Nextdoor: "We stopped everything we were doing and trained every engineer on mobile"

By Sarah Lacy , written on May 23, 2013

From The News Desk

Mobile first? How about mobile finally.

Nextdoor has finally released its long-awaited iPhone app-- long the most requested feature by users. Considering the trove of venture cash raised and how long Nextdoor has been in business, you can understand users' impatience. But Nextdoor's cofounder and CEO Nirav Tolia is hoping that taking the time to build it in-house and build it right will more than make up for the almost anachronistic lag in time to market. Particularly given that urgent notifications and communications around crime and safety has emerged as Nextdoor's clear killer app. (No potentially tragic pun intended)

First off, the app is exactly what you'd expect of a well-funded, modern social network. It's very pretty, with lush pictures of neighborhoods and some clever animations. It's easy to use, has simple icons, and there's a basic flow of news, like most social networks. Things like photo-sharing are emphasized, given how people use their iPhones.

But the story about how the app came to market-- and why it took so long-- is more interesting than the list of features. It wasn't launched on mobile first deliberately, Tolia says, because in the early days of Nextdoor users had to draw the boundaries of their neighborhoods. That was something you really need a full, desktop-sized screen to do. He points out that most of the major social networks were places that you'd sign up to on your computer, but use on mobile later. That's why networks like LinkedIn and Facebook had such rich profile information. By contrast, social networks mostly built out on mobile -- think Instagram and Path-- have very little personal information. No one is tapping out a work history on an iPhone; it's gotta be a few taps and you're done, or people will just give up.

The team didn't realize why mobile would be so central to its service, because it was really launched to be a more locally relevant Craigslist or Angie's List, not a way to make your neighborhood a safer place. You can wait to get home to get that referral for a plumber -- you can't wait to find out if there's an armed jewel thief loose in the neighborhood where your kids are playing. Sure, Nextdoor had text alerts and a mobile optimized site-- which made up some 30 percent of its activity-- but that isn't the same functionality afforded by a native mobile app.

When all this became clear, the company made a board-level decision not to optimize for speed, even though it was arguably already late to mobile. It went from not prioritizing mobile to believing that mobile was so important that it couldn't be outsourced or rushed. So rather than pay a design shop to develop an app, Nextdoor hired a team and even hired an agency -- Big Nerd Ranch -- to come in and teach all of its engineers to code for mobile so the whole company would have cross-functioning skills. Big Nerd Ranch is the same firm that Facebook used when it too belatedly got serious about mobile. "We stopped everything we were doing and trained every engineer on mobile," Tolia says. "We knew training our internal guys was going to take longer from a  development standpoint, but we thought it was that important to the company's future."

The development process wasn't optimized for speed, either. They started out by ripping out everything that wasn't necessary from the desktop version, focusing on building the quickest, cleanest, leanest possible minimum viable product. (Not exactly the Facebook approach to mobile...) As part of that, it wasn't possible to sign up as a new user in the app, because signing up as a new user is a pain in Nextdoor. You have to have your address authenticated, because the core organizing principal of the service is the idea that only your neighbors-- your actual, real neighbors-- are in your network. That allows you to share information you might not on a regular social network.

When its minimum viable app got close, it was Apple that recommended Nextdoor go back to the drawing board. The advice was that if the app did as well as expected, people would be frustrated they couldn't sign up immediately... and they likely never would. You don't get a second chance in the App Store, Apple warned.

Nextdoor listened and went back to the drawing board. It tried to come up with an easier, more frictionless way to sign up using the phone's GPS. Everyone loved the idea, but there were two clear problems with that. The first is that there's no reason to assume people would only try to sign up while sitting at home-- after all, the very point of mobile is that you can do it anywhere. The second is that people could enter other neighborhoods and pose as residents. Maybe only a few wackos might drive to Beverly Hills and pretend to live next to Ceelo, but you only need a few to destroy the site's painstakingly-earned user trust. The GPS version was thrown out, and the typical authentication methods -- entering a credit card or waiting to get a code off a snail-mailed postcard-- were added back in. While that will cut down on new user sign-ups, it's the only trade off Nextdoor could have made, given how important privacy and safety is to the site.

Regardless of the time to market and the challenge on user sign-ups, the app will absolutely change how people use Nextdoor. I live in one of the most active San Francisco Nextdoor neighborhoods and there's just a massive difference between me remembering to check Nextdoor when I get to my laptop (which I never do unless I need something) and being able to passively scan a newsfeed while I'm waiting in line for something-- or, better yet, get an urgent alert pushed to me. That's just become how we all consume social networks and newsfeeds, and if you're not in that game, you're not competing for a user's downtime.

To make the point, Tolia pulled up an example from his neighborhood feed of a guy who broke in, and the surveillance photo of him. In a desktop world, you might never see that photo or easily forget what the guy looked like. In a mobile world, you see a suspicious guy on the block, pull out your phone, and can match him to the photo and call the cops. And that's just one hypothetical.

Tolia expects photo-sharing to be huge on the app, and for it to be different from how other photos are shared online. On other social networks photos are a filtered, cropped, artistic representation of things like, say, your french fries or your baby's foot. (Guilty.) For Nextdoor it'll likely be tactical and about sharing information. Think: Look at this pothole; here's an accident that's closed off the rest of the block, so you might want to take another route to work.

One of the coolest features of the app is the directory. All of your Nextdoor neighbors are sorted into something that looks like an old fashioned phone book. You have whatever contact info they've opted to share with neighbors at your fingertips-- and can search by name, address, or see everyone on a certain street.

If your neighbor left a garage door open and you're walking in the neighborhood, you can instantly find out their name-- even if you've never met them-- and shoot them a message to let them know. It's rare that I can say with confidence that a social network is actually doing something first, but that is a first. I have no idea how I would find out a neighbor's name, let alone message them. Just like how email brought back an era of "writing letters," Nextdoor's app uses technology to hearken back to what neighborhoods used to be like.

Of course, given features like that, you can appreciate why Nextdoor had to take authentication so seriously. One funny tidbit from the demo emphasized this tug-of-war between ease of sign up and authenticity of sign up that the company is just doomed to wrestle with. Tolia was showing me the directory of his neighborhood, which is also essentially a directory of Internet entrepreneurs who've already had an exit. (My neighborhood is more like the directory of Internet entrepreneurs pre-exit...) Users can opt to obscure their street address, and he was making the point that very few do when he noticed that Matt Cohler-- a partner at Benchmark, one of Nextdoor's investors -- had obscured his house number in the system. (Note: Cohler is an investor in PandoDaily, as is Nextdoor's other investor Greylock.)

"Ohhhh really, Mr. Cohler?" Tolia said laughing. It turns out that Cohler had been one of the people arguing the whole network had to be opened up to all users early on. As an investor, he wanted speed and virality. But clearly, as a user, he wanted privacy. Nextdoor is a rare slow-and-steady Internet company that's voting with the users.