This article was published on March 5, 2011

Greplin Founder Daniel Gross on his amazing story behind building the company [Interview]


Greplin Founder Daniel Gross on his amazing story behind building the company [Interview]

Last week I spoke with Daniel Gross, the founder of Greplin, a Y Combinator (YC) funded startup that is generating a lot of buzz in Silicon Vally right now.

Greplin allows you to search though your social-networking accounts on Facebook, LinkedIn and Twitter, and online services such as Google Docs and Email. It recently announced a $4 million venture round funding from Sequoia Capital, arguably the best VC firm in the world.

Gross is one of the youngest people to launch a YC startup at the age of 18. We talked about the inspirational journey behind Greplin’s creation.

The tale is intense and powerful: Gross was flown in from a different country, where his original idea for YC was dumped, leaving him with no startup. Shortly after joining his co-founder abandoned him. When he did eventually find a product he believed in and built, Amazon ‘killed’ it 48hrs before Demo Day and when Greplin was born he almost ran out of money to keep it alive.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

Now Greplin is one of the hottest startups in the Valley. Here is his story:

Could you just give our readers a quick introduction?

I originally did Y Combinator about 13 months ago. I flew out here to the states from Israel. I think at the time I was actually the youngest YC founder. I’m 19 today. I was 18 when I did Y Combinator. Unfortunately I’m not the youngest anymore because I think they just funded a 17 year old.

At some point they’ll be hanging around the incubators in the hospital just poking at the next founder.

So I flew out here to do YC. When I submitted the application, I didn’t really expect to get in. I was looking forward to a paid weekend in San Francisco at the very most.

Were you in college at the time?

No. I was right out of high school. I flew out, did the interview. It went fairly well. And we were one of those people who, during the interview with Paul Graham, the idea that we were pitching received very unfavorable reviews from him. He sort of hated it.

When you say ‘we’ who else was with you?

I flew out here originally with a co-founder of mine, who is a buddy from high school. But who I am not working with today. We were in the interview and Paul was not really liking the idea and we were sort of getting bummed out. I remember after the interview feeling pretty much set to head home. As we were leaving YC, post interview, Paul comes out of the room and says “Hey, could you come back in for a second?” We went back in the room and he said “You sound really interesting and like you like building things. We don’t like what you’re working on now. How about you come do Y Combinator. We can build something different together”. It was a tough decision to stay and dump what we were building because I liked what I was working on back then. But in hindsight, it was not the best product.

How did you become a programmer?

I guess the story is rooted with my father. He is a CS guy from Columbia University and he teaches a lot both in high school and higher education. Computers were always something that were around at home. And I was always intrigued by them. Most of what I’ve learned today is self-discovery or me having  an interesting question, and his ability to answer it and educate me on the spot. I was probably around 10 years old I when I started programming, nothing major, just basic stuff.

What was the product you originally came to YC with?

It was a social marketplace a little bit like eBay. You could actively see who was bidding on certain items and who got a good deal and who got a bad deal. It’s one of the those ideas that sounds intriguing but no one really wants to use.

So we decided to start a new idea with the help of YC. We came into YC fresh, with no startup. We had our office hours with Paul and we started thinking about what to work on. And we iterated on a few different things with varying degrees of success. Nothing stood out. At the same time my then co-founder was getting tired with this cycle of testing ideas.

How long did it take you to go through office hours and decide on something?

The general space that we were working on at the time was around products built for shopping online and catalogues. Paul had a few different ideas in this space. Many of them were good. We probably just, in all honesty, didn’t execute very well on them. Mostly because we were building products that we wouldn’t consume. I feel like everyone’s talking about this with startups. It becomes very hard to execute correctly on the products when you’re not the user.

You’re sort of shooting in the dark. You don’t really know what’s good and what’s bad. For example, one of the things we were building was a way for people to build their own stores online. I’ve never built an online store and I’ve never had a need for one. And in a situation like that you end up making product decisions based on zero information and experience.

We were doing that for awhile. I think my then co-founder got a little frustrated with various things. When you read about startups when you’re not in the valley, it feels as if everyone comes out here, does something very simple and it succeeds. You don’t see all the hard work. So when you come out here you expect that very quick response and when you don’t get it, it can be very frustrating.

So he decided to leave. And then I was on my own. And I was working on this product that I was pretty excited about. It was essentially a new way for people to donate to charities online. What you do is you install this browser extension on Firefox or Chrome and once you install it you can pick a charity you want to donate to. And every time you bought something online, money would passively get donated to that charity. It was an online affiliate system, which essentially let retailers like Amazon or Zappos reward publishers for clicking on certain links. So if I show you a link to some pair of shoes on Amazon and you click on it and you buy the shoes, Amazon will give me a cut of the sale to charity. So that is what my extension did. That was quite popular and people used it and it made money. Things were going fairly well for the time that it was alive.

