(techno music) >> Narrator: Live from San Francisco, it's theCUBE covering AWS Summit 2017. Brought to you by Amazon Web Services. >> Hi, welcome back to theCUBE. We are live in San Francisco at Moscone Center at the AWS Summit. And I'm Lisa Martin, joined with my co-host Jeff Frick. We're very excited to have the chief architect and co-founder of Nextdoor, Prakash Janakiraman, on the program, welcome. >> Thank you for having me. >> For those of you who missed the keynote this morning, Prakash, you did a really fantastic keynote in the session that Werner did, and I loved how you positioned what Nextdoor is. If there's anyone out there that actually doesn't know, give us for those who didn't see the keynote just how you positioned it and talk to us about what you guys are doing, how you've achieved 70% of US neighborhoods covered, over 135,000 neighborhoods in the US. You're expanding globally. Give us in a few minutes this back story on Nextdoor. >> Yeah, so we think of Nextdoor as sitting kind of alongside Facebook, Twitter and LinkedIn. People are using Facebook to communicate with their friends and family all over the world, and Twitter to connect with people with whom they share common interests. And despite having hundreds of Facebook friends and thousands of Twitter followers, in my own neighborhood I only knew a couple of my neighbors, and it's sort of unfortunate. Your neighbors are a great resource that you should know, right, in a local context. So we feel like Nextdoor is bringing back a sense of community to the neighborhood by connecting people in the medium that we use today, which is this virtual medium, online social networking. So we kind of sit alongside of those. And as you think about the difference between what Facebook has been able to do, or an Instagram or a SnapChat, they really leverage existing connections, and when most people don't know their neighbors we have our work cut out for us, because we need to go build those connections. And once those connections are built, they're really, really valuable, right? All of a sudden you didn't know how to find the local electrician or a handyman, or how to reunite with a lost pet, or what happened with this rash of break ins in the neighborhood. Or who do you go to if there's a natural disaster that's happening in your community? How do you band together? How do you even communicate with other people? And so we think we're bringing that basic connectivity to these neighborhoods. >> What was the secret sauce? Because obviously you can't just authorize integration to my contacts, because by definition they're not in my contacts. So how did you kind of crack that code? >> There were a couple of things. So number one was we made a very deliberate decision, we'll get to why this was important, to verify each of our members as being actual residents of a neighborhood community. In our parlance, a neighborhood is really a geographically bounded region. Every single neighborhood in the Nextdoor (mumbles) is started by one member. The member draws the neighborhood boundaries, and then subsequent members-- >> Jeff: They draw the boundaries. >> They do, they do. >> They define it. >> That's right. And so every member that joins has to actually prove that they're residents of the community, and that raises trust. And all these neighborhoods have grown because of word of mouth. People go out, they tell their neighbors, "Hey, you should come on to Nextdoor. "It's a great place for us to connect. "I just heard something interesting "that happened on Nextdoor." But then we also use a variety of different mechanisms to invite people that are a little unusual for internet companies. For example, we allow our members to send out postcard invitations, because again, they don't have phone numbers, they don't have emails or electronic correspondence. So we've had to resort to some non-traditional types of marketing to get the word out about Nextdoor. >> Don't share any secrets, but what percentage of new members come through a postcard sent by a neighbor? >> Well, not isolating postcards, the vast majority of our members are coming in because of a referral from another member, of which postcards are one of those mechanisms. >> Are one piece. >> Yeah. >> With all apps and stuff there's kind of the moment, right, the CNN moment, whatever, where things kind of catch. Is there a particular type of activity in the neighborhood that is the one that kind of catalyzes people where you get that little spike? >> Well, it depends. In a lot of different cases, I remember an example from really the early days of the company, a community in Woodside, California, which is just right down the road from here, a small child in that neighborhood had developed meningitis, and meningitis if you don't know is hugely contagious. It's really bad news. And so lots and lots of members got the word out about this, and that brought new members in because they were saying, "Well, wait a minute, what do I do about meningitis? "Do we need to get immunizations? "Do we need to do something else? "Do we need to quarantine our kids?" So sometimes it's bad news like that. In other cases it's good news, it's about neighborhood events, it's about bringing people together. One of our very first neighborhoods organized a Halloween party, a Halloween parade in their neighborhood, and they used Nextdoor to organize the entire thing, and they had like 50% participation in the neighborhood. It was a small neighborhood of 90 households, but to see all the kids, and them taking pictures and publishing them on Nextdoor, it created this connectivity that just didn't exist before, so. >> And that's such great point, because we are so used to be connecting with everything, and we think we're so connected, we all have over 500 LinkedIn connections, and Twitter followers and Facebook friends, but it's very abstracted. It actually can be very isolating. So it seems like it probably was like a no brainer from an investment perspective to go, this is the, as you said in your keynote, paraphrase, this is the original social network. So talk to us about how you built that on AWS, how you've leveraged the power of AWS, the technologies, to gain what you have done so far. >> It's interesting, because prior to starting this company I was at Google, and we always felt like one of our technology advantages at Google was that we had near infinite computing resources, and that the computing resources were abstracted away from the developers so that they could just go in and build applications and not need to worry about what actually powered those applications