In San Mateo (CA), we meet the Co-founder and CEO of EverString, Vincent Yang. He shares his story how he came up with the idea and founded his company, how the current business model works, as well as Vincent provides some advice for young entrepreneurs.

The transcript of the interview is provided below.


Martin: Hi, today we are in San Mateo with EverString and Vincent. Vincent, who are you and what do you do?

Vincent: Hey, how are you doing?

Martin: Good!

Vincent: Welcome to the Silicon Valley.

Martin: Thank you very much. What are you doing at EverString?

Vincent: I’m one of the Co-Founder and CEO of EverString and what the whole company is doing is the predictive analytics, helping sales and marketing teams to predict who’s their next customers.

Martin: Great.

Martin: And what did you do before you started this company?

Vincent: When I started this company, I was mostly working in finance. So I started with my career in J.P. Morgan as an investment banker. So, over there my main job was analyzing public stock and then to decide which stock we should privatize. Later on, I moved to a private equity firm, called Summit Partners that was based here in Palo Alto. In my job again, it was to analyze companies, so I was analyzing around 25-30.000 companies in about two and a half years and this is kind of merely my experience before EverString.

Martin: And how did you come up with this idea of creating EverString?

Vincent: That’s a good question. So, I was originally mathematics major and when I was working in finance in J.P. Morgan I applied a lot of analytic skills on the analyzing public stocks. So this is all fine, and then, when I go to Summit, I realize for a human being it’s very difficult to analyze 10s of thousands of private companies and there is not too much information, it’s not public traded, there’s no revenue information. So, then I have the idea of is it possible to use national language processing to analyze companies for me, so that I do not need to go visit their website every day. And all I wanted it to do is for 10s of thousands of companies, and if you’re looking at the whole US, there are in total 5 million companies that have a website, I wanted to know which company is growing, which company is more interesting, which company fits the Summit Investment criteria. So, I started to do codings inside of the Summit and then do some trials and it works very well. And this really triggers me to think “Hey, there’s actually a new space to really apply the big data, the text mining, especially analyzing the unstructured data and to help business make better decisions”.

Martin: But still, there you analyze the companies which are growing etc., then, what is the leap towards marketing?

Vincent: That’s a good question. So, when EverString was first founded, we actually focused a lot on financial institutions, so helping venture capital, private equity, investing banks and even hedge fund. So we almost turn ourselves into an algorithm trading hedge fund. And it’s in about last year, I got a call from one CMO and then he was the one of the friend of friend and he heard about us and asking is it possible we apply a machine learning and natural language processing engine into their company and help him to analyze his leads, which is our company, too. Initially, that surprises us because we have never been to the space. And then we did a very smart trial, we were really starting to analyze. Because to us, analyzing company and analyzing leads it’s actually the same thing. So we did a trial and it actually worked very well. It actually converts about 300% more conversions than the benchmark. And after that, this really opened door for us, for all of us, which back then we were all bunch of mathematicians and neural network PhDs. So once we see the work and the variable for the marketing sales team, and we realize that the market is even bigger than the financial institutions, so we’re starting to now go all way into the space.

Martin: So now you’re only focusing on the marketing set, right?

Vincent: That’s right.

Martin: Ok, great. How do you manage your productivity of a daily day?

Vincent: My productivity actually varies a lot in different stages of the company. Early days, when we were at the seed round stage, there were not too many of us, in total we had about 10 people. So, I was basically doing, my main focus back then was focusing on algorithm design, focus on the backend architect and to focus more on getting the beta customers. But now, after Series A, our company is significantly larger, we hired 12 people in last month. And also now that we have 35 people in total and it’s much bigger team right now and we expect it to grow, like double in the size very soon. So in this case, my day was terrible. I now manage my day, I follow very strict discipline, so I get up close to 5:30am every day. I go to the gym every day just to make sure that I was well energized for the day. Then I’ll finish everything, go to the office in around 7-7:30am and before everybody arrives, so from 7:30am to 9am I try to focus a lot on product review, product design and making sure all the algorithms are in good track. And then, during the day from 9am o’clock to 5pm o’clock, some time is not actually my time. My time is divided by everybody and everybody put calendar invite to my calendar. So I was mostly stay in sales meetings, marketing meetings, product reviews, algorithm reviews, backend engineering, update calls, that kind of things. And from 5pm o’clock things start to cool down a little bit, I’m starting to get back my time again. And then grab dinner and then go back to do reply to email and finishing the day.

Martin: Ok, great. And how do you keep focus on all of the day? Because I can imagine if you have 10 open items that you need to work on. How do you focus at the specific item a certain time?