How long did it live?

This was a fairly short product span. It took me 2 weeks to knock it out. Building it was easy. This took about 3 and a half weeks. Getting the different merchants to sign me on as a certified publisher was difficult. But people seemed to like it. The exciting thing was charities could now send this out to their followers and say “Hey guys, install this extension. And then you’ll effectively be generating money for us.” So, it was pretty neat.

The bad news, unfortunately, came 48 hours before demo day just as the product had gotten a little bit of traction and hit TechCruch, Amazon revised their terms of service. They made one small change. And the change was that they didn’t allow guys like me to donate affiliate proceeds to charity. Which essentially made what I was doing illegal and against their terms of service. So that was a major bummer.

So what did you present on Demo Day?

I’m sure you know this from Immad [Immad Akhund my husband and YC alumni], nothing is more important than Demo Day when you’re a YC startup. So I was a little bit bummed out and I remember walking over to PG’s house. I had this stack of ideas that I was excited about. And he said something and this resonates, I think, with everyone. He said, “Don’t build something that you won’t use. Take all of these ideas, pick the one that you would use today if you had it.”

Greplin was very clearly the one for me. I felt the problem Greplin would solve was a problem I had that was getting worse. I couldn’t understand for the life of me why, if I want to find my stuff online, why I can’t search through it all. There’s this great tool and with one really simple search box called Google that I can go to and type in anything I want and get results. Why doesn’t that exist for my stuff? It should exist but doesn’t.

So I got excited about the idea. I built a really basic, broken prototype in 48 hours. Funny story actually, on how the name got picked out. I was sitting with Paul and the founder of now Hipmunk, Adam Goldstein was in the room. We were trying to figure out names for what to call Greplin. We were throwing different stuff on the board. Nothing was really resonating. Then Adam sort of just said it out loud and it caught. Another funny thing was the registration of the domain was so new that by the time I got up on stage on Demo Day an investor actually tried to email me, but the domain was so fresh that it actually bounced back. So I missed my first introduction from Demo Day, because of that.

What were you feeling and thinking during the Demo?

Oddly enough it was a weird time for me because I was building Greplin. A few other startups were interested in seeing if I would consider joining as an employee. This was all happening very much at lightening speed. So I got up on stage on my own. There I was, this 18 year old kid with my hair all messed up because I’d been up for 2 days. It probably wasn’t too awe inspiring. I demoed the product. I got, I would say, a lukewarm response at best.

I remember Immad coming back from Demo Day and saying that your idea was one of the best ideas that day.

Neat. Ok, cool. I’m glad Immad was always a believer. There was some quantum of that. I think that’s what excited me, that there was at least one other person besides me who liked the idea. That’s all I really needed to affirm that people actually wanted this.

What happened after Demo Day?

So after demo day I had 3 main issues. One was that the product didn’t actually work. The second was that here in California, I didn’t know anyone, and I was running low on YC’s check. And the third was that I didn’t have a co-founder. So I pretty much set out to try to solve all of those problems. Thankfully, they worked out.

So I spent most nights trying to build up a workable version of the product at random coffee shops and startups offices. And during the day I would try to meet angels. I spent about two months doing this. Initially, it took awhile because fundraising is very much a snowball effect and creating it is the hardest step. It got fairly close to all falling apart. The day I closed my first angel round was pretty much the day I ran out of money. The bank account was literally at $300.


How did you find your angels? It seemed like a very dire situation.

The first angel that invested came through Demo Day. It took them a lot of time to write me a check, thankfully it was a large amount. I went from zero to $200,000. For them to trust a kid with that amount of money, quickly, required learning who I am and who I was-quickly.

I would pitch people about the product and then ask them their opinion about it and sometimes they’d point out really critical things that I’d failed to note. For example, the initial basic version didn’t let you add multiples of the same data source. So once you want to add your Twitter account, you can’t add another Twitter account. An investor pointed it out and said, “Hey you should totally do this.” Then I came to the next meeting and it was done. I think that’s what excited a lot of people, being able to move quickly. In their minds, they’d draw the juxtaposition of this is what he did between three days ago and today, where would this graph [rapidly iterating] take you next year?

How long was it after demo day that you closed your first angel?

