Reaching altitude by being mobile firstPosted on by David Low
At the recent conference, I presented on the topic ‘Reaching Altitude’, about Skyscanner’s journey from being a mostly desktop business to being ‘mobile first’ and ultimately starting to embed our services in other products.
Essentially though, the talk was about making good flexible preparations for the future ; in the modern digital age, change is sometimes hours around the corner. Businesses simply can’t afford to be rebuilding and chasing all the time.
Skyscanner and Mobile First
A while ago now, we asserted ourselves as being a product led, mobile first business. That’s not a strategy in itself, but some basic assumptions and accelerators that underpin day-to-day and longer term development.
From making our internal networks available as simulated 2G and 3G experiences, to ensuring all staff have access to a wide array of devices, we try very hard to make sure everyone is up to date.
It’s important to understand what we mean by ‘mobile first’ though. For us it’s partly a behaviour shift to think differently about our users and how they want to use digital products like Skyscanner.
But it’s also about architecture, engineering and product development — and how to prepare yourself for a new world every few months.
In 2014 alone, Skyscanner saw a 77% growth in usage of our products on mobile devices — not a particularly new trend, but one we spotted early enough to act upon — and one that isn’t going away.
But in many of the markets we operate, mobile-first behaviour is almost the default. Indeed, some markets are already bordering on mobile-only, with consumer desktop usage and fixed-line connections almost unheard of in some emerging markets.
In Korea for example, 74% of users prefer to use the Internet on a mobile device. They may use it on other devices but the preference is clearly there.
Overlay on that, a high percentage of app usage and even then, a high percentage of ‘platform’ products like Kakao Talk — where you’re almost talking about using apps within apps — desktop is struggling to get a look-in.
Indian e-retailer Myntra dumped its desktop site last May in favour of being ‘mobile only’. Most observers expected a drop in revenue as a result, given there was still a chunk of desktop spending going on — but in fact, early indications are of a rise.
The mobile product had gone through various enhancements to make it more mobile friendly, tempting those users into longer sessions and more conversions to buy — and ultimately this enhanced experience tempted desktop users to migrate and stay.
And yet in that same market, a large proportion of new internet users are on 2G or similar connections, to the extent that Internet giants such as Facebook and Google are adjusting their products to suit. Facebook Lite, a thinned-down Android app, and Google Web Light, a web optimiser, play directly into this expanding market. Without these products the giants would not reach the sort of scale they’re used to elsewhere.
What is Mobile?
Having looked at all those things, what gave us most to ponder in going ‘mobile first’, was trying to define what we meant by it. What is mobile?
Take three scenarios — imagine three people on a train. The first is sitting looking at a phone in their hand — almost the dictionary definition of ‘mobile’, and the obvious example.
Next to them is someone using a laptop. Clearly they’re not on a ‘mobile’ but they will be suffering from limited, poor bandwidth, tunnels and other blackspots. Their experience will be as impaired as anyone else.
Next to them is someone without any obvious devices. But who are we to say they don’t have a wearable on their wrist, a phone in their pocket buzzing away, or that they’re waiting on a notification of some kind? ‘Mobile doesn’t just constitute active use.
Even if we did narrow it down to phone, what do we mean by that? Is it someone simply browsing mobile web? Or using an app? Are they deep inside a messaging app like WeChat — something almost a default behaviour in the Far East?
As application developers or content publishers, should we even care?
In truth the answer to ‘what is mobile’ is: ‘all of the above’. But in most of the examples above, we’re not talking about devices or platforms or applications, we’re talking about context. And that theory around contexts has been defined as the ‘mobile mind shift’.
In a very good book which dives into this topic, The Mobile Mind Shift, Ted Schadler, Josh Bernoff and Julie Ask start to describe (in example-led detail), what that mind shift is all about and how to address it. They define it here:
“The mobile mind shift is ‘the expectation that I can get what I want in my immediate context in my moment of need’.”
