“Holy shit, this thing actually works”
These were the words Dennis Crowley, founder of Foursquare, used to describe a moment in 2013 when he walked into Little Cupcakes Bakeshop to find a tip from their new geofenced notifications.
Earlier today Dennis went back to 2009, retelling some of the backstory of Foursquare and their new app, ending with a tease for the upcoming Swarm for Apple Watch:
Oh hey! Honored to see our first Apple Watch app show up on Product Hunt. So three things real quick… ☺
#1. This is exciting to us, not because you can tap tap tap and scroll thru a little UX on your wrist to find a great place for lunch. This is exciting to us because smart, contextual-notifications that you can read at a glance can be pretty incredible. And when they’re about the place where you are or another great place a hundred yards away, they can make you feel like you have superpowers.
#2. We’re been working towards this “when you walk around, your phone should buzz you and tell you about the awesome things you don’t yet know about” vision of the future for a long time. Honestly, we were thinking about this stuff with Dodgeball @ Google, and were motivated to build Foursquare because we wanted this to exist so badly. So, want a brief history of how we got from 2009 -> 2015?
2009. Check out this screenshot from Foursquare 1.0 — http://cl.ly/image/2z1c2F2p2u2n. From the very beginning, you’d check-in and “tips” for places nearby would pop up with the intent being “we should all have a piece of software in our pockets that teaches you about all the awesome things nearby”. But back then, the only way to trigger these “awesome things nearby” messages was to press that check-in button over and over and over again. A bunch of you probably remember seeing these “pop-up tips” in Foursquare (and we still have them in Swarm). Our goal was always to make these messages pop-up without the user ever having to *do* anything…. make them pop-up by just walking around. Turns out this is really hard.
2011. The iPhone 4S comes out w/ iOS5 (first version of iOS to offer passive geofencing in the background). It was a more limited than we wanted/expected — you could only have a couple dozen geofences set at once and meanwhile we had 100,000 things we wanted to teach you about NYC (not to mention every other city in the world). This is where we started to think “if we really want this to work, we’re going to have to build this engine ourselves, from scratch — geofencing, stop-detection, contextual-awareness”. And that was when we started working on a project called “Pilgrim.” (Google it for more info)
2013. We finally get Pilgrim working the way we want it to. Using a “virgin” Android phone (no friends, no checkins) I walk into the cupcake shop down the street from Foursquare HQ and get a ping telling me to try the “Dreaming Princess” cupcake (which was delicious, btw) — https://instagram.com/p/dmjF7jmvtv/ — which lead to my first “holy shit, this thing actually works” moment. The geofences we were using this time were not circles, but rather abstract patterns created from the GPS/wifi signal shapes that we have derived from our billions of checkins. We also realize that places have “fingerprints” (open/closed hours, popular hours, types of people that go there) and people have “fingerprints” (places you’ve been, categories you’re likely to go to). We discover that it’s the combination of these three things — venue shapes, venue attributes, user patterns — that allows us to have a high confidence about the place you just stopped at and what’s else would be interesting nearby. We launch this slowly, testing to make sure people dig the notifications, that people tap on them, that we’re not sending too many, that people don’t turn notifications off or delete the app because of them. All signs point towards it’s working and people dig it. We launch to millions of users. We roll out a Pebble version end of 2013.
2014: Google/Android Watch comes out and this “buzz your wrist instead of your phone” story starts to feel more real. I get a Moto360 and the first day I wear it, I walk into a restaurant deep in Williamsburg (Brooklyn) and it buzzes me with a suggestion of what cocktail to order and what dish to try — https://instagram.com/p/suROb4mviv/ I have a “holy shit, it works on your wrist… and it actually feels right” moment. We start feeling super confident that we’re sending good quality pings to folks when they stop at interesting places. We gain the ability to detect when people are in new neighborhoods vs familiar neighborhoods, and learn to turn up the awareness when people are visiting new cities. I took a trip to Tokyo in December, was walking thru Harakuju (neighborhood) and Foursquare buzzed my phone telling me about a hipster coffee shop and an arcade 100 yards ahead of me. I had my “holy shit this works everywhere” moment.
2015: Apple makes the watch that people want to wear. We’re feeling pretty good about the Pilgrim engine, the venue shapes, the tip quality our the stop-detection algorithms. We’re now sending tens of millions of pings to people’s phones every month. We start getting really excited about sending tens of millions of pings to people’s wrists in the same way. Is it perfect? No. Do we get it right every time? No. Does it get smarter and more accurate every week? You bet. Will it be smarter/better/faster 3 months from now. Of course. And, more importantly, are we actually building and launching the things that we dreamed up 6+ years ago? Hell yeah.
@dens / co-founder & CEO
ps: Oh, I almost forgot…
#3. Yes, yes, yes… Swarm app coming soon. Prob in about two weeks? We’re working on a bunch of pretty fun changes to Swarm and we wanted to group them all together. Swarm Watch app will be pretty basic to start, but like everything else we work on here, we got big plans ☺
Oh, one more thing, before you start playing around with the Foursquare app, do me a favor and make sure both your iOS Settings & Foursquare Settings are set up properly. And remember, it takes a few days from when you first download the app before you really start getting pings (we take some time to learn about your travel patterns). Thx!
Settings here: http://cl.ly/image/461u2t1r0r2l
“Holy shit, this thing actually works” is an amazing feeling. Congrats on the launch, Dennis and team, and visit Product Hunt for the full discussion.