Probably a month and a half. My first angel was oddly enough, not well known. They’re called SK Telecom. They’re basically what AT&T is in South Korea and they have an investment group here. They got us a fantastic deal. They got rewarded for their investment and early belief in us.

What happened next?

Once I had their check I was able to breathe. I kept on iterating on the product. The co-founder problem was unresolved and an interesting one. I’d raised about half a million dollars. Which was enough to get an initial feel for the product. And I basically had given up on finding a co-founder, because it becomes really hard to find someone who’s incredibly technically proficient. Which is really what Greplin needed.

In all honesty, I needed someone who was much better than me at engineering. Building a search engine, as you were saying before, is really hard. So I needed that. I also needed, the other end of the spectrum, this guy, whoever he or she would be, to be someone I could spend 20 hours a day with. And finding the intersection of these both was really hard. I practically gave up on it. But out of the blue, Harj [Harjeet Taggar], who was one of the partners at YC, made the introduction between myself and Robby [Robby Walker].

I had read a little bit about Robby before. And the back story on him is really interesting. Robby’s basically one of those kids who started college at 9 and graduated at like 20 with a Ph.D. I don’t know the exact age. He applied to Y Combinator in 2007 and was working on basically, what is now Google presentation. Google, I think, saw them once and was extremely impressed by them. He was acquired by Google in 2007.  So, Harj made the introductions. I met with him a few times. We hit it off. And to his credit, he took a big jump. He quit Google. Joined a 19-year old. Things have been going well since.

I’ve noticed the Greplin pitch has changed in the press. Now it’s more social, before it was search everything. Was Robby part of the change?

This is sort of a weird perception that came of out a TechCruch article. It’s interesting because we’ve been very tacit about this change. People have started calling us different things based on how they perceive us. So we’re effectively doing a few things. We let you search any single data source you have online that we support. That’s our goal: Imagine next time you’re on route to a party and you need the location of it but the information is buried deep in a Facebook event somewhere — You really should have this box where you can just go to and search every single data source you own. Sort of like what ‘find’ was on a computer before everything moved to the Internet.

So in 1999 you fire up a computer and you look for something and it’s probably in your computer or its not there. Today, that’s not true anymore, because nothing really important is on your computer, right? Everything is posted on some other service. Somewhere in the cloud. So with that in mind, it seems that people get, or Silicon Valley, gets really excited about us, because it happens to be on a totally separate tangent: search within a lot of these social applications was executed really poorly. So a lot of people just use us to search their tweets on Twitter especially because Twitter’s search function isn’t great.

The product vision is pretty much the same. But the way people use us and perceive us is constantly changing. And the thing is, it’s really a learning process for us because we’re defining the market. No one has really done this before. So we’re very attuned to ‘what do people think we do.’ It’s probably very different from what we think we do and we should probably listen to that more. I think that because the product is so complicated, people distill it into these terms.

Do you wish you’d made something more simple?

I think you said it very well. We were doomed with this problem of messaging [what Greplin does]. It is really hard for us. We’ve distilled it to a very simple functionality. We get a lot of feature requests that don’t involve search. People want to browse their online data, or provide some sort of journaling or catalogue. In my mind that’s overdoing an already hard to execute feature, which is just to let you search everything you own.

We focus a lot on making the actual product vision simple. I think that we’ve done a fairly good job with it. But it’s really hard to do with our product because we’re not making a better X or an improved Y. We’re creating something new. So it’s going to be a real challenge for us and especially when it comes to press and more mainstream media of distilling what we do into something that everyone understands. Mostly because I don’t think everyday people appreciate the notion of cloud data.

Everyone’s very excited at the moment about Greplin. Why do you think that is?

Personally, what I’m really excited about and I saw this happen with Twitter a lot, is that users were flowing into the system and staying. There also seems to be this “Ta Da!” moment when someone tries out the product and just gets it. The home page doesn’t really explain what we do. Again this is the same problem: How do we distill with images what we do? It’s kind of hard. So we just shuffle through some screenshots of the product or try to show a video because it’s a good first pass.

When people sign up for the product and when they log in and they add a few data sources, then they search, they have this incredible realization of “Wow, now I realize why this thing is useful.” It’s fairly frustrating for us because we’d love for users to have that moment of realization before they sign up. We’re trying to figure out ways to do that. This will ultimately fall on the very good design of a story compelling without having to register.

However, that doesn’t really answer your question. I think people [press] are experiencing this same magic moment. This is not a very well formed thought, but I think people used to have this model in their mind of information, and information lies in specific cabinets, right? You have a filing cabinet in which you have files, and you have a home in which you have furniture, and this information used to apply to the digital world. You have a physical hard drive that you can actually feel and you have information on it. Lately that model is ranking down a little bit, because you have information but it’s not tangible in any way and you can’t see where it is physically.