under the hood. You said, "I needed some RAM. "I needed some disk. "I need some CPU," and boom, you're off to the races. So coming out of Google, and this is a long time ago now in 2008, seeing that AWS was just getting started with EC2 and S3, and that the playing field was leveled for developers in such a way that you no longer had to rack and stack servers, you didn't need to have a physical data center, you didn't have to have people managing that data center, and that you could use hardware as software really, you could interface with it as software, was hugely powerful. So we started out on EC2 and S3, and over time, and as we fast forward to today, we're using almost 30 different AWS services, including DynamoDB, ECS, Redshift, Kinesis, and so all of these components fit together in running our business. >> Right. Early lines, right, for AWS. It's a data center as an API is one of my favorites, just those little lines, but one of the, so security, all that stuff is kind of done talking about security for a publisher in the cloud, but one other thing that still is out there is at some point, you know, you rent, rent, rent, and at some point you pass a milestone where it's more economical to buy. Clearly you guys didn't buy. Netflix hasn't bought. There's plenty of cases, but from your point of view as a founder and a businessman, and also an operator, when you looked at that, has that ever come up in your discussion as your Amazon bills get bigger and bigger and bigger, and if so, kind of what's the internal discussion and why are you where you are today? >> So we, like a lot of startups, we really value speed and we want to focus on the things that we're really good at, and the things that we're good at are building great products. So we are so advantaged by the fact that we do not have to manage our own infrastructure, that we can use these component parts from which to build our applications, and nowhere is this more apparent than in prototyping new features, right? When you don't necessarily know if something is going to work, you don't want to go make a big investment in procuring resources for that. You want to just be able to spin something up on demand and try it out, and if it doesn't work you shut it back down. And so for us it's a TCO, total cost of ownership, type of question. And along with having on premises data centers and infrastructure comes overhead for maintenance and having teams that do that, and right now, for all of these neighborhoods that we support, 135,000 neighborhoods in the US, right, we only have two people that manage our underlying infrastructure, two dev-ops folks. >> And how many people do you have in the whole company? >> The entire company is about 150 people, and we have 60 of them in engineering, so we're a really, really, really lean team, and we're supporting massive scale with that small team because of the help of something like AWS. And this would've never been possible 10 years ago, so. >> And you'd rather spend that next marginal dollar on another developer, or marketing, or sales, or community, or something, than a piece of metal and a rack. >> We want to put that money towards making a better user experience for our users, and whether that means going into new communities, places where we aren't already, or improving engagement in communities where we already have the product, in either case that's a better dollar spent for us. >> Jeff: Right, right. >> So along those lines, how do you do that? How are you taking user feedback and determining kind of what the next steps are even within the US, but also is there, as Andy Jessie actually talked about in his fireside chat, the GOs that they're expending, they're in 16 now, is there maybe a model to kind of follow along where AWS has a footprint? I'd love to know what that customer engagement is to help really refine that user experience. >> Yeah, there's a few different ways that we can get customer feedback. One is sort of implicit. So we have lots of data that we collect and look at and trying to understand like what features are really popular, what features are taking off, what features aren't really as popular as others, and we do a lot of A/B and split testing to understand how these features perform. But the other way is just good old fashioned qualitative feedback from our users. So we have a couple of different ways that we do that. One, we have a neighborhood operations team that's always in contact with our communities. We have leads in our communities that serve as sort of moderators and the folks that manage the community's activity, and so we lean on them heavily for feedback. And then we have a national leads forum where they all come together and they exchange their own experiences building their community. And from that, it's a treasure trove of information where we communicate with them, we participate in those forums, and then we build the features that we think that they want. >> It's funny, this theme of competing for speed, you're competing against time has come up a number of times today, in Werner's keynote, you mentioned it, it came up again in Andy's fireside chat. I thought it was interesting in your keynote piece how the measure of speed and deployment has so radically changed. It used to be one push a day was crazy. And you talked about a concept of democratization of deployment, which I thought was interesting. Because in the big data world we're always talking about democratizing big data, get it out of the hands of data scientists, let everybody make better decisions based on data, but you talked about democratization of deployments which I've never heard before. Wondering if you could share that story, because again, one a day, that's no longer the measure of success. >> No. So the way that we think about this and the reason that I call it democratization of deployment is as you invest more in automation and you turn down the kind of batching of releases together, the batches of commits together, you can now isolate their impact, and you can move faster by doing more frequent, smaller releases. And part of that is each developer is responsible for their own code going out to production. So what we've invested in is a lot of automation that makes it possible for any one of our developers to push a button, a metaphorical button, and actually push code out to the site, observe how that code performs, and then we can move onto the next one. So in the old world people would often hire a release engineer, and the release engineer's job was really to coordinate all of the activities of the developers, build packages, get