https://www.linkedin.com/in/janmaat/

This interview is about the future of sales and marketing research and intelligence. Bastiaan Janmaat, along with is team, is creating that future.

Bastiaan is the CEO and co-Founder of DataFox. DataFox is an AI platform that helps businesses bring more intelligence to their sales and marketing process. This means they often bring unstructured news and data into a structured format, pretty amazing.

They’re backed by Goldman Sachs, Google Ventures, Stanford and Slack.

Here are some other things we talk about:

-How did you start DataFox?
-How do you gather and structure the data?
-How would a client use Datafox?
-How do you get away from work?
-How do you deal with stress?

 
 

Transcript

Dave Kruse: Hey everyone. Welcome to another episode of Flyover Labs and today we get to talk to Bastiaan Janmaat. Bastiaan is the CEO and Co-Founder of DataFox and DataFox is an AI platform that helps businesses bring more intelligent to their sales and marketing processes and they are backed by Goldman Sachs, Google Ventures, Stanford and Slack; so not too bad of a list there. So I’m excited to learn more about DataFox and how they are exactly using machine learning and how their processes are better than the traditional sales and marketing research and qualification. So Bastiaan, thanks for coming on the show today.

Bastiaan Janmaat: Thanks for having me Dave. I‘m really excited to be on here today.

Dave Kruse: Definitely and so tell us about – before we get into DataFox and what you are doing now can you tell us how you got to be where you are now?

Bastiaan Janmaat: Sure, sure. Let’s see, I’ll give you the short background, but a better personal context I guess. I am originally from the Netherland, so I am – I’ve only been in the States for about five years. Before I moved here, my first job when I graduated from my Under Graduate Degree in the Netherland, my first job was in London. I worked there for an investment bank. I was 22 years old and I was basically – I was an investment analyst, which basically means you are researcher and part sales rep, in the sense that my job was to comb through a massive range of data to find companies that seemed to be taking off, that was the research part. And then I put my sales hat on and hit the phones, tried to reach out to the people wanting those interesting businesses, often times trying to reach CEOs, CFO or other kind of senior folks there to explain to them who we are and why we might be able to help them grow and it was all about building a relationship. And it was interesting job. I didn’t know anything about the world, so I learned a lot about interesting industries and businesses, but I was always kind of staggered by the fact that so much faith was put in me as a 22 year old to kind figured out where to find the right data, what actually motivated these enormous investment decisions. So I always found that kind of weird but I didn’t really have a big book and plans and I didn’t know how other businesses were. I did that for about four years and then I moved to the United States to get my MBA and I did my MBA at Stanford, which is an exciting place for a business guy like me to go do an MBA because it’s such a great technology school and learning ground, right. So whereas one of my classmates were spending a lot of time at the business school, I actually ventured over to the computer science department and was hanging out there all the time and this is where I started learning about some of the advancements, incredible advancements in Artificial Intelligent and other computer science and engineering methods. Basically what I saw was there are all these great methods to do what I had been doing manually, but to do that automatically and – there was one other classmate of mine who also I could tell was spending a lot of his time over at the Computer Science department. His name was Mike Dorsey and for me he had also been working in finance for a couple of years. So over many of the years and hackathons we kind of compared notes and realized, ‘Hey, this is future. We can –I think big companies, organizations, investment firms they are putting too much you know risk in their method just by relying on young people to find the right information. So we set out to build a computer driven method to basically do what we have been doing manually, not to replace those young analyst, but to make them far more effective. And so kind of full circle, four years later, now the investment bank I used to work for at Goldman Sachs is a customer and the 22 year old there now, they don’t have to sit in spreadsheets all day trying to manually figure out what the data is to look for. They use DataFox to find information on millions of business at the same time, so that they can wear more of their sales hat and actually go out and build relationships with these companies. So yeah, we’re kind of fun invention.

Dave Kruse: Nice, and so what type of data do you deliver to that 22 year old investment bank?

