Biz Stone describes building Jelly, a social search engine that pivoted away from its original vision, then unpivoted back. Anonymity drove rapid engagement, but viral press brought low-effort questions that strained the system. Stone explains why they stayed small while scaling on the cloud, and how voice assistants shaped their product direction.
Chapters
With Jellia, you've started out with this original idea, then you switched it up, now you're back. Talk about that process of when you decide that everything is— maybe the way you're approaching it is wrong. I mean, how do you come to that conclusion? How difficult is it to— and then you unpivoted.
Yeah.
You went back to your original idea.
I'm pretty sure that's going to end up in a Silicon Valley episode because we made up the word unpivot just on purpose to actually get some attention.
Yes.
So what we did was we made a rookie mistake. Ben and I had great lives going. Ben was— Ben's company had been acquired by Twitter, and he had like all this stock and all this stuff. He was a really great engineer, and he had a great team at Twitter who was doing really fun stuff. And he would have made millions and millions of dollars. And I was working with Ev doing— well, it ended up becoming Medium, but before it became Medium, we were just experimenting on lots of things. And I was— for a period of time, I was only in the office 3 days a week so I could be home with my new son at the time. And I wrote and directed a film with Ron Howard and Cannon. I was doing all these cool things, and then Ben and I came up with this idea that we just couldn't get out of our heads. And we were like— but then we ran it by all the smart people we know, and they were like, you should go for this, you should do it. And I just got sort of sick to my stomach, like, oh God, we're going to do this all over again, and I might ruin Ben's life because he could get all this money. Anyway, so we did it, and we were totally into it and super psyched and thought it was definitely going to be a thing. And then we did a total rookie mistake, which I tell other people not to do all the time, which is we let ourselves be talked into, this isn't working, we should change. We should just build something completely different that people want. And so we did that. Even though at the same time I was advising other companies like, no, no, no, you shouldn't build something else. Focus on your main thing that you originally thought was a great idea and just make it better. And meanwhile, I wasn't doing that. So we pivoted away from Jelly, which was our original concept, which was let's merge mobile and social and create a different kind of search engine.
And— Basically, it's a search engine where anyone can answer your question.
Yeah, it's different now than it was. The first gen depended on your social network. The new gen doesn't. Asking questions is anonymous, and it's more tech-centric, but all of the computer science is just to get your question routed quickly to a person who knows the answer so that they can answer it. Of course, it's like on-demand answers, so it may take a few minutes, but you'll get an answer, and you wouldn't have had to look Anyway, we got talked into switching, just pivoting into like a silly— we did notice on Jelly 1 that for every 1 question there was like 30 answers, so people loved answering. So we made this service that was like just answer an unasked question, basically. So we would start them off with things like the best, the worst, the craziest, the sexiest, and then they would just Finish that. And we made it really weird and wacky and wild and cool looking, unlike anything else. And then we found— Ben and I went into a room. The guy that had sort of convinced us to switch no longer worked at Jelly. Ben and I were in a room thinking about how we would shut down the Jelly service, because we never actually shut it down, which is telling. And then we came out of that room more pumped than ever to return to our original vision and try to make it work this time. And actually, something good came out of giving up on it for a while, which was the perspective needed to completely redo it. We could have kept at it, and we probably would have done all these little tiny iterations. That kept not working. And we wouldn't have taken such a huge step back and just been like, rip it all out, start over. So this gave us the perspective to see what was completely wrong about the whole approach entirely and reapproach it anew by saying, okay, what is the current media landscape, technology landscape? What would you do? How would you make search now given such things like mobile phones are practically an appendage? The new 6 degrees of separation is 4.
The rise of bots and Alexa and Siri.
Yeah, there's like a lot of— I mean, the UI between people and computers is going to change. And we're already seeing it with things like Siri and Alexa and whatever the Microsoft one's called.
Well, I'm gonna ask you one more question, but I want everyone to— after this question, we'll take yours, so start thinking about what you wanna ask. So now you have this smaller company. How do you decide how to scale it? When do you know that the time is right to hire?
Some people just get a lot of money and they just hire a lot of people, and then they hope that those people have a lot of work to do. And that's— sometimes that works. But Ben and I, when we decided to start the company, we decided that in the world of Amazon Web Services, you can have a small You can have a company of tens with participants of tens of millions or more. So we decided we wanted to stay small. So we're 8 people and 5 are engineers. But now we need way more people because we've— we've bitten off more than we can chew. When we relaunched this product, We were just flooded with way more questions per minute. The problem with Jelly at one point was nobody was asking any questions. So we completely changed that, and now there was like— we were just being flooded with questions per minute. And it was all—
You think because it's anonymous?
Yeah, it's totally anonymous. It's just like using Google. So ask away, you know? And also, we got a lot more press than we thought we were going to get, and we got— Featured by Apple, so we were just getting flooded. And a lot of the questions were kind of like kick the tires type questions, like why is the sky blue? Just because people wanted to see what it did. So we kind of had to scramble to create systems to take care of those types of questions and not route test-like questions to people who indicated that they'd be interested in. X, Y, and Z.
Right.
So the way we decide to scale is we now have several jobs that we could fill that would take up a lot of work. So that's how we're deciding to do it. But that has its drawbacks too, because then you're all of a sudden, you're kind of behind all the time. Because we— right now, we need to hire like 5 engineers, but we don't have time to interview any engineers because we have to constantly work. So what should we do? You know what I mean?
Right.
Should we spend— should each engineer spend an hour every day interviewing someone, or should they build something that's going to make the service better? So it's kind of like we're stuck. Until we can really get running ahead of the game, we can't really interview anyone properly. Because like I said before, you really gotta hire the right people. So maybe, I mean, when Ev built Medium, he hired a whole bunch of people. And for a while there, there were people that were like, kind of like, I don't know.
Hanging out.
I guess I'll just build something up on my own or whatever. And that was a problem, but, you know, he got through it and then Medium started taking off like crazy. And everyone was super busy and then they needed to hire more, but they hadn't They had the power to catch up, whereas we now are sort of falling behind because it's getting a lot more use than what we expected, and we need to put people behind that. So—
