Why Mobile ROI Is So Hard
Those of us who are in the digital advertising industry are well aware of the gap between mobile usage and mobile ad spend. This gap is most clearly illustrated by the chart included in Mary Meeker’s “KPCB Internet Trends 2013” presentation from earlier this year.
As the chart illustrates, although US consumers spend 12% of their total media consumption time on mobile devices, mobile advertising only accounts for 3% of the total ad spend. (BTW, the other take-away from this chart is that, at some point, print advertising will be toast! It’s just a matter of time.)
Why does this gap exist? One of the primary reasons for the lag between mobile usage and ad spend comes down to one word: ROI. If advertisers were able to understand the value of advertising on mobile devices, they would spend more on mobile ads.
So what’s preventing advertisers from understanding the ROI of mobile advertising? Two things: (1) consumer shopping behavior on smartphones, and (2) fragmentation of consumer Internet usage. The first item affects the digital conversion rates that advertisers see from mobile device usage; the second item affects the ability to measure conversions from mobile devices.
Consumer shopping behavior on smartphones
There have been multiple recent studies that have reported the growth of mobile commerce on smartphones. For example, in August 2013 Comscore reported that in the first half of 2013, nearly 10% of digital commerce spend is happening on mobile devices. Year-over-year growth for m-commerce is 24% over one year ago.
While on one hand, we should celebrate the growth of mobile commerce, it’s also useful to look at that growth in context.Although consumers are spending 32% of their digital time on mobile, only 10% of digital commerce is occurring there. Consumers don’t spend a proportional amount of time buying on smartphones, relative to how much time they spend on mobile overall.
There are two reasons why consumer shopping behavior on smartphones results in fewer commerce transactions: (1) form factor and web connectivity; (2) usage of smartphones “on-the-go” for offline shopping.
Smartphone form factor
Smartphones are fundamentally different devices from desktop PCs. The form factor is different, the connectivity to the web is different.
A study by Webcredible in Aug 2012 reported that:
Device security, small screens and the inconvenience of entering card details are still among of the main barriers holding back m-commerce
These findings were corroborated by research from Google’s “Our Mobile Planet” 2013 study, which found that the top 5 barriers to mobile commerce are the following:
- Cannot trust credit card security on mobile device (40%)
- Screen size is too small (40%)
- Cannot see detailed product/service information (27%)
- Hard to type (25%)
- Hard to compare prices and options (22%)
As a result of these barriers to commerce, although traffic to commerce websites from smartphones is growing, conversion rates remain low.
As the chart above illustrates, mobile devices now contribute about 21% of the traffic to commerce websites, with smartphones bringing in about 10.4%.
However, the barriers to mobile purchase noted above have kept conversion rates from smartphone to be less than 33% of the rates from desktop and tablets.
The industry may make some improvements to mobile commerce over time to improve the shopping experience on mobile—larger smartphone screen sizes, mobile-optimized websites, easier (and more trusted) mobile payments, etc. However, in the short term, we will probably need to assume that these barriers to shopping will continue to exist on smartphone for the foreseeable future.
Usage of smartphones “on-the-go” for offline shopping
In addition to the smartphone form factor, another reason why smartphones results in fewer commerce transactions is the heavy consumer usage of smartphones “on-the-go.” Smartphones are always with you, wherever you are (including when you are out and about). They are often used for short bursts of activity for immediate needs, rather than for lengthy browsing sessions. As a result, smartphones are often used for local search and in-store/offline shopping. So the “share” of digital conversions from smartphones is lower, whereas the share of offline conversions is much higher, than traditional desktop PC.
Google and MARC research published a study in April 2013 that analyzed mobile shopper behavior. They found that 79% of smartphone owners are also smartphone shoppers (defined as using a smartphone to assist with shopping at least once a month).
In addition, the study analyzed how shoppers use their smartphones for pre-shopping activities. Given the fact that smartphones are always with you, they are more often used for offline pre-shopping behavior than a desktop PC. In the chart below, the Google/MARC study reports that the Top 2 (and 4 of the Top 8) ways that shoppers use their smartphone for pre-shopping apply exclusively to offline shopping.
A different Google study published in March 2013 analyzed the subsequent actions of mobile shopping searchers. The study found that 28% of mobile searches resulted in a conversion (defined by store visit, call, or purchase). However, many of these conversions were conducted in non-digital ways (offline, in-person).
Mobile searches resulted in purchases 17% of the time, store visits 17% of the time, and calls to the business 7% of the time. (The conversion rates add up to greater than 28% when added because some searches resulted in multiple conversions). If one considers that store visits and calls to the business do not result in any traditional digital activity, then only purchases from mobile searches are trackable as a traditional digital conversion.
However, the Google study reported that over half (52%) of the conversions from mobile searches were done in person—again, in a way that did not result in a trackable digital conversion
What all of this data tells us is that consumers are primarily using smartphones for offline (non-digital) shopping activities. As a result, relative to PCs, the share of trackable digital conversions from smartphones is lower and the share of un-trackable offline conversions is higher—resulting in lower conversion rates seen from smartphones overall.
Fragmentation of consumer Internet usage
So far, our discussion has been focused on smartphone user shopping behavior, and how the smartphone phone factor and on-the-go usage have resulted in the lower digital conversion rates that advertisers see from smartphones. Let’s now analyze why it is difficult to measure the smaller number of digital conversions that are occurring on smartphones.
Think back to seven years ago, to the times before smartphones and tablets. The primary mode of consumer Internet usage was a browser on the desktop PC. In most cases, the consumer completed their entire path to purchase on this browser on their desktop. They saw an ad, clicked on the ad, visited a website, and completed the purchase—all on the same browser on the same PC.
