Getting more butts in seats

Yelp’s foray into the reservations market was a shrewd choice. In the mid-2010s, Yelp’s leadership saw a surge in popularity for online reservations platforms like OpenTable and Resy and thought that it would be perfect for Yelp users to be able to book reservations in the same place they look for restaurants. Yelp initially partnered with OpenTable for a time, but eventually that business relationship soured, so instead Yelp acquired a San Francisco-based startup called SeatMe for $12.7M. For such a young company, SeatMe had a stellar front-of-house table management app, but they weren’t able to drive much traffic to their restaurants.

The Yelp Reservations table management system circa 2016

In the summers of 2014 and 2015, I was working as a Yelp intern on the SeatMe team right as they were integrating their platform into Yelp and eventually rebranding into Yelp Reservations. I worked on creating some simple but important parts of the experience, like a printable ‘cover report’ that summarizes all of the guests that a restaurant can expect in day and a configurable widget that restaurants can put on their website to drive additional traffic. Both of these features are still in use today, with only a few changes introduced along the way.

When I started full-time at Yelp in 2016, I was tasked with focusing on the diner experience for making reservations. There had been a lot of focus on making the front-of-house app a stellar experience so that it was easy to pitch it to restaurants, but beyond a simple module on a restaurant’s Yelp page, there wasn’t much happening to nudge our users towards making a reservation.

The first few projects I worked on were quick wins meant to improve conversion and do what restaurants need most: get butts in seats. One was a no-brainer: add a search filter allowing users to look for reservations at a specific time across all restaurants in an area. This puts users into a high-intent reservation search flow, where the specific restaurant is less important than finding a slot for a certain party size at a certain time. We made the choice to call out which reservations are available around the specified time directly from the search results screen, so it’s easy for users to scan all the listings and make an informed choice with minimal back-and-forth. We needed this high intent search to work seamlessly on both desktop web and mobile, since plenty of people use their computers when trying to make reservations, so we maintained the same approach with some visual differences on each platform.

We also learned from restaurant feedback that the widget experience was falling short of expectations because of some technical issues. Many users were complaining that the widget was broken and not completing the reservations they were trying to make, so we needed to rebuild the widget on new architecture and make a more reliable solution. We also decided that instead of handling all of the booking interaction and edge cases inside of the widget, we would redirect users into a very speedy (and responsive) microsite where users could complete their reservation and see even more available times.

The Yelp Reservations widget for restaurant websites

While opening a microsite does lead to a context shift from one site to another, we saw from our A/B testing data that booking conversion from the microsite experience was far higher than from the widget alone. It was also good from a development point of view, since it was easier for us to maintain the microsite instead of building many complex interactions into an iframe widget (which is essentially a website in a small frame on another website). Since we were already rebuilding it, we took this opportunity to make our widget more refined and include a dark mode variant, which was great for restaurants using a darker color palette on their websites.

Our responsive microsite for reservations. Super fast and consistent across device types.

The last significant improvement I tackled was the reservations module on a business details page, which is basically a business’s main presence on Yelp. Reservations are important to many restaurants and bars since they drive paying customers into the location, but they aren’t the only way Yelp can help businesses. A large percentage of Yelp listings also feature food delivery options, and users might also want to write a review, look at a menu, look up directions, go to the website, call the business, etc. While reservations are undoubtedly important, they’re just one piece of a bigger pie, and we needed to find a way to feature reservations prominently without undercutting our other business verticals.

We experimented with a few different variations of the reservations module, but many of them ended up being too prominent and took up a big chunk of space on the page, pushing other important actions and info down further. In the end, we opted for a simple module that links to the same microsite the widget uses, but with one key difference: logged in Yelp users would have all of their information prefilled and would be able to complete their reservation in a single click.

We collaborated with the delivery teams and our VP of Product to establish a stack ranking system to prioritize how we display Yelp features, including reservations and delivery. Since reservations drive pure profit for restaurants, we stacked that first, with delivery below, and all other actions further below that. This also carried over to search results, where we focused on one primary action (i.e. ‘Book a table’ or ‘Order delivery') for each listing and gave other possible actions a minimal visual treatment.

A waitlist with less waiting

Keeping up the theme of shrewd acquisitions, Yelp acquired a company called NoWait in 2017. NoWait’s product was simple: an online waitlist for restaurants that users could join from afar. Imagine planning a brunch at a local spot that doesn't take reservations, but you can see that they have an expected wait of ~30 minutes and join the waitlist while you take the bus over. NoWait was surging in popularity in San Francisco and a few other urban hubs, so it looked like a perfect complement to reservations and food delivery as a way to get more people using Yelp to get food.

As with all acquisitions, it takes time for the products to become truly integrated, but I found the process of integrating NoWait into the Yelp consumer experience to be very smooth and very fun. We changed the name to Yelp Waitlist for the sake of simplicity, and integrated modules for waitlist customers into business pages showing live wait times and a button to join the waitlist. We also looked at combining the waitlist and reservations modules into one dynamic block, but some quick user testing showed that it this was just overcomplicating something that should be simple.

Yelp Waitlist kiosk. Diners can join the waitlist either from their phone or at the kiosk.

other contributions

Credits