In Mountain View, we meet CTO & Cofounder of BloomReach, Ashutosh Garg. He shares his story of how he cofounded this startup and how the current business model works, as well as what the current plans are for near future, and some advice for young entrepreneurs.

The transcription of the interview is uploaded below.


Martin: Hi. Today we are in Mountain View with BloomReach. Ashu, who are you and what do you do?

Ashutosh: I’m the CTO and co-founder of BloomReach.

Martin: When did you start this company and what was the reason for starting it?

Ashutosh: So we started BloomReach in late 2008 – early 2009. The vision of the company was to actually make the web a lot more relevant and personalized to the people. So like, over the last 10 to 15 years, the amount content of the web has grown by leaps and bounds. The amount of content that is available is huge.

And while Google and Facebook after all are trying to make it easy for people to navigate some of that web, what happens is once you leave these portals and land on an international website, whether it’s or eBay, or Amazon, or even New York Times, finding the content on these websites are huge challenge. The amount of content that exist in any of these websites is more than the entire web, what the entire web was in 1998. So navigating to that piece of content is really-really hard. What we want to do is to make it very easy for people to find and discover what they are looking for. So we started the company with that vision in mind.

Martin: So, Google would be the entry point, and then once people are on the websites automatically…

Ashutosh: It could be anything. Today if you look at any, to build a website, right. very small portion of the people, like 5 – 10% come directly, and the rest of the people will come either through a social network, through an email, through an affiliate campaign, through Google, through page search campaign, or any other channel.

So, people are going to land from where ever, whatever channel they picked. So it’s not that much about, like what channel they are picking, right. But they are coming from outside, they may land on any individual page of the website and not necessarily the homepage. So, what can you do, what can you do, all you know about that person, from where they’re coming, how they are coming, why they’re coming, and from their past behavior, and you use all that information to see what on your website is the most relevant piece of content for that user at that time and the name of that whole thing.

Martin: Understood. So what enabled yourself to start this company based on your professional background?

Ashutosh: So what does it take to make this thing happen?

  • One is, like, being at Google, like, they found out how people are navigating the web, what are the challenges that they are running into, what are the struggles they have, right. So that was one big area of pain thing for me.
  • The second is, working at a company like Google, I had access to lots of data, allowed to their systems, so became an expert in that system.
  • And third is, my PhD was in Machine Learning and Data Mining.

So, at some level all these things came together, where I could process a large amount of data, to make sense of that data and then present it to the people.

Martin: Okay. Great!


Martin: Let’s talk about the business model of BloomReach. As you said, the users come through different channels to the website and then what happens next?

Ashutosh: So if you take any website, right, what you want to do is you want to bring people in, you want to keep the people in over there and you want them to come again. So, bring the people in, that is marketing, right.

  • So one focus is, how can you make your marketing a lot more effective bringing more people in.
  • The second, so what have done is, we work at the website to understand how they organize their content. Is their content getting crawled easily or not, is this content in the same language as how people are looking for that content. So for example, the way people will describe a product in Europe is different from the way the will describe it in the US.

But, now if you’re reading the content for European audience but launching it in the US, that won’t work or vice versa, right. So you need to make sure that language is the same with how people are describing it. So we do a number of things to analyze content to make sure that it is really accessible by various, like either to social networks or search engines or anyone. So that is one focus on the marketing side.

And the second thing is, that once you are on the website, the way you navigate the website, you use the navigation search for the website, you go look at it, you go to search pages over there. So we provide a SNAP solution which is Search Navigation and Personalization.

And actually, our third product now which is on analytics. What it does is, it helps analyze all the content of a website and the user behavior and it will see what’s working and what’s not working and what you can do to improve it.

So the business model for different products is different. It is the combination of the usage and the performance, like how much customers are using it and what they’re getting out of it.

Martin: I totally understood that you have this kind of recommendation engine that you’re trying to sell and then the analytical part. Regarding the recommendation, is it that you have some kind of snippets that you just enter or you totally rearrange the whole website and categories, etc.?