With the rise of smartphones, we live in a much more fragmented world now. I’m not talking about the fragmentation of OS/device. According to Comscore’s most recent August 2013 numbers, in the US, iOS has roughly 40% smartphone market share, Android has 52% share, with the rest being split by Blackberry, Windows and Symbian. Even among Android OEMs, Samsung has enormous share in the US at almost 24% share of the total smartphone market.
What I’m referring to is the fragmentation of consumer Internet usage. There are two types of fragmentation happening today: (1) on-device fragmentation, and (2) cross-device fragmentation.
On-device fragmentation
In the pre-mobile world, the user saw an ad on the desktop PC browser, clicked, and converted.
In the mobile world, the user may see an ad within a mobile app, but when they click the ad they navigate to a mobile website. Users are much more likely to see ads within a mobile app, since according to an April 2013 study by mobile analytics firm Flurry, consumers spend about 80% of their Internet time within mobile apps and only about 20% on the mobile web. One interesting observation from the Flurry study relates to how browsers can just be considered apps, that how some apps—like Facebook—have become browsers.
Studying the chart shows that apps (and Facebook) are commanding a meaningful amount of consumers’ time. All mobile browsers combined, which we now consider apps, control 20% of consumers’ time. Gaming apps remain the largest category of all apps with 32% of time spent. Facebook is second with 18%, and Safari is 3rd with 12% Worth noting is that a lot of people are consuming web content from inside the Facebook app. For example, when a Facebook user clicks on a friend’s link or article, that content is shown inside its web view without launching a native web browser (e.g., Safari, Android or Chrome), which keeps the user in the app. So if we return to the chart and consider the proportion of Facebook app usage that is within their web view (aka browser), then we can assert that Facebook has become the most adopted browser in terms of consumer time spent.
Let’s examine the case where a user sees a mobile ad from within an app. When the user clicks the ad from within the mobile app, they are taken to another context (mobile website, most likely within a webview), where they complete the transaction. During the transition from mobile app to mobile web, the user’s identity is lost. There is no way to share context between mobile app and mobile web, since these two applications are basically sandboxed from each other.
Conversion tracking on the desktop web relied on the user having a persistent browser cookie between the time that they clicked an ad, and the time that they converted on an advertiser’s website. However, on mobile, this persistent identity does not exist. The user’s identity within the mobile app remains there, and when he or she clicks to the mobile website, there is no way to connect the identity.
Until the industry finds a good solution for sharing user context across mobile apps (or across mobile app and web), the challenge of on-device fragmentation will continue to make ROI measurement difficult.
Cross-device fragmentation
In addition to the on-device fragmentation, there is also the more challenging issue of cross-device fragmentation. A user may see an ad within a mobile app on their smartphone, but then completes the purchase process on a different device altogether (most likely PC, though sometimes tablet).
In the March 2013 Google study from above, the research found that 31% of the digital purchases from mobile search were conducted on PC or multiple platforms. This is directionally consistent with another Google study from May 2012, which reported that:
Research on smartphones also leads to purchases across channels. 37% end up purchasing online and 32% prefer to purchase in-store after having conducted research on a product or service on their smartphones.
So even if an advertiser is able to measure conversions that happen on the same mobile device as the user fluidly transitions from mobile app to mobile web, what happens to the roughly one-third of users who research on their smartphone and then purchase on their desktop PC?
This is another problem that the industry faces — how can we accurately and anonymously share user context across devices for measurement purposes? There are a number of startup companies that are tackling the problem using algorithms and triangulation techniques, with various levels of accuracy. Larger consumer Internet brands that require users to login to access their services may also have an ability to connect user identities across devices, albeit with limited coverage. Similar to on-device fragmentation, until the industry solves the cross-device fragmentation problem, ROI measurement for mobile advertising will be difficult.
Many in the digital advertising industry are aware that there is a gap between consumer mobile usage and mobile advertising spend. In this post, I have analyzed the top-level reason for the gap—namely, lack of perceived ROI from mobile ads.
Consumer shopping behavior on smartphones—due to the smartphone form factor, and usage of smartphones on the go—limit the conversion rates that advertisers see from mobile ads. Consumers face a number of barriers to purchasing on smartphones, and they tend to use smartphones for offline shopping much more than they do on desktop. Therefore, the share of digital conversions relative to desktop is much smaller on smartphone, while the share of offline conversions is much larger. All of this leads to lower observed conversion rates from smartphones.
The fragmentation of consumer Internet usage is also making the measurement of those digital conversions that are occurring from smartphones difficult. There are two types of fragmentation—on-device and cross-device—that makes it challenging to connect a user’s identity from the point that they saw an ad, to the point that they converted. On-device fragmentation is due to usage of browsers as apps, and apps as browsers, with no app taking more than 20% share of the consumer’s time. User identity/context is not shared between apps and browsers, so there is no way to track conversions between the two. Cross-device fragmentation is due to users being exposed to ads on smartphone, but then completing the purchase event on desktop (or tablet).
Although there are a number of challenges, there is reason to be optimistic. Smartphone usage continues to grow, as does traffic to retailer websites from mobile devices. Mobile commerce also continues to grow, experiencing year-over-year growth of 24%. And a number of players—startups as well as established consumer Internet brands—are racing to solve the problems outlined in this post. As a result of these and other phenomena, eMarketer predicts that mobile advertising will grow from $8.5 billion in 2013 to $32 billion in 2017.
As with many other types of new media (including the desktop web), there was a lag between the growth of usage and the growth of ad spend. However, until the challenges around consumer smartphone shopping behavior and the fragmentation of consumer Internet usage are solved, the ROI will not be understood, and therefore the true potential of mobile advertising will not be achieved.