Vincent: That’s a good question. I did a couple of things. One is: every month, I will write down what my top priorities are. I’ll list 5 priorities to do for the whole month. And then, once I start to break it into the weekly priorities and then, this is all I have to achieve and I assure it with the top executive team in EverString. So, they all know what I will be focusing my time on. And then every day I will write down today there are three things that I must focus on and everything else is not as important as those things. Because I find out a lot of the CEOs that their mode was, what I call it, interruptive tribal. So, they are sitting there, solving issues, putting out fires everywhere. This is very tough for an young company, so we have to be really focused, so that’s why I make sure I only focus on the right things. I do some meditations, and I think meditation helps a lot to, that people focus on the fundamental issues. And that fundamental issues, I ask myself every day, I ask most of the team members here, too, why we are doing EverString and why we are here. Most of the teams we hire are really top stars in all kind of industries. And then, there is no point joining a startup, so this is a question I ask myself here, why we are here doing EverString, so that really helps me to focus on what I’m supposed to do.

Martin: And how do you manage your team? Do you also know that they have some kind of 4-5 top priorities over month, are they communicated between the executive team?

Vincent: Because I’m a big fan of planning and I started doing my own planning for a total, like 15 years now, so since I go to high school. I plan my day every day. This is the culture that I set up in the company and I was wishing that everyone, in our weekly meetings, everybody will list down: here are the top 3 things that I’ll be focusing on. And 60% of their time will be focusing on top priorities. That moved the company to the next level. 40% of the time are putting out fires, solving issues, there are a lot of unexpected things coming up. Not everybody was used to that in particular, engineer, scientist, they are not born in a world where you have so many things. But gradually, I think, after working with me for some time I think they are starting to getting used to those. We use a lot of tools internally, as well, the collaboration tools, the goal setting tools that get everybody into the same page.

Martin: Ok. Can you recommend a goal setting tool for example?

Vincent: Yeah, so we use Asana. We’re hoping everybody, there’s a section everybody are putting their weekly goals, and then we will review their goals. So that’s a very helpful. And Trello is another helpful tool. Trello is much easier, you can put your week into different days and everybody could review what people are supposed to do during the days. And we are building product, we use Hackpad. It’s an online writing pad where people could collaborate. This is also very helpful for your team to do things and making sure people are on the same page.

Martin: Ok, great.


Martin: Vincent, let’s talk briefly about your business model. How does it currently work?

Vincent: Well, our business model is actually, if you think about it in high level, it’s actually very easy. We take a look at our clients’ customers and we find a look-like, and that’s it. The biggest pain points that I find when I was talking with a lot of CMOs, they were like “Today I know we have a lot of revenues, our revenue will grow”. But if you ask them “Who are your ideal customers?”, this is the question where most of the CMOs don’t have a clear answer. And when I ask them “Who are the leads that you don’t want to talk to?” they are like “I don’t have a clear answer”. So, our fundamental goal in EverString is really to use data science to help you understand who are your ideal customers and bring you more of them. So, this is kind of the general high level concept. In terms of the way how it works, it’s we plug in our engine inside of their CRM, marketing automation and all the internal lead capturing tools they had been using. We first analyze who are your ideal customers and who are the bad leads that you don’t want to talk to. But, all the information we have is just a company name or url. This is the same thing when I was back in the finance part of the world is, back then I would have a whole list of 30.000 company names, and I have to do all their analysis to come up with among them 30.000 who shall I talk to. So, the same thing. Once we have that, and then we started to crawl the web, and it would focus a lot on analyzing unstructured context. So we have the technology built internally to turn the whole webpage into mathematical signals. A webpage is unstructured, a mathematical signal is structured. So, there’s a lot of work internally that we’re doing there. That’s why, in the space that we are able to generate our features, that are around 50 times more than all our competitors. So, our predictive accuracy is much higher as well. This is generally how we had been doing in building the model. Once the model is built, and then we apply the model to do a lot of things. So we apply the model to tell to sales team, every day, when all of the leads come to the website, what are the leads you should focus on. And also try to grab more leads into the funnel, whether it’s upper funnel or the lower funnel. So this is generally what we are doing.

Martin: Do you only tell “Ok, you should focus on these leads” or do you already tell some kind of story or some kind of pitch that you could use? If you see potential lead, went to a conference, and assuming that the company did not have this information before, than you could either pitch “Ok, he was on the conference, maybe you can use this as our pitch for making the contact”.

Vincent: Yes, this is something that’s important, that we have built a prototype, but we haven’t put into the product & environment, this is in the product roadmap. Not only helping our clients to know who to talk to, but also help them to engage with them.

Martin: I understand.

Martin: Let’s talk briefly about the corporate strategy. Can you tell us a little bit more about general overview of the product strategy going forward?

