% higher conversion rate than tablets, which is stupendously better than the smartphones’ UX performance and confirms our tablet user research finding that most desktop sites work reasonably well on tablets.)

Exacerbating the problem, average sales were $107.72 to desktop users compared with only $88.70 to mobile users. Thus, dollars made per visit were 372% higher for desktop compared with mobile (i.e., sales per desktop user were $4.72 for every dollar in sales per mobile user.) This, of course, is one reason advertising rates are much higher for desktop search than for mobile search: if you don't track your own analytics data in sufficient detail to tell otherwise, a good default strategy is to make mobile bids 20% of the desktop bids for each keyword. (A bid adjustment of -80% for mobile clicks.) However, it's obviously better to track your own analytics, because some of your keywords may perform better — or worse — per mobile click than the average. Queries relating to finding store locations may be more valuable on mobile than on desktop if you run an omnichannel operation.

I was going to say that desktop sites sell almost 4 times as much as mobile sites, but more likely the reason for these terrible mobile conversion rates is that many of these ecommerce companies don’t even have a special mobile design but try to show the same design to users regardless of platform. Scaling a user interface across different platforms doesn't just involve stretching or shrinking the same info to fit different screen sizes as with a naïve implementation of responsive design (RWD): much more is required to optimize a design for different screen sizes and input mechanisms.

Other reasons for the poor showing of m-commerce in 2014 include:

• Some sites don’t even follow the guidelines for ecommerce usability, and the resulting impoverished user experience hurts more on a small mobile screen than on a big desktop computer. For every one of the 884 ecommerce design guidelines a site violates, its sales to desktop users will certainly be somewhat reduced, but its mobile sales will drop much more.
• Other sites may follow general ecommerce guidelines but ignore the additional guidelines for mobile usability. Thus, their desktop sites are OK but their mobile sites are poor.

Whatever the reason, this fresh data should serve as a wakeup call for ecommerce sites to get cracking on mobile usability, because many customers obviously want to shop on mobile devices.

Non–ecommerce sites shouldn’t get too smug: likely they have at least one of same 3 problems: no separate mobile design, bad general usability, or lacking in compliance with mobile usability guidelines. These problems may be less easily quantified for other genres of websites, but if you try to measure your conversion rates for each platform you'll get your own wakeup call.

While “t-commerce” (tablet-based ecommerce) performed much better than m-commerce in the new  data, ecommerce sites still have the potential for a 31% lift in tablet sales if they would bother following the special guidelines for tablet design.