them staged and ready for QA, and the overhead on that just does not scale well as a company gets bigger and bigger. And so what we've said is, "Listen developers, "you guys are responsible. "You guys know how your code works. "Let's make it possible for you "to minimize the amount of time "between you committing a piece of code "and that code being live to our users." And so that's what I kind of was talking about in terms of democratization of the release process, making sure that everyone can do it. >> What I thought just was fascinating is on one end everybody is talking about smaller, marginal units of compute and store, etc., right, and yet on the other hand you were basically doing, basically duplicating your entire production environment in kind of this red black strategy, build it up, make that one go, crash the other one. Build up a fresh one, make that one go, crash the old one, and leveraging what are basically infinite resources behind the AWS screen in a really innovative way that you could never do that with your own hardware. >> Well, a lot of credit goes to Netflix. We saw a presentation that they did a few years ago at Reinvent that inspired us to build our own kind of version of the red black deployment system. But again, because these resources are ephemeral, you can spin them up and shut them down, that's what makes any of that possible. So it's really, really awesome for us, because we're moving fast. In fact, I don't even know how many deploys we've done today, but every time a deploy is done we get a message in a Slack channel that says, "Hey, somebody pushed." And it says who the conductor of the train that actually pushed out the commit is, and so you have full accountability in this like living log of what's going on in your production environment, and anyone can do it. Even our interns are allowed to do it, and so it's really, really empowering for our developers to be able to do that. >> I was going to say, that was the word that I heard when you were talking earlier. It's really this empowerment, which is huge for productivity. And speaking of productivity, you guys have really achieved some pretty significant technology achievements that are presumably in this feedback circle going back and making the experience for your mother in law, your mom, probably my mom, even better. What are some of those technology achievements? You mentioned a few of them on the keynote. I'd love for our guests to hear some of those big things that you've achieved. >> Yeah, I think when we think about the technology stack there are two things that we're trying to do. Number one is build for our developers so that they can push features out faster, but then the second and the most important is that we improve the user experience for our users. And most recently, we've done a few different things. Number one is we've introduced this concept around personalization, and it's very lightweight at this point where we can try and understand, what is your experience? If you're not a parent, do you really care about play dates and all of the information about kids? If you don't have a pet, do you care about what's going on with pets? Do you not like the classified ads? And so we're starting to look at data, this is where the big data stack becomes very interesting, and say, "Okay, let's see if we can "match the content in your neighborhood "to the things that you're interested in," and sort of making a lot of investments in that area. And then the second is we're making a lot of investments in search, because we think a lot of people will come to Nextdoor with an intent in mind. They'll say, "Hey, you know what? "I need to find a handyman or a plumber." And what's the natural way to do that is to go into a search box and say, "Handyman," right, and look for handymen in your local community. And so we're making a lot of investments on both fronts that will make it easier for our users to get connected to the information that's most important to them. >> That personalization is so key because we sort of, as consumers and as people who expect things immediately, we want what we want, we want what's relevant. And we know a lot of companies, whether they're banks or whatnot, retailers, spend a tremendous amount of time and dollars investing to know how do I make this experience unique to Jeff, or unique to Prakash, or unique to Lisa. And so I think that there's so much capabilities. I have to ask you though, you must have the most connected neighborhood. You must have the best block parties. (laughing) >> Me personally? >> Yeah. >> You know, it's interesting. The product takes on the identity of each of our communities. And so in an urban community like in San Francisco I'll say that I still don't know all of my neighbors, but it's really, really comforting to know that I know by name and by face who they are even if I haven't had in person interactions. But one of the things that we always talk about, which is pretty cool, and it's happened to me in my own community, is we talk about this concept of bits that move atoms, and using our platform, which is really exchanging information electronically, to facilitate in person interactions, and those are the atoms. And it's definitely happened to me where I've seen someone on the street and I just say, "Hello," right, because I'm like, "Oh yeah, I saw your post. "Did you get your dog back? "Were you able to sell that television?" And so it's really interesting that this is actually happening, and ultimately that's our goal is to bring back that sense of community to neighborhoods and to make neighborhoods stronger and safer, and the way you do that is by connecting people in the community. >> Right, right. >> Absolutely. >> It's a neighborhood watch on steroids kind of, right? >> Totally, yeah. >> Yeah. >> It's fantastic. And I'm sure they would be blown away if they knew they were talking to the guy behind Nextdoor. >> Prakash: One of them. >> One of them. >> One of them. >> Well Prakash, thank you so much for sharing your story with us on theCUBE. >> Thank you very much for having me. >> It's so exciting to get back to the original social network and really kind of, I love that. I think people might even be learning how to write again rather than type. >> My handwriting is terrible. >> We can only hope. >> Reaching, you're reaching there. >> Well Prakash, thank you so much. Continued success with Nextdoor. >> Thank you very much. >> We so much appreciate your time here. >> For my co-host Jeff Frick, I'm Lisa Martin. You've been watching theCUBE. Don't go anywhere. We'll be right back. (techno music)