Google's crusade to make mobile web apps less, well, horrible
No chance of making people on the web less horrible, though
Chrome Dev Summit At SFJazz – the first freestanding jazz arts center in the US – on Thursday in San Francisco, Google riffed on the virtues of Progressive Web Apps as a vehicle for efficient and engaging content delivery.
"Our mission is to move the web platform forward," said Darin Fisher, VP of Chrome engineering, at the Chrome Developers Summit.
It would be another hour or so before a video acknowledged what most developers have known for years, that the web platform buckled under the weight of premature enthusiasm and ever bloating ad pages: "Let's face it, until now the alternative to native apps hasn't been great."
Back in 2009, Google declared that the web had won, optimism as shoddy as recent election projections. Not long after that, it became clear that native apps offered a better experience on mobile devices than web apps.
Facebook touched on the issue in a 2012 post that describes the company's rationale for shifting its development focus from web apps to native apps. It wasn't quite the repudiation of web technology that many took it to be – Facebook said it would continue to use HTML to allow it to update its app without requiring a repeat download – but it affirmed that web technology needs further attention.
To frame the issue in numbers, a recent study from Google's DoubleClick advertising arm found that the average load time for mobile sites over 3G is 19 seconds. It also found that 53 per cent of visits to mobile sites are abandoned after 3 seconds. Moving the web platform forward mostly means making mobile web apps lightning fast.
There have been efforts to address page load performance, in the form of Google Accelerated Mobile Pages (AMP) and Facebook Instant Articles. Where AMP and Instant Articles focus on getting content to users faster, Progressive Web Apps (PWAs) have more to do with engagement, through parity with native app expectations.
PWAs were proposed last year by designer Frances Berriman and Russell as an attempt to help web technology compete with native code. PWAs would be responsive, network-independent, and discoverable as applications in search. They would observe app-oriented navigation and interaction conventions. And they would be always up-to-date, safe, installable, and linkable.
PWAs are about behavior rather than APIs. Google has a checklist that describes the way PWAs should work. Making a compliant PWA may involve technologies like service workers, which cache content so the app isn't network-dependent. But developers could equally well write a custom caching routine, were they so inclined. Similarly, developers might choose to use the Credential Management API introduced at Google I/O earlier this year, because it makes authentication easier. But other approaches work too.
"Progressive web apps are not just about any one technology or any set of technologies," said Fisher. "It's more about a new way of thinking about building web experiences."
It appears to be a good time to be having such thoughts. Native mobile apps have lost their luster, apart from those backed by the marketing muscle of companies at the intersection of social and mobile communication, like Facebook. As some tech pundits like to say: "The app boom is over."
Perhaps more meaningful than the overabundance of apps among smartphone users in the US and Europe is the situation in emerging markets. Elsewhere in the world, where network conditions are often constrained – 60 per cent of mobile traffic travels over 2G connections, according to Fisher – and downloading hefty native apps isn't practical, companies like Flipkart have been early adopters of emerging web technology because there isn't a better alternative.
Thao Tran, head of global product partnerships for Chrome and the web platform, cited the case of Housing.com, an Indian real estate website. The cost for the company to acquire a user for its Android app is about $3.75, Tran said. The cost to acquire a user through via the mobile web is about $0.07.
Tran also recounted how Lyft recently rolled out a web app and ended up receiving five times more ride requests than expected. She said Lyft's goal is to have feature parity between its native and web apps. In markets where downloading native apps weighing in at 17MB and 75MB respectively might prove prohibitive, a web app of less than 1MB could entice new customers.
The appeal of the web is its vast reach, and much the web is seen through Chrome. Fisher said there are now over two billion active Chrome browsers across mobile and desktop devices, reach that rivals Facebook, albeit without the social data.
Even so, Google Chrome should appeal to marketers because, as Google product manager Zach Koch explained, people trust Chrome by letting it remember personal information like credit card numbers and home addresses.
Easy access to personal data contributes to the way investors value companies. Apple observers, for example, sometimes cite the number of iTunes accounts as a financial bellwether for the company, based on the notion that spending comes easily when payment information is already on file. Google is in a similar situation with Android. And if it can turn its mobile browser into a shortcut to spending, so much the better for its advertising business.
Toward that end, Koch discussed PaymentRequest, a recently introduced payments API for the web. Introduced in July, PaymentRequest is a standards-based API to make it easier to pay for things from the browser. It's not, Koch stressed, a new payment method or an attempt by Google to become payment processor.
"Our goal is to help users pay the way they want to pay, and to do it quickly or efficiently," said Koch.
In some circles, that counts as moving the web forward. ®