Bastiaan Janmaat: Yeah, the way to think about it is, we use algorithms to find all of the information that’s publically available about a business and we structure it all into a clean dataset. So a way to think about it is, if you or I were to go do research on one company, where would we go look. We’d start on their website. Probably we would read their About Us page, we’d look at their products, maybe there’s information on their blog. We would read the recent news articles about that business. We check LinkedIn and Twitter and other social feeds and you put that all together and we paint a pretty interesting picture of that business. But that’s kind of the work that’s required. Now DataFox does all of that but across millions of companies at the same time and automatically. So we use a technology called natural language processing, which basically just means it’s an algorithm that goes and looks for work patterns and key words the same way you or I would, right by reading a new article and there is one sentence that talks about the company’s recent funding round or there is an sentence about a company hiring a new CEO. There are certain key works that we are looking for, certain phrases that we are looking for and our algorithms do the same thing. And so the way that then plays out for our customer is, you know our market has really broaden significantly since those early days when we delivered this solution to the folks we used to work with and so now days big companies like UPS and Amazing and they their sales departments and investment banking and investment departments, they get access to this data set. So you might be looking for information on one business, but what usually happens is we actually surface the opportunities for you before you even ask for it. So we’ll know that Dave likes to taking to, I don’t know fast businesses that have raised at least a couple of million dollars, that are based in the United States, that have been growing quickly and then you’ll just get a feed of opportunities like that from DataFox.

Dave Kruse: Interesting, okay. So are you actually going out to you know like individual article, or articles and saying that hey, this company raised money and putting that out and then putting it into a structured data set.

Bastiaan Janmaat: That’s right, yeah, that’s right.

Dave Kruse: Wow! That’s not easy.

Bastiaan Janmaat: Yeah.

Dave Kruse: So…

Bastiaan Janmaat: No, I think you use so much content out there right, which is why you can’t do it manually.

Dave Kruse: So are you, and so do you. I mean, I image you must be like a number of areas doing that right, because you don’t exactly what these arterials are saying, but maybe if you tuned it enough over the years.

Bastiaan Janmaat: It’s a good question, not to get into too much technical detail, but one thing I’ve learned is from the forks I work with here, there is a tradeoff in machine learning between what’s called precision and what’s called a recall. Precession is when you just talk about footage; is the algorithm right. So is it correctly identifying that this sentence is about company actuating a funding round. On the other hand recall is the algorithm for finding everything that’s out there or is it missing certain sentences. So you can see why there is a tradeoff, right, because if you try to be exhausted and have high recall, but then you find everything then see you might sacrifice a better precession, you might actually find a couple of sentences that weren’t actually right. So that’s the constant tradeoff that our engineering team is ratcheting higher and higher and as we get more and more experiences we build out more and more algorithms, you end up getting high precessions and high recall. But you know one thing we do is we invest a lot in the quality assurance. So we constantly run tests across our dataset and certain samples or subsets we actually have people look at and scrutinize and the nice thing about it is it’s not a one-time effort. Like if you actually find in a mistake somewhere it’s actually a gift that keeps on giving, because the fact that that was a mistake becomes an ingredient in making the machine learning algorithm more efficient and more effective. So if you ever heard of the concept of supervised machine learning, that’s kind of how that plays in. You’ve got these tests for people supervising the algorithm and therefore making it better and better over time.

Dave Kruse: Got you, okay. And before we keep diving down there, can you maybe give us a little overview of DataFox; like the money raised and employees, kind of like your some examples of clients. You mentioned a couple I think at least now, but yeah anything you can mention will be interesting.

