Big Data & Brews: Jake Flomenberg of Accel Partners Shares His Insight into Startup Success
In the second half of our conversation (you can catch the first half here), Jake shares his thoughts around what makes a successful startup in the Big Data market.
Stefan: Welcome to Big Data & Brews. We continue with Jake from Accel, and a wonderful strong double IPA beer with 8% alcohol by volume. Cheers Jake! Thanks for joining.
Jake: Cheers! Thanks for having me.
Stefan: So we just joked a little bit about Big Data jumping the shark. Well, first of all, I am German, I don’t even know what jump the shark means. What is that? There was a TV show or something?
Jake: No, not so much with a TV show. I think the expression just refers to, Big Data is everywhere. So every company that walks through our doors is a Big Data company. They can be doing network function virtualization, there would be-
Stefan: It’s Big Data.
Jake: Big data. Accounts payable? Big data. Everything is Big Data. You know what, I’ve been trying to rationalize that. How do I make sense of what I’m hearing the market try to tell me? Are they just trying to latch on to a fancy marketing term? Probably, but the way I frame it is that all software needs to be data driven. If it’s not data driven it’s just not good software. And so, I think we’re going to see data driven software just become the way that we think about software going forward. And if you’re not-
Stefan: So just leave the Big Data out and just tell me what you do.
Jake: Yeah, tell me how you actually leverage a data asset to provide value to an end user, not that you store lots of ones and zeros.
Stefan: Yeah. So, I mean there is so much noise in the market, and you have to sift through that all, right? I guess you see a lot of presentations?
Stefan: I’m sure there is a lot of fun, but a lot of not so fun. What’s the problem with all that noise? Is it a problem in the media department that TechCrunch and Gigaom are hyping this too much up, or how can we have the dust all settle and come back to do a good job building great technology? [3:02]
Jake: I don’t know that it’s necessarily a problem. We typically go through these cycles of rapid innovation in a particular area followed by a little bit more of a maturation and perhaps even consolidation. I’d say it’s great having all these different companies come in because most of the folks who come in are experts in their field, and have spent a lifetime developing a particular knowledge base, far more than me about it, so I’m excited to meet them and learn from them. I think really where it become a problem is for the companies that are really on to something interesting to get up above the noise, because some of the spaces become so crowded that they wind up blending together or bleeding in to just a noisier crowd.
And so they might have a very good technology and they might have a very good product, and now they need to all of a sudden go spend all these marketing resources just to poke their head out and get up above the noise. So that’s where I’ve seen the most frustration. And it’s particularly hard for seed stage companies that don’t have the type of marketing budget that a larger company can command.
Stefan: So that brings us back a little bit more to the entrepreneurial conversation we had, and I thought was very fascinating. So what’s, maybe, if you think about companies before they get funded by guys like you, what are the three most important, maybe steps or processes or parts, steps in their lifetime before they come to you guys? There is a guy in a garage saying, “I have an idea,” but what should they do, the first three things, before they even show up?
Jake: Before you show up?
Stefan: Yeah. And then we’ll talk about after. After you wrote a twenty million dollar check.
Jake: Yeah. So, I mean it depends what stage they’re coming in, but I’d say again, the most important things that they should think about are what is the team that they’re assembling, and what is the problem that they’re going after? Validating that the market is meaningful and large, and that they can go and tackle that market. At the beginning, before any code is written, really that’s all there is. It’s a team and a market, and that’s the only basis of evaluation we often have for the very young companies. And then, you should just figure out the right way to reach out. I think asking for, figuring out how to get into that first meeting can be tricky at times. If you know anyone that is connected that can endorse you, that’s obviously than sometimes coming in cold, but we’ll take meetings with almost anyone. [5:35]
So if you have that team and market, then it’s just, “What are the steps you can take to de-risk?” Is building a prototype? Is going and getting that first piece of revenue? Is getting usage more important? And it sort of depends on what kind of business you’re intending to build. Are you building a social network? Are you building an enterprise software business? Right? And so, there’s no magic bullet.
