Everybody wants to enjoy their free time. Lots of apps are trying to solve this problem but nobody has quite cracked it. What's right for you might not be right for me, and how am I supposed to know the difference? At one point our founder, Justin Parfitt, was traveling in Italy and saw a one-star review for a beautifully converted medieval mansion. The reviewer warned travelers that the hotel was too old and that the staff "don't speak English good". They recommended the Sheraton near the airport instead. Justin thought, "why am I seeing negative reviews from people I have nothing in common with? Why can't I just see recommendations that are right for me?". So we set out to create a local and travel discovery app that connects people based on their shared interests and the experiences they recommend.
What it does
HeyLets blends the community engagement of social with the powerful utility of personalized discovery. Other apps concentrate on places, with any user content on secondary screens, but HeyLets is more personal and teases out each user's experience at a given location: the wildlife on their favorite urban hiking trail, the mind blowing pistachio gelato at a gelateria they came across in Rome, or maybe that hidden gem that serves killer hortobágyi palacsinta. On HeyLets you see a feed of user generated experiences - unique social stories told with 200 characters and a photo. A powerful machine learning algorithm populates your feed with experiences recommended by people with similar interests and demographic profile, so what you see in your feed is relevant to your age, gender, relationship status, interests, location, time of day and day of week. HeyLets is integrated with Open Table, Uber and Booking.com, so once you find an experience you'd like to check out you can make a booking and get there without leaving the app.
Everyone is always hearing about a new restaurant with a signature dish that sounds incredible, or a karaoke night that's an absolute blast, but more often than not it'll go in one ear and right out the other. HeyLets finally solves this problem. As you browse through your feed you can Wishlist the experiences that you'd like to check out. You can filter by category as well as location, so you can see what art and entertainment is recommended for you in Paris, or what clubbing experiences are trending in New York right now. Build your Wishlist in your hometown and in destinations you'd like to visit someday. You can search, filter and curate your Wishlist so you'll never be at a loss for things to do, wherever you are in the world. What's more, if you're out and about and you're near a location on your Wishlist you'll get a notification that there are some fun things to check out nearby.
What we felt really captured the essence of our mission–positively connecting people around the experiences they love–was closing the loop, or: user A recommends an experience, user B Wishlists that experience because it sounds awesome, then user B actually goes and checks the experience out - eats the burrito, watches the sunset from the park, whatever that might be. That moment, when user B consumes the experience user A recommended is a big point of differentiation for us. We didn't want to do check-ins because that was too manual, if you're checking in, you're playing on your phone, you're not really experiencing it. So we integrated Alohar to detect when a user visits a location on their Wishlist. This allowed us to send notifications to both users, telling user A that they have inspired someone to do something fun, and reminding user B that someone else loves this too, and offering them the opportunity to thank user A for sharing their experience. This was a big win for us because it's a hard problem to solve. It also gave us the opportunity to send recommendations to users when they visit a location and handle other place-local insights.
How we built it
We had big ambitions for HeyLets from the start. We wanted to build a robust, beautiful experience on the client and power it with meaningful content delivered quickly. We wanted to remain constantly up to date, because we're social, but we also wanted a good experience offline because we're travel. We settled on RestKit for its out of the box power and convenience for network and persistence which allowed us to focus on great interfaces. We centered our content around some classic social relationships (following members, comments and content creation) but also added a recommendation engine which applies machine learning to our member graph to determine relevancy, as well as a Wishlist allowing users to save things they want to do. We used AWS to host our back end for its wide array of services, flexibility and it's balance between cutting edge technology and relentless stability. We used Alohar to provide insights into a user's local world and connect them back to their in app experience.
Challenges we ran into
Recommendations are hard, and there's always room to improve. We've learned a lot about what people like to share, what people have in common and where people's similarities really lie. We came up with solutions that blend statistical learning, sociological observation, and just enough curatorial intervention to deliver the best content to our users.
Accomplishments that we're proud of
After testing a lot of different interfaces, we think we've nailed a design language that feels simple, but handles the complexity of our messaging well–and the session lengths we're seeing tell us it's fun to browse too. We knew that existing discovery products were failing to engage their users (DAU/MAU = 5 - 6%), and we kept on hearing how Yelp and Trip Advisor were a chore to use, and that local and travel discovery was broken. We felt that a window on to all the amazing experiences your city and the wider world has to offer could be and should be fun, entertaining and engaging, so for us success would be measured by how our users engaged with HeyLets. Was it fun to use, would it be entertaining as well as useful? After a lot of testing and countless iterations we are thrilled that over 10% of sessions are over 30 minutes long, median return rate in the first month is 18 times (more than every other day), DAU/MAU is nearly 40% (that's Instagram territory) and over 13% of registered users post content. We were aiming to fix discovery by making it entertaining, fun and engaging, and the early signs are that we've succeeded.
What's next for HeyLets
After lots of product iteration, we're in our growth phase. Building our user base and spreading the word, with a lot of exciting partnerships in place and more in the works. We're also about to start raising our seed plus round! Come have a conversation with us about anything.