People have forgotten that they still have this collective group of information. So when they log on to Greplin and they start searching, and they start seeing, you know, two years worth of, lets say, Facebook messages. They realize, “I actually have this and it’s an actual hard drive that I own. Part of our job is to let you become aware again of how much information you have online.

How does Greplin work? Where’s all the data going and how safe is it?

It’s actually quite simple. You head over to Greplin and you make an account for yourself. Then you can pick which data sources you would like to have us search for you. You can keep on adding more and more everyday. For the most part you can click on a data source. Let’s take for example, LinkedIn. We’ll do this really neat thing where we’ll ask you to authorize us on your behalf to access your data from LinkedIn. I could get more technical, but that’s how we basically work.

For the bulk of the services that we offer for indexing today, we never actually capture your username and password, which I think is important to a lot of users. So you know, if one day, lets say, somehow, our accounts are compromised your password was never exposed to us. So it would still theoretically be safe.

Once you let us in, we’ll suck up some of the data from that site. We’ll try to build a really fast searchable index. The way to think about this is, imagine that you have a book, and right now, without Greplin, if you want to find a specific chapter in the book, you pretty much have to go through the entire book and scroll to the right page that you want. What we offer, essentially is an appendix on that book. And that’s what we get and build from your data on that site. So we’ll build this neat appendix of “this email is located here and this file is located there” and so forth. That’s a very distilled overview of how Greplin works.

Will you be building a mobile app?

Oh, yes, work begab for that quite awhile ago and it’s difficult, because you have a lot of constraints. We’re building first for the iPhone. The processor on the iPhone is not that powerful compared to a server. Full text search is an extremely processor intensive activity. So it proves to be a little difficult to do this to scale with a very large number of documents. Work has definitely begun and I think as we expand, mobile will become an increasingly important facet of how you access your information and Greplin will definitely be a part of that.

Being able to access data offline would be important with mobile. Is this a feature we can expect?

It’s very important to us that it’s accessible offline, because when you fire it up, you don’t want to have to wait another 2 minutes until AT&T can find an Internet connection. What if you’re underground or have no reception? For Greplin mobile to work at its best it would be best for it to work offline. However, as I said before, the mobile project is a hard one to build.

There are four of you on your team. Tell me about them.

It was myself and Robby for the most part and then we hired our first engineer, this guy named Kevin Clark. Kevin worked at a company called PowerSet that got acquired by Microsoft. The interesting thing about Powerset is that they didn’t really have a product when they were acquired. But Microsoft bought them for $100 million primarily because the engineering talent there was incredibly valued. One of our angel investors introduced us to Kevin Clark. We pretty much hit it off right away.

Our second guy is named Shaneal Manek. Shaneal also has an interesting story: We put up this programming challenge to Hacker News. It had this tone of “Are you a good enough engineer? Try to solve this.” We got a tremendous amount of responses. I think we got about 450 responses in total. Shaneal was clearly off the charts, the best. The only bummer was that he was in Chicago. We were like, we’ll do anything we need to do to get you. Sure enough, we flew him out here, and he started about a month ago and things have been going great since.

You announced 4 million in funding from Sequoia recently, but that was very soon after you raised your angel round. Why did you raise another round especially since your angel money was probably not dented, and why did you raise so much?

There were two things that we realized after we saw the incredible amount of inbound traffic. Firstly, the amount of people who want to use it put us at a point where we realized that to make this work there’s actual meaningful cost, both in servers and what we were doing. Google may not say this, but search is expensive to do.

Secondly, we needed money to be able to do this upfront and needed manpower. We’re a very small team at this point. Everyone is incredibly overworked. I know every startup has this story of everyone working to much, and could always use another hand. We [Robby and Gross] really felt that a lot of the very good engineers out there would be afraid of joining a startup that only had a few hundred thousand dollars. Some of these guys are more established and older and they’re looking for a company that would be around for awhile. Those are the two financial reasons to get funding.

There are deeper reasons that are not necessarily associated with money. The first one is brand. Sequoia is arguably the top venture capital fund in the world. So having them on board meant two things. It meant that our brand improved for users so I think we’re slightly more trustable now than we are as a company some angel decided to write a hundred thousand dollar check to.

The second reason is again focused at the engineers. Engineers were excited because they thought building Greplin was pretty neat. But now Greplin has essentially been validated as a real thing by people much more intelligent than me.