Stefan: And then, you’re excited, you guys do your process, you guys invest. Boom, I’m an entrepreneur, now I have twenty million dollars. Should I go to Brazil, and run away? What should I do with twenty million dollars as a new entrepreneur? I mean, you have that experience in multiple companies. What are the next steps? Is it investing it all into engineering and build a product, or when should I start hiring a VP of Marketing, VP of Sales?
Jake: Yeah. And again, sometimes there’s not generic answers. I would say, if there’s anything I’ve learned from my brief time in this business, it’s that the best leading indicator of future success is the quality of the team that you’re assembling. So I think assembling, in the case of a lot of these Big Data, whether it’s the management platforms or the ops on top, that core engineering team is really critical. And hiring almost in advance of need, because as soon as you release a product into the wild, people are going to use it, and then future requirements are going to flood in. And then you’re going to be inundated. So building up engineering wisely is the first step. Then waiting for that validation, that product market fit. [7:03]
The CEO and founder often has to wear a whole bunch of hats out of the gate. They have to be, sometimes they’re the tech lead, they’re the sales department, they’re the BD department. And going and proving that something is working, and collecting that first piece of revenue. And once you start collecting a couple checks, you get a feeling that something is working. And then you can figure out, how do you want to step on the gas? Do you need sales and sales engineering? Do you need marketing? What are the ways that we can accelerate and scale this business, and make it productive use of capital? You run the risk of building a sales team too early, when there isn’t that product market fit, and you can waste millions of dollars that way. By the same token, a lot of the technical teams often have never hired a salesperson before, and that process can take a period of time. [7:45]
Stefan: Yeah, how is that salesperson even look like, how is he talking? What’s a good salesperson and a bad salesperson? I certainly made that experience very painful.
Jake: I guess the guidance I’d give, you know I could give you a lot of generic advice in terms of how to go find one who’s young, hungry, up and coming. You probably don’t want to grab someone with gray hair who has a cushy seven figure salary, somewhere else. But at the end of the day I think it’s a great opportunity to rely on your network. If you come from a technical background, you’re going to know what a qualified engineer looks like, you’re going to be able to make that decision quickly. And the most, the best thing that I can recommend, is build a, whether it’s a board of directors, or a board of advisors, or just friends. People, you know find people within your network who knows what a great salesperson looks like, and ask them to get involved and help you in that process, so that you can find someone and act and move quickly on them.
Stefan: So the wisdom of the crowd, basically pulling-
Jake: I don’t know about crowd. The wisdom of a portion of the crowd.
Stefan: A trusted group of the crowd.
Stefan: Maybe switching gears a little bit. What’s one of the technologies in that whole stack that excites you, where you think there will be another rapid explosion of innovation? I mean, there’s a lot of talk about SSDs, and there’s a lot of talk about in-memory and fiber channel, what have you. What’s an area where you’re really excited, where you really think, especially the top layer, will be very much driven forward by that?
Jake: So I think there’s a whole bunch of analytics companies that are going to make people’s life a lot easier. Whether it’s a more modern BI solution, search solutions. Ways to roll out particular industry verticals, whether it’s retail analytics, marketing analytics, etc. One of the broader trends that I’m seeing and still trying to make sense of is just machine learning at scale. And the opportunity to democratize some of those insights to the end users. So a lot of people play in these two buckets and can build some algorithm better and badder than everyone else, and I think that’s really exciting, and it’s going to create a lot of value for a lot of large enterprises. But the real prize in my mind is the ability to democratize that, and to share that with someone who doesn’t have to know how to program in R or Python, and can now get some value back from that system. [10:12]
Stefan: And will that be vertical focus, or is that a generic thing?
