Rather than waiting for users to come to Google, Google went to them, with the Social Search Bot - helping you search Google, right there on Twitter.

Say you’re busy tweeting away, and you started craving a burger. Just tweet @Google with a burger emoji, and we’d send you a Google search result URL showing great burgers near you, as well as a fun little GIF.

We did this through a little known feature of Google: searching “[blank] nearby” told Google that you wanted results based on your GPS location, and showed you the optimal results, wherever you were.

In total we indexed over 200 emojis, from burgers to barbers to banks.

We built this so that anyone, anywhere, could use the same link, and when they retweeted their result, their followers across the world could see it, click it, and find local results near them too.

On December 7, 2016 we launched the bot with a single tweet, telling followers of the @google account that “We Speak Emoji. Let’s talk”. With the launch tweet we embedded a short how-to video explaining how to Tweet a request at the bot.

We timed the first 24 hours to include a lifted embargo on PR, and took advantage of the family of owned Google Twitter accounts (e.g. @GoogleMaps, @GooglePlay) to help drive the behavior by tweeting relevant emojis at us (? and ???, respectively).

We also encouraged users to break standard Twitter conventions and tweet "?? @google" vs the standard reply format of "@google ??", ensuring that all of their followers saw the request, ultimately driving more interest across Twitter.

In one week we drove 35,095 1:1 interactions, leading to 269,187 Google searches. This multiplicative growth was due to 1 in 6 users retweeting their replies, earning 13.5m social impressions, and each person we responded to drove 7.6 additional engagements via their own social graph.

Press coverage was incredible, achieving 21 million media impressions, with global coverage outside the tech press bubble.

25% of users interacted with the bot more than once, and we were overjoyed to find 1 in 8 users who clicked a link made a similar search on Google within the next 7 days, proving we were able to influence meaningful behaviour change in our audience.

It went so well that we took the data showing the most popular requests and built them directly right back into search - try searching “?? near me” from your mobile device to see it in action.