Bastiaan Janmaat: Yeah definitely. So the business is close to four years old and we are all based in San Francisco. It’s an exciting place to be just because there is so much talent, especially when you are trying to build an artificial intelligent company, you got great universities around here, but also great companies that serve us. Phenomenal training ground for folks trying to get in and to build their expertise in that area and also it’s nice to build the business here because there is so many great investors around here right. So we’ve raised closed to $10 million in total from Google Ventures, Great Advisor and Slack and Goldman Sachs. Slack is an interesting investor because we actually partnered with them first. Our data can be accessed in multiple ways and one of them is if you use Slack. The messenger used kind of the business messaging app, then you can get instant insights from us right there and your messaging tool and so Slack turned around and said look, our users are loving this. Can we invest in the company? And so they deepened their partnership with us that way. Yeah, I mentioned a few customers earlier. A way to think about it is that we sell into – as a startup you got to be focused right. I think some day anybody in a job that involves doing research on other companies are finding businesses to work with should use DataFox. That could include recruiters, and real-estate brokers and equity analyst and what have you, but for now we are stuck to areas that we just know really well which is two-fold. Number one, financial services, banks and investment firms they use DataFox to find opportunities and can track all their clients and then the second is, anyone working in sales and marketing and that’s where companies like UPS and Amazon use us to find customers. It’s fun because we are really turning what used to be a very manual process into a highly data driver one. So for example, for UPS, you see it’s a logistics company and so they have an anomalous sales and marketing department looking for opportunities to grow that business and it used to be a little, a little manual. If you were to create a sales rep or a marketer at UPS you’d have to hit the phone book just to find potential customers. Now look, they figured out using DataFox is that actually there are certain things about companies that make them right for becoming a potential customer and one of those things is when a company opens a new office, right. You can imagine when a company opens a new office one of the first things you are going to have to figure out is logistics by sending through mail and partnering with a logistics provider to actually interact with their customer base and so through DataFox we provide a feed of exactly when that happens.

Dave Kruse: Wow! Interesting!

Bastiaan Janmaat: It’s not just big companies we work with. There is a very similar use case. There is a company in New York called Managed by Q. They offer office cleaning services. It’s also a really cool company. If you are opening it in an office, the pains, kind of figuring out which cleaning service to work with and to manage that service, Managed by Q does all that for you. It’s a really cool service. The same thing, they want to know when companies are opening new offices, but they can’t kind of figure that out manually, so our algorithms are basically monitoring the Tri-State Area and any kind of business that opens an office, we let them know and it’s great for both sides. Can you imagine you are opening an office there and that exact day or week, you get a call that says hey, you might need some office cleaning services. You’re like ‘great yes, we got me at the right time.’

Dave Kruse: Yeah, it will make me think about the kind of work that we talked about before the interview the different clients. Just kind of think about what would be ideal and you might then, usually somewhat ideal like opening an office is impossible, but with you guys now it’s maybe possible. So that opens a world of possibilities; yeah interesting. Can somebody come to you and say hey we really want to figure out X. Would you guys put together a custom program or if they paid enough.

Bastiaan Janmaat: Yeah, that’s a great question. Every – I’ve learnt every business is different, right. Even though there is some similarity in how I just described how a business like UPS and a business like Managed by Q want to go to business. There is always some idiosyncrasies and every business is a little different. So this happens all the time. People come in and they say, ‘I’m especially interested in knowing when a company hires a VP of Sales’ or ‘I want to when a business files a new patent’ or ‘I want to know when a business is sponsoring a conference’ and this is golden feedback for us to get. This is exactly how we prioritize the 70, now we call them company signals that we were able to track across millions of companies. So we’ve got this running, huge running list of space because of events like that, that people have asked us to track and a little over time incorporate into the product.

Dave Kruse: Interesting, all right. And can you take us back to before you even started DataFox. I’m curious, you know definitely had a lot of experience you know to say, you know this is a good idea. But how did you – how did you kind of get it off the ground? Did you raise some money in order to build the initial product? Because it’s kind of a – I mean if you had just one signal, you mentioned 70 now, but at the beginning you probably didn’t have a lot. So how did you get enough so that a client would actually want to buy this?

Bastiaan Janmaat: Yeah, that’s a good question. You know we spent quite a bit of time. I think it was about a year that Mike and Ben and all of them; they are our engineering co-founders. The four of us spent probably about a year in a dark incubator somewhere working on a prototype. That’s the reality of building an artificial intelligent based technology right; it does take time. So what I described to you earlier about how we used training data to supervise their algorithms and make them smarter over time; you know it took us the better part of the year to actually get to a point where it was blowing off interesting insights. Its just – yeah, it’s difficult. You can’t built a data company, certainly not an AI company over night. But I think what gave us hope was two things: First of all Mike and I went on the business school. One of the best things about business school is that you are surrounded by hundreds of people who have done all kinds of interesting jobs before business school. So they were former investment analysts, and bankers and people who had run sales teams and people who had worked in marketing. So even though the product wasn’t up and running from day one, we were able to show them our prototypes and our thoughts and get the feedback on those prototypes and design. So that’s how we knew, if we could build this, it would take people a lot of time and drive huge benefits for people who needed search like this. The second thing was, we raised a nice seed round even before we launched, we had built those algorithms. So Google Ventures was one of our seeded investors and you know this, if you are building a data company there is no one better than Google to have on your side and one of the best things about that was they introduced us to some of their data experts. For example, some of the people behind Google Finance and their search algorithms and things like that. Wow! That helped us avoid some major pit falls that we wouldn’t have know about if we hadn’t had access to people like that.