Jake: I’m not sure, time will tell. I think out of the gate it should be verticalized. Because it’s such a hard, gnarly problem, that unless you contextualize it a little bit, to frame it in language that people are used to seeing, it’s just going to be really hard to execute on. And that doesn’t mean that a horizontal company can’t be the one that does it, but to showcase the value of the solution they’re building, they need to move up the stack a little bit and say, “Hey, here’s how you do this better than you’ve ever been able to do before.”
Stefan: So, completely different question, but as you sit in all those meetings every day, and again I find that fascinating that you see so many people, where’s your BS meter going off? And like, “Eh, that will never happen?”
Jake: There’s probably a couple different answers to that question. One is just, market is too small. You can find a really passionate entrepreneur just going after a problem that you don’t think the rest of the world suffers from to the same extent. And that’s unfortunate, because it’s usually a very well intentioned entrepreneur who cares a lot about the space they’re going after. Another one is just failure to talk to customers. People that talk a little bit too much about the technology, and not enough about the value that they provide for their end customers. And then finally, it’s just not having the data. A lot of times, if you probe, you’ll get sort of, the deflection answer of, “I’m not going to tell you how much revenue, how many customers, where they exactly are in the customer contract cycle.” It makes it really hard to do business, and I think the best thing that a lot of these entrepreneurs can do is just be open. [11:55]
Stefan: Everybody will find out anyhow, right?
Jake: We’re not going to tell them, but-
Stefan: Do you guys sign NDAs with … ?
Jake: We don’t sign NDAs, but I mean, our entire reputation is at stake. We wouldn’t be in business if we didn’t respect the companies that we speak with.
Stefan: I mean, I’ve never signed an NDA with any single investor I’ve worked with in the last, almost ten years.
Jake: Yeah. I think the reputation matters, at least, I think more than the piece of paper.
Stefan: So what’s your process, then, look like, when you get someone in on a high level? You know, this is a great team, there is a big market, technology is interesting. And then I guess you double click on all the assumptions that the entrepreneurs and companies provided to you. But what are you guys doing? You pull in the guy from Google and have him do code review? Most likely not.
Jake: Sometimes. Again, it’s about risk identification. So at the end of every meeting, we’ll circle up. We’ll say, “What do we think about this business potential?” If it’s interesting enough. “What are the key risks that we see in the business or assumptions that we need to validate?” Or, “How can we validate those assumptions or reduce those risks?” And sometimes that can be done within the partnership and, you know, we do that, and other times it’s bringing in industry experts, or talking to customers, or doing tech due diligence, or whatever it is to address the particular risk of those businesses. And we try to move quickly, address those risks, and then come to an assessment or determination in terms of whether or not we want to move forward. [13:26]
Stefan: Tell me the top three books entrepreneurs should read before they meet you.
Jake: Top three books. I think a lot of the work that Steve Blank has done over the years, you know, “Four Steps to the Epiphany” and the more modern iterations are, is very useful in terms of how to go conduct those marketing experiments and product market fit really quickly. Other books …
Jake: All those books are good, but I mean, nothing replaces-
Stefan: There is no blueprint.
Jake: There’s no blueprint. Go spend time with your customers. Go figure out how to hustle and get something done, maybe read the cliff notes, I don’t know.
Stefan: Good. So where is the Big Data market going in the next five years? Where do you think?
Jake: You know, so as I said before, I think all this, whether it’s the storage or the data management layer, I think there’s more and more opportunities, but the Big Data movement is moving towards the top of the stack. That last mile of Big Data. How do we give people who are now used to these consumerized apps, where they have the power of things like Facebook Graph Search and Google at their fingertips, and give them similar experiences through the form of data driven software.
We have this rare asset, these data scientists, who are now inundated with tasks, where they have to go munch all the data and do the analytics. How do we take them, and move them to the higher order problems by democratizing a lot of the tasks that they do, so that people can just pull down an app for whatever it is. Whether a manufacturing, or ERP, or IT ops, whatever it is, and just use that tool, and now their life is more productive or more efficient. I think that’s just the value that a lot of these platforms can create at the end of the day. [15:20]
Stefan: Do you think the value created at the top of the stack will be bigger than the value created on the bottom of the stack, or that a life cycle thing, as we put on more and more layers, the lower tier is getting more commoditized?
