Google Allo

Google Allo failed because too few people used it, by Google's own account. It launched in 2016 as a phone-only app that worked on one device at a time, in a market where WhatsApp already had a billion users, and Google shut it down in March 2019.

What it was

Google Allo was a free messaging app for Android and iOS that Google announced at its I/O conference on May 18, 2016, as "a smart messaging app".

You signed up with your phone number, not your Google account. The pitch was machine learning. Smart Reply would learn over time to suggest replies "in your style," Google said. Allo was also "the first place users can interact with Assistant," The Guardian reported at launch.

Google's announcement offered end-to-end encryption in an optional Incognito mode.

The Verge called Allo a fresh start for Google's new communications division. "It's really liberating to start from scratch sometimes," Erik Kay, Google's director of engineering for communications products, told the site.

What happened

Allo launched on September 21, 2016, The Guardian reported. Google's Pixel announcement on October 4 said the new phone would come with Allo preinstalled.

It reached 5 million downloads in five days, then sat below 10 million for two months, Android Police reported when it crossed that mark on December 13, 2016. By January 31, 2017, it had dropped out of the Play Store's top 500 free apps, Android Police found.

In February 2017, Nick Fox, Google's vice president of communications, tweeted a screenshot of a desktop version, The Verge reported. Allo for web started rolling out in August 2017, at first on Chrome only, paired with an Android phone through a QR code, 9to5Google reported.

In April 2018, Anil Sabharwal, the new head of Google's communications group, paused investment in Allo and moved its team to Android Messages, The Verge reported.

On December 5, 2018, Google said Allo would keep working "through March 2019" and that users could export their chats until then. Its farewell page says Google said goodbye to Allo on March 12, 2019.

Why it failed

Google said Allo didn't get the traction it hoped for and that Messages had the momentum. Reporting at the time added the privacy defaults, the one-phone design and WhatsApp.

What Google said

Sabharwal told The Verge in April 2018 that "the product as a whole has not achieved the level of traction" Google had hoped for. Google had wanted a product that "hundreds of millions of people" would get excited about and use, he said, and "we're not feeling like we're on that trajectory."

Android Messages, which he put at 100 million monthly active users, looked to him like "a product that has tremendous momentum."

Matt Klainer, Google's vice president of consumer communications products, gave the reason for the shutdown in December: "Given Messages' continued momentum, we've decided to stop supporting Allo to focus on Messages." Google had already moved Smart Reply, GIFs and desktop support into Messages, his post said.

Neither Sabharwal's April comments nor the December post named the encryption default, the stored chats or the one-phone design as causes.

What reporting showed

After the I/O preview, Edward Snowden tweeted that turning off end-to-end encryption by default was "dangerous, and makes it unsafe," ZDNet reported. Allo's AI "won't work if it can't scan a user's messages," TechCrunch explained.

At the preview, Kay had told The Verge that Google stored message data only "transiently". On launch day, The Verge reported that Allo would instead "store all non-incognito messages by default", a change Google said would improve Smart Reply.

Ron Amadeo wrote in Ars Technica's review that Allo "works and receives messages on a single device at a time" and had no web or desktop version. In February 2017, The Verge's Dieter Bohn wrote that the missing desktop client "seriously hurt Allo's chances among early adopters".

At the May 2016 preview, Bohn called Google's plan for getting users "a little, well, unformed." The people you want to talk to are your friends, he wrote, and "right now, they're using something else." WhatsApp had announced one billion users that February.

In April 2018, Bohn reported that fewer than 50 million Android users had installed Allo, against over a billion installs each for WhatsApp and Facebook Messenger. He wrote that Allo "had no such strategy for acquiring users" as those apps had, and that it "likely suffered from Google messaging app fatigue."

What later analysis found

In a 2021 history of Google's messaging apps for Ars Technica, Amadeo called Allo "Google's dead-on-arrival WhatsApp clone."

He argued that a minimum viable product can work against little competition, but not as a copy of an established competitor. "There's no reason to use a half-baked WhatsApp clone when regular WhatsApp exists," he wrote. Allo also never did the fast iteration an MVP is supposed to buy, in his view. The desktop fix came a year after launch, when "Allo was already dead."

Amadeo also wrote that the phone-number sign-up suited WhatsApp, while people trying a new Google product "usually already have a Google account."

The stage it reached

Allo died in the introduction stage, which the product life cycle guide describes as the launch period, "when awareness is low and sales are small."

It was free, so downloads are the nearest thing it had to sales. They came in a launch rush and then stalled, by Android Police's count.

The introduction post says a product manager's job at this stage is "to measure activation and early retention rather than sign-ups." Every public Allo figure I found tracks downloads, installs or chart rank, not active users.

It never reached growth, so I count it as a failure, not a decline. Google's attempt to take on Facebook, Google+, never left the introduction stage either.

The lesson

I think Allo shows what happens when a team decides distribution can wait.

Before launch, Kay told The Verge that "if you don't have a great product that users love and are willing to recommend to their friends, then there's no sense in worrying about distribution." Amit Fulay, a group product manager on the team, put it as "just nail the product."

That order works when one person gets value alone. A messaging app is worth little until the people you talk to are on it, and I'd argue Allo's one-phone design made each invitation harder to accept.

If your product only pays off once someone's friends join, plan how the second user arrives before launch. Then judge the first months by how many conversations keep going, not by downloads.