Vincent: The industry we are in is a very early stage industry. In general, people call it predictive analytics, but it’s just very much like a buzz word. It’s like people let’s call it big data, people call it Hadoop in the early day. People call social network, it’s just buzz word. But, in reality what we’re doing is just use data science to help specifically the marketing team, that the main generation people, the inside sales team, the lead qualification team to do better on their jobs. Right now there are a few players like us that are in the space that are even early stage. And then what most of the people are doing is building the model. This is where we are right now. So, in the future, what you can imagine is, once the model is built, there are a lot of places where you could apply the model, to help them to do better lead scoring, to help generating better leads, and there are a few more things that it could help them further. So, for us, it’s no different. We are constantly exploring new applications that could be built based on a model we apply to the clients. So, right now we have built a few very interesting applications, so we are working with our clients on those, so we haven’t announced that yet, but very soon we’ll be letting the world know that the superpower of the model and once it applied in the right way, whole business efficiency and your revenue will have a big boost.


Martin: Can you tell us a little bit about industry structure or the market structure in general and maybe some trends you have identified?

Vincent: Yeah, sure. The way how I think about the industry is in three parts. It is data, insights, and action. If you look at three parts. So, in the general, what I call the marketing or the sales intelligent space or the sales enablement space, there are a lot of data providers. They focus on generating proprietary data and then feed it to the right customer or to the modelers like us. Then you have the middle part which is the insights, like us, like some other players that, what we focus on is building predictive model and integrating the third party data, the first party data to do a good job on telling what our audience will look like. And then there’s the action part where you have the CRM, you have the marketing automation and you have the DSP-SSP part of the world that helps you to really engage with the client. So this is what I see the ecosystem of the market, but definitely the trend that I would see those three parts will come closer and closer to really even become one place. This is a very interesting trend that I’ve seen happening a little bit in the market, it’s too early in the games, but definitely in 5-10 years down the line you will see some players that are able to handle all three things at the same time and provide end to end solutions.

Martin: Ok. And what would be your estimated forecast, do you rather think that the data companies will integrate all of the other steps or do you really think that they would just organically try to build something similar?

Vincent: Yeah, I think all the three parts of the industry has the potential and ambition to take over the whole industry.This is just generally true for most of the industries, so on the lower stream people wanted to go up stream, the data vendors where I see they’ve starting to build models, they’re starting to even go more and more into the action part. And similarly for the people, the marketing automation, the CRM provider, they try to go a little bit backwards as well. Everybody might have a chance to win, but to me, I feel like the eventual winner, it really depends on how good you understand the real work flow of your clients and how strong team you have. It’s easy to say we do predictive analytics, but it’s actually very hard to do it. A lot of the people that I know just basically uses Excel sheet to do some simple logistic regression, where you can do it and I can do it, it’s very easy. But for us, in day-one that we hire our PhDs in Stanford that focus on neural network, and they are analyzing about 10s of billions of data points they build on models. So, these are the people that it’s very hard to get and it’s very hard to make them work as a team. So, we will see who will eventually win the game.

Martin: Let’s take a step back and talk briefly about machine learning. How does machine learning work basically for somebody who has never done it before?

Vincent: You mean, for first time entrepreneur interested in space?

Martin: Yes, right, for example, if I’m an entrepreneur and I’m interested in doing some kind of machine learning startup. What should I know, how the basic machine learning world works?

Vincent: I think the basic concept of machine learning is very, very simple. It has two parts. You understand the past and you try to use this to predict the future. So, the past and the future. In the technical terms, the way how we understand the past is by building a model. We call that training set. So, we need some of the data that tells me a little bit about your past so I have to build a model of. Second thing is what we call the evaluation. It’s very important to know how good the performance was for this model. We need sort of evaluation data set to test how well the model could perform. I think for a first time entrepreneur, engineers or anyone interested to come to the space, definitely try to go watch the Andrew Ng’s lecture about machine learning. Andrew Ng is famous professor in Stanford. He does the whole Coursera videos on machine learning, I think this is very helpful for people that doesn’t know the space, and the other thing is more to understand where you want to apply the machine learning. I’ve seen, when I was back in finance, so when I was an investor myself, I’ve seen numerous companies that claimed they do machine learning. It’s more like a typical failure example, you have a technology and you look for a solution. Most of the cases that doesn’t work. So, I’ve seen companies apply machine learning to industries that would never make any money, or to industries that are not growing, so this is also, I think it’s one of the lessons that I’ve learned in a hard way, as well, to really first think about what is the solution you want it to solve rather than think “I’m just going to run a machine learning company”. I feel that might be a little bit tough.

Martin: Sure. Ok, great.


Martin: Let’s talk briefly about your learnings and advice that you can give to first time entrepreneurs. What have you learned over the last years?

Vincent: I’ve learned so many, many things in a past. How do I start, so the couple of things.