What they’re trying to explain is the ‘mobile moment’ — that point in time where provision and need just happen to meet, by matching together some good data points and contexts rather than pot luck.
To use an example closer to home — Skyscanner Google Now cards provide flight price changes to users, when Google believes they’re most likely to be actioned.
Google’s knowledge of the user gives us deep context we can’t know, the benefit to us is learning about those conversion times, how valid an alert is at a given point in time, and what types of messaging work best for each user.
Thinking about the end points that suit those different ‘mobile moments’ is critical.
Degrade to Desktop
Companies that simply take their web experience and replicate it on phone, even if they use accepted techniques like responsive design, aren’t quite getting that principle of context. Making a big thing smaller doesn’t cut it any more.
Adaptive development — only showing blocks of content relevant to the device or context — goes a long way towards fixing that, but as we’ll come to later on, it doesn’t answer some of the micro-contexts and non-visual ways people are starting to get used to. Nor does it necessarily answer being an ‘app within an app’ such as Kakao or other Asian experiences depend on.
If you think of a modern smartphone and the mixtures of touch, geography, movement — it’s both the hardest one to get right, and the most rewarding at the same time. From there, you effectively drop capabilities as you head backwards desktop sized experiences — and the visual design isn’t really that important if the user doesn’t get what they wanted along the way.
It’s almost progressive enhancement in reverse, and something we’ve been calling ‘degrade to desktop’.
And it’s not an easy thing to get right , but by thinking about small blocks relevant to particular contexts , and only building and serving those blocks at the right time, you give yourself an efficient way to meet these emerging challenges.
A principle I call ‘leaning on the platform’ means you’re only building the bits you actually need for your own purpose and context , built on a solid underlying platform and data structure. There aren’t that many new problems to solve so why solve them again?
It’s a fact that being featured in an app store is mostly dependent on this methodology — it’s both efficient as a development and quality-assurance process, but also showcases the power of the core platform, which is crucial exposure for handset / OS manufacturers.
Growing big by thinking small
From a flights perspective, maybe someone doesn’t want a flight price, maybe they just want to know ‘who flies’ to somewhere. Or whether they can fly somewhere at all.
These are small queries that can be delivered by microservices to fragmented clients. Ideal building blocks for modern low-friction products — voice, messaging, wearables or IoT. And all delivered from a platform that still serves legacy and existing needs without the need for monolithic systems.
Having got this far, you should be able to take your concept of a mobile moment, and deliver it almost anywhere — it’s now about the nitty gritty.
The good people at Forrester Research, describe how to enable all this through a ‘four-tier engagement platform’, which puts it very well:
If you think about it that way, previous responses to mobile such as m.dot sites, responsive and adaptive, are lucky if they deliver one or two of those tiers.
You need to work your way back from the mobile moment and that fragment of expectation, all the way down your application structure — to the point you can deliver that moment’s worth of information quickly and efficiently.
By starting small, you’ll find you can scale to the billions of devices and endpoints you can address.
A world of possibilities
This attitude towards development also opens up those other platforms mentioned earlier — voice, wearables and IoT — and also embedding yourself in the wider Internet.
Opening up these systems, platforms and fragments for people to adopt into other products and services, allows you to take those good foundations and scale way beyond what you might achieve on your own.
At Skyscanner we’ve adapted our platform from things as simple as presenting price changes within Google Now to the more complicated integrating of our flights API into Amazon’s Alexa. We’ve created an Apple watch app that directs you back to your hotel in a strange place, and, at the bigger end of the scale, created full blown travel search for MSN users.
Lots of variations, big and small , all based on the same underlying platform, and a variation on that four-tier architecture. Hopefully there will be a lot more to come from us with those foundations in place, and through Skyscanner for Business we’re using those foundations to do some exciting things. Watch this space.
That, amongst many other things, is part of how we plan to ‘reach altitude’ at Skyscanner.