Dave Kruse: Interesting. How much was that round, the initial seed round?

Bastiaan Janmaat: The first round was $2 million. We knew we needed – there is another bigger side of seed round we think, but we knew we needed some time.

Dave Kruse: That’s nice. Yeah, that is a hefty seed round, especially for the Midwest. Out there probably not unheard off, out here before having a product in the market, that’s hard to do.

Bastiaan Janmaat: Yeah, there are pros and cons. Its life’s expenses, you know anyway. We thought about opening offices elsewhere, we probably will at some point.

Dave Kruse: Nice. You should come to the Midwest, come to Madison.

Bastiaan Janmaat: Yeah, yeah.

Dave Kruse: No Madison, is a…

Bastiaan Janmaat: I’ve heard great things.

Dave Kruse: Yeah, it’s a good place but – All right, and so I was curious, where you kind of want to take the, I guess the AI part. I mean your vision for next three or five years is a lot around increasing those numbers of company’s signals from 70 to more. What else do you want to kind of build into your platform?

Bastiaan Janmaat: Yeah definitely. You know at times it feels like we are building – at times it feels like we are building three different products in one. Those three components are the data piece, the insights or algorithms piece and the works well piece. And one or two of those without the other would – really wouldn’t do it for our customers, they really go hand and hand. What it means is obviously you know first and foremost we have to have good reliable data. Both algorithms that go find information about millions of companies, if we are wrong, it all stops there, right. So we have to come up with very efficient novel ways to find out how many people work at each company, and how quickly are they growing, and do they have multiple offices and what are these milestones that the business is hitting, that’s all got to be right and we put a ton of time into that. Number two, it’s the insights or algorithms based on all those insights. You know so when is it really – when is an impart milestone hit. Like okay, we are tracking all this data on a business. Are they now suddenly hiring a lot more people or did they just open their first international office. It’s turning a humongous data base into the insights that are most important around the businesses. And then thirdly is workflow integration. You know we all – this isn’t about some people talking about millennials and how their workflow expectations are different. I don’t think it’s about millennials. It’s that we are all used to information being pushed our way. Not just information but the right insight at the right time and we are used to this as consumers and the example I always think of is, if you have a South West flight booked tomorrow, you expect to 24 hours in advance get a little notification or an email that says, ‘hey David, it’s time to check in for your flight’ and then you expect it to be a couple of quick passes on your phone, your checked in and now the ticket is saved on your phone and you just breeze through the airport and we all were increasingly expecting our business apps and our business tools to work the same way. So for DataFox as a data provider you know we have to be mindful of this and we see it all the time. Like people shouldn’t have to log in to DataFox.com to get a key insight. We should push it to them before they even ask for it. So if you’re a sales rep working at you know Managed by Q, that cleaning service company, you should get an alert in the morning that says ‘Hey David, of all the companies that you have been talking to recently, here are three that just opened a new office’ or ‘there’s a business that you met with a year ago that you should perhaps get back in touch with’ or ‘you know here are two companies you don’t even know about yet, but that look very similar to businesses you have worked with in the past’ that’s what people are beginning to expect and so that’s why I would say, your question was where do you see the storm for the next couple of years? We have to keep investing in all three, more and more reliable data about more and more companies. More and more of these interesting insights that are actually actionable and then these great workflow integrations, whether it’s Slack or an email or in your CRM like a sales force for example, we have to do the work for you.

Dave Kruse: Interesting, okay. And do you share a cost to do this for whether its – well, you don’t have to give an example of Managed by Q or how much they might be paying. Do you say that publicly at all?