  • One is I think in hiring, I made a few mistakes in hiring. And it’s very, very important for first time entrepreneurs to have a really strong principle and discipline in hiring. Don’t hire your best buddies in your high school, don’t hire your uncles and cousins just because that you know each other for 15 years. Or don’t hire people because they’re just good in interview and never do reference check. So, really have a discipline in hiring. If you hire the wrong people, this could literally kill the company. Product strategy, if that wasn’t right, that’s fine, you could eventually pivot, but hiring could kill a company. So hiring I think is the fundamental part.
  • And second is really to dream big. Obviously, a lot of the people say “Oh, just think big!” but it’s actually very hard to do. When I’m interviewing with so many people, now we are hiring a lot of executive, and you ask them to think big, you ask them how the industry will look like. Well, it turns out most of the people do not have the capability to think big. You do not even practice. So constantly you should ask yourself the question “Right now, there’s Microsoft, there’s Oracle, but what are the new products that could come up that could totally disrupt Microsoft? What are the new products that could disrupt the total CRM industry? What are the new products that could disrupt whatever the traditional DSP technologies?” So, these are the questions you really need to ask yourself every day as a brain exercise. Thinking big is very, very important. In the past, in the early days, we actually did not think big. We tried to see, we just narrowly focused we’re all into this thing so eventually we’ve become a big ship in a small pond. We can never go anywhere. So, we would rather be, let’s find the big ocean. Even if you start as a small ship, because the ocean is so big, there’s a tons of ways that you could gradually become really, really big, gigantic ship.

Martin: It is very interesting what you just said, because in Europe, a lot of investors say “Maybe you should start in the niche market where you could create the big market share and maybe once in a time later on you can move to other markets”.

Vincent: Obviously, you need to start small. I understand, that’s definitely true. But, in my opinion, you really want to make sure that industry is big enough that could have so many different players in the space and there’s so much room to grow and then there are so many modules and features you could really bring to the industry. And also, industry is old enough that you could disrupt. If you really start with very, very small, without even looking at the industry, you might end up taking a 100% of the market share of the industry that’s in total 5 million dollars.

Martin: Vincent, you said that hiring is key, I totally understand. What are focusing on when somebody sits in front of you and you want to test whether he’s the right guy or girl to join your company?

Vincent: It depends on the level that I’m interviewing, whether it’s a junior level engineer, a senior level data scientist, or executive level product officer. I’ll focus on the mid level and above, because junior level is more skill set testing, not too much on the other things. Middle and above, I focus on the mistakes that they have made. I would usually ask how many mistakes that you have made and what have you learned from your mistakes? So, if people say “No, I’ve not made any mistakes”, I would rather see people making lots of mistakes and learn from that, even in EverString, we make mistakes, too. But this is a new space. If you don’t stretch yourself too much, you will never be able to grow. This is the thing that I think I benefit the most from is always, every day to go outside of my comfort zone. Every day to try something new, because that’s the evolvement of yourself, as I think is the true purpose of human beings, and I find that I learned a lot achievement by doing that. So, coming back to the hiring question, the question I always ask is “How many mistakes you have made?” The second question I usually ask is “If I do a reference call to your supervisor and to your peers, what would they say about you? What are the critiques that you think they will say about you?” This is really to test whether they have good self awareness of them. If you join Oracle, Microsoft, that’s fine, you go there and you work, but if you join an early stage company, the things that we’re doing nobody have done before. For sure, we will make mistakes, but we want someone that have a very strong self awareness, that every day could reflect and know what did I do right, what did I do wrong, and with this the whole team could grow very far away.

Martin: Are there any further learnings you would like to share with our audience?

Vincent: I think that the other learning that I have is really just to have the courage to start, if you talk about the general starting up a company. And most of the people that, if you dare to think, and most of the people have very high safe net. Similar to my former background working finance, working Wall Street, you have relatively high income, and as long as you don’t make mistakes, you have a very safe track to be a managing director in a private equity firm or something. You have a high income but if, at the end of the day, when you die, you ask yourself what an influence did you bring to the world and you might think “Maybe not too much”. So, this is the one that, most of the companies that I see or my friends, their companies, a lot of them fail in the seed round. And the main reason why they fail is not because the product doesn’t fit, it’s that founders give up. The reason why the founder give up is they’re constantly evaluating “If I go to work in McKinsey here’s how much I’m going to make, and in company I make nothing”, so every day they’re evaluating their opportunity costs and their salaries in the start. I think that’s not the right way to think about it. Burn the bridge, never even think about the opportunity costs and keep going. I think that’s kind of one of the strongest lesson that I learned. If you could really be committed, think about the question when you die what are the things you really leave the worlds. What are the things that really pushed the mankind and all the organizations forward, even a little tiny step? And if you’re thinking this way, you will find running a company is the single best carrier in a life. Don’t do consulting, don’t do banking, that’s the waste of your life.

Martin: Great. Thank you very much, Vincent. And now you should think about what should you do in your life. Maybe start a company.

Vincent: Thanks, Martin.

Comments are closed.