Jake: I think some pieces of the tier may get more commoditized, but I think having things on top of the stack, quite frankly, makes other pieces in the stack more valuable. They can’t be truly valuable in the long run without an ecosystem. And that’s why we think of, just the platform opportunity for Big Data, right? We really need places where people can go to leverage, whether it’s Hadoop or non-relational data storage, to build things on top. I think there’s room for, again, a handful, a half a dozen vendors to become huge multi-billion-dollar companies. At the end of the day data-driven software has room for hundreds of multi-million-dollar, multi-billion-dollar companies. [16:11]
Stefan: What’s the Big Data apps, and those kind of stakes, that you know you touch every single day? I guess you use maybe Twitter or Facebook, right, so they obviously have a Big Data stake, but is there anything on your day-to-day business life, on a daily basis, you know, “This is here, where their technology is actually making a difference in my life?”
Jake: Yeah, I think every day I touch a little bit of Hadoop, a little bit of Couchbase, just through the consumer services that I use. I use marketing analytics tools like Origami Logic to look at how, or trends in social media, or things of that nature. I used to use a lot more of the data tooling in BI tools, unfortunately I don’t get to use them quite as much anymore. But quite frankly, some of the tools that I use are probably the least interesting market, because venture capital is this small little corner of the universe, and I’d rather much think about what the average business analyst or average marketer or average IT ops person is going to use on a day to day basis. [17:20]
Stefan: Does the world need CDOs? Chief Data Officers?
Jake: I think it’s a great question. I think at the end of the day it actually depends on what it is that they’re going to do, and whether or not they’re going to create value for the business. There’s a lot of three letter terms like Chief Innovation Officer where people go prance around and go to meetings and speaking engagements and don’t actually drive value within the business. I think for those companies that are equipped to actually drive more data driven practices into their business, a Chief Data Officer could make all the difference in the world. There’s great companies that, credit card companies have Chief Data Officers that help do, drive new risk models and all these things that are really exiting. So I think at the end of the day, as long as they’re good at turning insight, the insights they have, into action, and pushing it throughout the rest of the business, it’s very exciting. So, yes but only the good ones.
Stefan: So I saw with a lot of investors, big companies coming to the investor and say, “We want to build a Big Data stake, show me all your portfolio companies.” Those kind of things. Do you see this working out very well, where the big banks having a VP of Big Data and then they, touring Silicon Valley, right, and see all the technology, and put it in their labs, and juggling around the next twenty four months. Do you see those projects being successful in the long run, or is there the different approach being more successful?
Jake: They fail much more often than then succeed. And I think they suffer from a number of problems. The first problem is just changed normal business practice. Large businesses are all about a reputable business model and business practice and now you’re trying to introduce change. How can you do that more effectively? The easiest way to know that a project is going to fail, is if someone, a CIO, comes in and says, “I need a Big Data strategy.” That’s a recipe for disaster, it’s a recipe to start investing in the bottom half of this stack, do a lot things, and not have any idea how you’re going to get value out through the end of the day.
The right way to solve the problem is to identify a key area where you think you can extract value, from leveraging the problem, and find a purpose built, or build your own, solution to address that problem. And more often than not, while you’re in the process of building that solution, you’re going to start inspiring people within your companies to start asking divergent questions about other fields, and see how you can leverage that technology in a new way. But without that first ROI story, that first use case, it never works out. [19:59]
Stefan: So I guess, see the success story and then grow around that. Well, Jake, thank you very much for the really good beer. We should continue in the dive bar down the street. Always good catching up. Thanks for joining us for Big Data & Brews. See you soon.