These days, it’s no longer a question of ‘why’ marketers should be optimizing for mobile.
According to Mary Meeker 2015, an average of 2.8 hours is spent surfing the web on mobile, compared to 2.4 hours on desktops and laptops.
So what marketers need to be thinking about is the ‘how’ of mobile optimization, as WiderFunnel CEO Chris Goward notes that the arrival of the year of mobile seems only true from the consumer side. He observes that a lot of web sites still can’t truly be considered mobile-friendly.
In an episode of Landing Page Optimization, Chris and SiteTuners CEO Tim Ash sort through factors to consider in making mobile work.
1. Device versus Context
On mobile, there is the smaller screen and the different operating system to think about.
Chris stresses, however, that mobile optimization is beyond different device types – it’s really more about context.
When thinking about how your site should be laid out, you have to consider the restrictions mobile has.
- Someone in a mobile context, for instance, has environmental distractions that people in a desktop/ laptop experience don’t have to deal with.
- They also have different needs.
- People in a mobile context are, in many cases, looking for something location-based, so the task might have a higher degree of urgency.
Consequently, because of that urgency, the response channel and the call-to-action (CTA) will be different. Customers may want to call you instead of waiting for an online response, for instance.
So, when optimizing for mobile, one thing you have to pay attention to is page load time, as that’s critical to meeting the user’s urgency.
2. Marketer Bias
For many years, marketers have gotten comfortable with web done on a large screen, but mobile turned everything on its head.
On mobile, the interfaces, user experiences, and the interactions are different, so marketers have to be so much more disciplined about what and how they communicate. To that end, one pitfall to watch out for is the inside-out view.
A typical web developer works at a desk looking at a desktop most of the day without looking at his/her phone, and consideration of how the mobile site actually works is limited to a few minutes of QA.
The problem with this set-up, Chris points out, is that marketers don’t really get the full experience of their mobile site and won’t stumble upon critical insights they might have in their desktop. There’s no “dogfooding” in the same environment where something is expected to be used.
The personal context of actual users is different – they’re out of the office, and they’re not thinking about work. So, marketers need to have the discipline to build in unstructured times.
3. Mobile Urgency
With conversion optimization, the details are critical.
The large concept is one aspect of optimization, and that’s more about the top-end of the funnel.
But for mobile, a lot of the opportunity is in the bottom-end of the funnel. And with transactions, it’s more intense – load speed, tiny interactions, just a word here or there, a few pixels can spell the difference between success and failure.
For more background, read up on the difference among the Persuasional, Information, and Transactional ends of the conversion funnel.
4. Conversion Spectrum
Mobile covers both users in the early stage and the late stage of the conversion funnel. Majority of the time, people get on their mobile device when they’re ready to buy because they have an immediate need. However, people also get on their mobile device when they have an idea or thought they want to research on.
Chris notes that this is a challenge, as marketers have to provide that context in a very restricted environment. At the same time, it’s an opportunity, especially for marketers with high traffic, to test dramatically different interfaces to figure out what mobile interfaces should really look like.
5. Mobile Testing Tools
The good news is that tools now enable optimization with ease. Almost all major testing tools have mobile capabilities, and many have mobile app capabilities.
Chris observes that mobile app testing used to be traditionally restricted and very difficult to do without pushing new versions of the app. Now, through software development kits (SDKs) or javascript, it’s possible to test within apps without pushing new versions.
Optimizely, for instance, allows mobile app testing. Also, with Artisan, a mobile app tool, you can just use Google tag manager to insert some script and get tests running.
So, now, it’s possible to test even with just a little technical know-how, which opens up the possibilities for businesses that rely on apps to generate revenue.
Tim agrees that this is critical because marketers used to be tied to software release cycles, staging servers, and everything else. Now, at least the non-functional, the visual presentation parts of the experience can be handled and tested by marketers independently of the relatively slow pace of software rollouts.
Chris adds that “waiting for consumers to update their apps is like molasses in January.” It’s not going to happen quickly because people don’t like to have their bandwidth used up by updates they deem unimportant. So the current mobile tools allow marketers the freedom of testing anything, anytime, with the rapid iteration that’s really required for continuous optimization.
6. Mobile Segmentation
In the web world, segmenting by traffic sources when you’re running tests is important, as audiences from different sources react differently. On mobile, there are additional constraints like device type and operating system.
Chris observes that on mobile, segmentation is exciting because typically, we know more about the user.
There’s of course the device type, but there’s also the usage. For this, marketers can still look to the age-old direct marketing methods- things like recency, frequency, and monetary value (RFM). You can segment audiences based on the three factors of …
- how recently they purchased
- how frequently they make a purchase
- how large their purchase is
There’s also the notorious gulf between iPhone and Android users. They’re very different users and a variety of different segments, and the kinds of purchases they make tell you who they are.
Chris stresses that there’s not just one way of segmenting. He says that marketers often think of segmentation as simply creating and implementing pre-planned segments.
Sometimes, however, it’s better to run tests that highlight different value propositions or different approaches, and then look in the backend at the different segments to identify where the differences in performance are. This can reveal which segments are actually important to focus on rather than guessing and planning beforehand.
Tim points out that for businesses that are big enough and have a data scientist, they can create predictive models.
They can also collect all the data, then sift through it to find clusters of high value types of visitors, and after the fact, create models of who’s got the most value, and see if it’s possible to identify those people ahead of time. You might find, for instance, that iPhone 6 users that came back within the last 24 hours have 10 times the value to you than an Android user who hasn’t been back in a week.
Chris points out that there are certain great things to be said of predictive models. They have to be, however, based on previous usage data – there has to be to some data to pull on to predict.
What marketers have to avoid is considering something as a good segment based on detailed research and personas which they’re developed, which according to Chris, are merely interesting stories that we create for ourselves as marketers.
Tim agrees that segments should be based on more durable roles and tasks – the visitors relationship to the site and their specific intent. That means user scenarios are better than personas.
Putting It All Together
Marketers today can’t just worry about the offer and the transaction with the user. They need to think about context, and avoid marketer bias. They need to understand the varying tasks, the stage in the buying cycle, and the relative task urgency. They need to segment the data in ways that have not been done before, and test on mobile, something that everyone’s just getting accustomed to.
That’s a lot of new tasks, and since most marketing teams don’t exactly double in size to meet the new demands of the job head on, most online marketers need to diversify in their specialization. That creates new challenges – but it also opens up a ton of opportunities for the ones who get it right.