The third reason we took Sequoia’s money, specifically Sequoia and not another VC, was we deal with a lot of these different API providers. It’s very important to us to maintain very good relationships with a lot of these companies because we have a higher than average volume of requests that we have to make to them. And finally, we thought Sequoia was the best place to go for this because they’ve had their hand in or know someone at almost every interesting tech company in the world, and that network is super important to us. They funded Google and Apple to name just a couple.

Were there other venture firms involved at this stage, other than Sequoia, who wanted to fund Gremlin?

Funnily enough, it’s quite fascinating to see how it took me two months to raise half a million dollars and three weeks to raise 4 million dollars. I think what happened is that we generated some excitement. There were other equally exciting, equally interesting venture firms that we were talking to. Pretty much I would say, the top bunch [of VC firms]. We had some type of either verbal or paper term sheet from [the top VC’s].

I would add one more thing about Sequoia which is what totally sealed the deal for us. If you go to their website you totally understand the way they think about search. It’s pretty much the same way we think about it. People need the most simple intuitive experience to explore data. Sequoia flat out provides that better than anything else. They have a really strong view of this vision. Probably mostly because they invested in Google, and had one of the most record breaking returns in history. But mostly they have this idea, that people need a box that they can put information into and get results from. We had that vision too.



So I wanted your opinion on other search engines, like Blekko, who is generating quite a bit of excitement. What are your thoughts?

It’s weird. First of all you have to make the very clear distinction between web search and what we’re doing. There are technological differences. For example, we don’t crawl websites the way Google, or Blekko or Bing might. They do it based on links but if you look at your personal information online none of it really has links that leads to other different personal information. So the way we think about crawling and indexing is entirely different.

Will you be looking to move into that area at all as a lot of people are saying you will rival Google?

I think right now, we have this massive cookie to chew on that will probably keep us busy for the next few years. Down the road it will be interesting to see how we can help improve the quality of web search results because we have a very good profile on who you might be. But for now I think we have our hands full. Plus Google is doing a pretty good job at this point.

Everyone is saying that the future of search is social and while your approach is quite interesting, to mesh the two would be exceptionally interesting.

It would be very compelling, but it would also be very easy to get wrong. So we’re very sensitive to not hurting something that’s not broken. But that doesn’t mean that we won’t try at least.

Great to hear. Google should watch out. Back to the topic of you. You were born and raised in Israel, how does it differ to San Francisco?

My parents were born and raised here and moved to Israel shortly before I was born. So I’ve always had a very strong American upbringing both in culture and in language. It’s been really interesting to see the different style in both work ethic and culture here as opposed to anywhere else in the world. One of the things that’s pretty unique here is that there seems to be this value in the valley that all that really matters is what you can personally do.

This is something I find really refreshing. It’s one of those places in the world where, all the other chains that get locked onto you, whether it’s your age, your race, your religion, seem to disappear in the Valley. It literally boils down to you, the human, what can you do? What can you execute on? So that was pretty exciting.

Final question, what kind of advice would you give to young entrepreneurs?

This was my question to Mark Zuckerburg when he came in to YC. And I would love to give his answer except I don’t remember it, so I’ll just give you mine. I think my main advantage, that comes with my age, I would argue is true for people my age is: because you’re young, you have significantly less baggage than older people in life. You have less people to care for. You’re basically just caring for yourself.

So you have very, very little to lose by trying. So my advice would be just whatever you might be thinking about doing: just do it. You might be shooting down the idea in your head because the voice in the back of your mind might be saying it’s crazy. Ignore it and just do it. As long as it’s obviously legal and moral there’s very little to lose. I really like that attitude. I think that’s what Y Combinator really encourages, that incredibly scrappy attitude of you know, coming out here, living on a futon and building something because, what could go wrong? Probably not much.

One of the things that’s a bit of a bummer to me, as Greplin becomes more and more successful, inherently I have more and more to lose. So when I came out here, my strategy was, I’ve now basically won the equivalent to the golden ticket in Charlie and the Chocolate factory. Pretty much the worst that could happen would be for me to die or obviously get sick or injured. People I think, get frustrated, when things that aren’t really bad happen to them and they consider them very bad because they’ve gotten used to a certain level of stature. When an angel rejects you people get very dejected and I guess stop persuing what they’re doing at some point.

It’s very important to constantly tell yourself if anything, how lucky you are to be here in the first place. Even if you aren’t in the valley, remember how lucky you are to live in the 21st century.

Get the TNW newsletter

Get the most important tech news in your inbox each week.