Bastiaan Janmaat: Well, I’m not sure. Its anywhere from five to six figures a year, but there is so many dependencies and sometimes we have a scheme and what’s the volume of data and work that you’d be doing through DataFox, there are all kind of structures. And then yes, there is so many ways to interact with our data now. Are we using the integration? A lot of sophisticated companies now are actually using our API. So they just interact directly with our database to get what they need. So there are many, many different ways to structure an engagement with.

Dave Kruse: Interesting, okay. And what type of issues have you ran into around machine learning. I think what I’m trying to get at is that often people read about machine learning and AI and think click a button and it kind of magically all happens, but that’s probably not quite the truth. Just curious you know if you have any insights or if you have any lessons learned that you can share that aren’t too top secret I guess.

Bastiaan Janmaat: Sure, you mean incentives from our perspective building it or incentives in terms of from our customer’s perspective of using it?

Dave Kruse: More from building it.

Bastiaan Janmaat: Building it, yeah you know it’s funny because especially here in the valley we see the term AI that talks around pretty uniquely nowadays and I can’t blame people because it’s a big theme. Companies big and small are talking about it. You know IBM blogs and we see that everywhere nowadays. I think something I’ve learned is that I think generalized AI, which is you know kind of the science fiction image of a robot that hears and thinks like a human and kind of does anything that we could, I think that’s quite far off. But use case specific AI is already very advanced and so that’s where I think the most powerful AI solutions are those that are designed for a specific functionality, for a specific use case. I think a lot of it comes down to again the training data that’s required to build these algorithms that actually make the verification. You know if Google can train its image recognition algorithms to recognize when a picture is of a cat or of a dog, but its taken years and just millions of instances of training data to help get those algorithms to the point where they might recognize a cat in a picture. Now that same algorithm cannot be used to find UPS a new customers; you know what I mean. So I think that’s an odd lesson that in terms of how AI can actually help our customers and so what we actually need to build it, but is very concrete about how dysfunctional we can help our customers and build the algorithm that’s specifically pay loaded to getting that done. So within our stack we found different instances of machine learning based algorithms that do different things. We’ve got one that is able to extract these milestones from a piece of text; we’ve got to totally different one that tells you which businesses most closely compete with a company that you’re looking at. So it’s just different specific functionalities.

Dave Kruse: Interesting, okay. And you know when a company starts using it, do you have an idea of how much time you might save them? I mean this is a pretty loaded question, but you probably looked into a little bit of the kind of the ROI or I’m just curious if you have any insights on what you found?

Bastiaan Janmaat: Yeah, it depends on the use case. If you’re a sales rep and you do a lot of prospecting, so you’re constantly looking for new businesses to reach out to, then we save that person time in a couple of different ways actually. One of them is purely on the automation front. Your probably having to go find it – probably having once you find a business, you now have to pull all of the information about that business into your CRM. Your manager wants you do that, otherwise – and they are right to want you to do that, because otherwise you’re going to lose track of all these prospects. So that can take anywhere from five to 20 minutes for you to fill in – just fill in the information on one, go find it and fill it in, so that’s the first thing we save someone time on; just the automated flows of this information. What’s a little harder to quantify is finding the right opportunities in the first place and that’s where I think when people look at how DataFox helps them in that regard, its perhaps even more importantly than saving them time. It’s just finding them the right opportunity and what people have found with DataFox is that they end up finding businesses that are a better fit; therefore they are more likely to get a response from those prospects and these tend to be happier prospects over time because they are a benefit, and they tend to be customers from longer and pay more, so it’s just better for both sides. So depending on who we are talking to, we might look at it more from a time saving perspective or we might look at it more from a just driving revenues frankly perspective.

Dave Kruse: No, that makes sense. If you can drive revenue, people are always a little – people get happy pretty easily.

Bastiaan Janmaat: Yeah, that’s right.

Dave Kruse: So all right, so we’re almost done, but I’ve got a couple more on the personal side of things just if you’re game, just to get to know you a little bit. I’m always curious like what you do outside of work.

Bastiaan Janmaat: Yeah, yeah.

Dave Kruse: I mean I know you probably work a lot, but what do you do when you’re not working.

