What Marketers Should Take Away from Google I/O 2017

Read Time: 5 - 10 min
Google wrapped up their I/O conference last week in Mountain View. The emphasis on the keynote was squarely on AI and AI-powered experiences, with the Assistant taking center stage. They also presented two strong but different visions for VR with Daydream and a totally new take on augmented reality (AR) with Google Lens.

As is common with Google I/O keynotes, these tend to be very aspirational in nature - remember Google Glass and the Nexus Q taking center stage in past years - but they give a good idea of where Google thinks the industry will be heading down the line. Because Google can be exuberant futurists, I’m going to try to break down the conference takeaways in three tiers: Things you probably should start thinking about right now, things you can probably wait and see on, and visionary future ideas that will probably remain niche for the time being.

Things to Think About Now

  • Android O - There's a lot of new things in Android O that you should be thinking about. In particular, for apps doing their own location processing, you’ll need to start looking at compatibility with the new direction they’re taking with battery life optimizations.
  • Notification Channels in Android O - Notification Channels will let your users opt in and out of specific types of push messages, and set separate preferences for the interruptiveness - like sound and vibration - for different types of notifications within your apps. This won’t change the kinds of messages you send, but if your app takes advantage of channels, your users have options for dealing with your marketing notifications other than just disabling all notifications. That’s a big win for users and for marketers alike.

  • Kotlin! - This fledgling programming language is now officially supported by Google for Android development; most Android developers I’ve spoken to are very excited about this. It can coexist nicely with existing Java code, is highly compatible, and adds relatively little to most app’s overhead and app size. The overly simplistic way to think about it is “like Swift for Android” -- but don’t say that out loud to a developer, or they’ll roll their eyes at you. Along with some other exciting new Android architectural tools like Room, this is the most exciting I/O conference in years for Android coders.


Wait and See

  • Google Assistant Actions - If you have a service with a strong use case for low-UI interactivity, consider building a Google Assistant Action. These actions can now run on the Android phone Assistant, not just the Home Speaker. There’s still a discoverability problem for developers targeting these sorts of platforms, though - and an even bigger re-engagement problem, since notification capabilities are very limited.

  • Android Go - The new slimmed down “Android Go” experience will be coming to lower-end devices; if you’re targeting the developing world or the lower end of the market, you should consider building and testing compliant apps so you can be featured on the Android Go Play Store.

  • Android Instant Apps - Google has been promising Instant Apps at I/O for two years now, and they’re finally ready for prime time! They come with some distinct caveats, though: Most importantly, they’ll only work on Android 6.0+ at first - that means two thirds of all Android users won’t be able to take advantage of your Instant App. They’re also tough to discover, rely heavily on Web search and deep linking, and have no re-engagement or push notification capabilities. Developers with a great app experience, highly monetizable one-shot content, and great SEO should be jumping on the bandwagon here; most other developers should simply focus on a good mobile web experience and their primary mobile app experience instead.

  • Android Wear - Google’s smartwatch OS was barely mentioned on stage and only warranted two sessions during the conference. Their flagship watch announcements for Wear 2.0 with partners earlier this year received largely lukewarm reception; barring major new hardware announcements, it seems like Android Wear just isn’t a cornerstone of Google’s consumer strategy this year. On the back of news that companies have been backing off from working on Apple Watch apps as well, this may signal cooling down on smartwatches, at least when it comes to third-party developers.

  • Daydream VR - Google announced new standalone Android-powered VR headsets, but consumer adoption for the brand of low-end headset-based VR Google is pitching seems to be slow. Daydream was center stage at I/O last year, and still got a lot of stage time this year, but the smash-hit breakout success stories just aren’t there yet.


  • Google Lens and Visual Positioning System - Google demoed some great new image recognition capabilities for their own software, and what they’re calling the Visual Positioning Service (VPS) for specific compatible Tango sensor-enabled phones. Whether third-party manufacturers take up the new hardware requirements - so far only one partner, ASUS, has released a compatible phone - is very much unclear, though, so put this squarely in the “slim chance of market maturity in the next 12+ months” bucket. Google’s story here - light augmented reality centered around smartphones rather than headsets like the ill-fated Google Glass - could be the most realistic AR vision we’ve seen, but mainstream adoption and developer tooling may be years away still.

  • TensorFlow - This qualifies as a "someday" item for everyone but experts. While Google is clearly aiming for the head of the pack with the low-level tools and raw silicon for building highly scalable machine learning tools (TensorFlow and Tensor Processing Units), there’s still a lot of hard data science companies need to bring to bear to take advantage of these tools. We’re very pro AI at Localytics (our Predictions product uses machine learning to predict what users will do in the future), but to take advantage of what Google has to offer today you’ll still need a very savvy team of data scientists.

  • Progressive Web Apps - PWAs promise a no-mobile-app-code-needed way to turn web content into a packaged Android app. But they’ve still got a discoverability problem, and the primary problem is that they’re still Android-only for the more advanced installation and notification capabilities. Android Instant Apps fill the same gap as PWAs but with full native experiences and capabilities, and the clear content focus at I/O this year was on Instant Apps instead of PWAs.

Wrapping It Up

At Localytics, we’re excited for the future of mobile computing - and the future of “beyond mobile” experiences on display at I/O this year. I personally always love seeing the keynotes and sessions from Google I/O - from the practical nitty-gritty developer content up to the pie-in-the-sky visionary hardware pre-pre-preproduction announcements. Clearly Android apps are still where the rubber meets the road for the company, but the vision for the Assistant and an AI-powered future full of AR phones has us excited to see what’s next.

You can catch all the talks on Google’s I/O conference website, and let us know in the comments what you’re excited about, skeptical of, or starting to develop for today!