Bastiaan Janmaat: Yeah, yeah, yeah, sure. Well, during the week I try to – I’m a pretty – I’m definitely a morning person, so I get into the office really early, probably like 6:30 in the morning or something like that. But I try to as often as possible find an hour in the afternoon where I go for a run. I mean San Francisco is a beautiful city and it’s rarely – the weather conditions are rarely prohibitive. So our office is pretty close to what’s called the Embarcadero. It’s a road that runs along the bay. So I actually block that time on my calendar to make sure I kind of get out of the office and go from a run. But the week honestly Monday to Friday is pretty centered around work. The nice thing though is I get to interact with some of these people, right, our customers and the sales leaders and marketers and investors, so I guess work borders on fun pretty closely for me there. But then on the weekends you know it’s like I try to spend as much time outside as possible and enjoy sports, so it’s kind of that. I used to be a rugby player, but I’m too old for that now and that would be almost life threatening at this stage and a lot of times I hike. There’s beautiful hikes around here and I hop on the bike every once in a while.

Dave Kruse: Got you, interesting, all right. Any my next question is how do you find a little happiness each day. You might have already answered it with your run, well and talking to customers I guess and other stuff, but yeah, anything else that you do with in your life.

Bastiaan Janmaat: Yeah, just one other thing that comes to mind is my family is pretty spread out. My folks, they live in Singapore. My brother lives in Belgium and my sister, she used to live in the Netherlands, but she now lives in Singapore as well. So kind of – kind of I know the early morning or evening is the right time to call Singapore. So I kind of know kind of which time I can touch base with folks there and that gives me happiness when I get to touch base with my folks.

Dave Kruse: I like that one. All right, and lastly, you know obviously you enjoy what you do and you give awesome good energy here, but I’m sure you get stressed sometimes too, you know running a startup with or raising a lot of money, you have a lot of employees, you know when you get stressed out, you know how do you cope with it. You know do you talk to people, do you read something, do you just like punch something. I’m always curious how people you know…

Bastiaan Janmaat: Yeah, I guess there’s three things that come to mind to me. There are the going outside, going for a run, clear the hear, that always helps. Secondly, I find a lot of support in my fiancé. She is incredibly supportive and can really help me think through things and then thirdly, there are a lot of other people building businesses around here and even though a lot of the publicity is always you know so rosy and positive, companies growing quickly and you know everything just humpty dory, when I see the character of folks, everybody is figuring out similar challenges right and everybody goes through ups and downs. So I find a lot of support from this other Founders and the community here who you can kind of build a relationship with where you’re a sounding board to each other. I have a friend who is running a company here and every once in a while we pick up the phone and we’re like, how are you doing? Are you in a peak or a valley and that’s the reality. It’s a high data job you know. It’s like when you walk out of a meeting with a customer and they are happy and you close a deal, there is no better feeling. When you hire somebody who is great, there is no better feeling or you know conversely if the sites down for 15 minutes or you know your investor asks you a tough question and you didn’t have an answer for it right away, those are hairline challenging moments and so you kind of – its good to have a sounding board of other founders where you kind of be for each other, equalize each other out, you know what I mean.

Dave Kruse: Yeah, no that’s a really a good point. I think people often see the media and expect everything to be – I mean occasionally I read about things going bad, but yeah a lot of ups and downs, that’s good.

Bastiaan Janmaat: Yeah, yeah, yeah.

Dave Kruse: Well, I think that’s a great way to end this podcast. So Bastiaan, I definitely appreciate your time and your thoughts and it was great, and you speak English quite well; my goodness! I could never tell that you grew up in Netherlands, that’s pretty amazing.

Bastiaan Janmaat: Yeah, well thank you. I really appreciate the great questions and your time really.

Dave Kruse: Definitely and thanks everyone for listening to another episode of Flyover Labs. As always I greatly appreciate it and hope you enjoyed it as much as I did, because I think what DataFox is doing is pretty interesting and Man! You guys are putting together quite a – I was going to say data set, but quite a data intelligence company. So excited to see where you guys go.

Bastiaan Janmaat: Thank you very much. That’s awesome.

Dave Kruse: All right, bye everyone. Bye.