Ashutosh: It can be both. So depending upon the authentification and what our customers want and will they see the maximum value, it could be a small widget on their website or could be an entire landing page or a combination of both.

Martin: And how did you find the first customer?

Ashutosh: The initial first few customers are always these early adapters, and either through family and friends. So working with our investors, their relationships, through our personal relationships we found some customers. One of the questions that their value proposition has to be strong enough that even though they understand that a startup is a risky value proposition, but by working with a startup their value is going to be so large if it works, then I’m willing to give it a try.

Martin: Can you tell us a little bit more on how the technology behind this works, especially in terms of the recommendation? So what type of data sets are you taking into account?

Ashutosh: Yes, absolutely. As I said, right. There’s a lot of data. So for the data, we look at 3 key sources of a data.

  • One is, any customers that you’re working with, we Google and crawl and mine all the data that exist on their website.
  • Second, that is crucial about the data is that broader web. Just to give you a context of why that matters is, if you look at any product as an example, you call a product a certain thing because everyone else calls it that. But any name is meaningless. It is the context of the rest of the web that name become meaningful. So by crawling the web, we’re able to understand and associate what each word means or each entity means. And then we’re able to relate the content of the website.
  • And the third thing that we look at is the users’ behavior. That ultimately if you’re trying to satisfy the users, you really need to see whether they like certain things or they’re not liking it.

And by bringing these 3 things together, the content of a website, understanding of that, understanding of the users’ behavior, understanding of the product, and then picking the right machine learning algorythms to learn on top of it and collaborative conditions, search and that allows the data processing.

Martin: Regarding the analytics product that you are selling, what type of actionable advice can you give your customers based on these insights?

Ashutosh: So if you take any website, I mean you guys have one for yourselves.

  • One thing you decide is what to write.
  • The second, once you’ve written it, have you written it correctly or not. Are you missing things or not, right.
  • Once you’ve done that, then how you are organizing the content on the website? Are you getting the right users’ behavior from that or not, right. Are there gaps in the content that you’ve written?

So, we actually give advice on all these things. What to write, how to write it, who to go after, how to lay out everything, so it really fits.

What is the purpose of a website or any of these things is connecting the right user to the right piece of content. It’s not about, shall I go with email or shall I go with SEO, shall I go with SEM, shall I do page search, shall I build the website, do I do that traffic, no. Our business, in your case, it is a set of articles, for a retailer a set of products, any of those things, right. First is you need to sort out your entry, so we can tell you what is performing, what is not performing, how you should go sorting it.

Second is, we see you’re writing an article. First I tell you, you should write on security, then within security what people are concerned about, how you should be describing that, then how should you be laying it out, to what audience you should be reaching out to, should you be reaching out to the people in a certain age group, certain demographics, or certain psychographics, what will work, what works? And once you’ve done all that, what work, what tends to, and help you to predict what’s going to happen tomorrow. So we do all of it.

Martin: Can you tell us how the revenue model works?

Ashutosh: For this product?

Martin: For analytics and for the recommendation.

Ashutosh: On analytics, it is a function of the amount of the content that people are searching and the amount of the users that we are processing. So, if a website has a large amount of content and a large amount of user base which are a lot more than, because the amount of processing that we have to do and analyze and storage and everything, is a function of that system.

Around the recommendation and personalization, on that it’s a function of the usage, like how much are users getting, are they putting a lot of, are they putting on a few pages. Based on that I mean, it costs us percent to work there. And there’s an element of how much value they are taking out of it as well.


Martin: Let’s talk about corporate strategies. What competitive advantage do you perceive of BloomReach over other companies?

Ashutosh: So, I think I will expand that slightly more and I say that, as a startup, what is a competitive advantage and how do you build that over them. And there’s number of things that you can do. One is when you’re starting on day one, it is clearly technology and product. Your product that no one else has, so that is your competitive advantage. All your piece of technology that no one else has.

So BloomReach, what we have done is a kind of tools with algorithms that are very-very unique. There are very-very few companies with a handful of people in the world who can do this kind of things. So that is one advantage that we have, that no one else has.

The second thing is that, over time the amount of data that we have collected is unique to us. So the customer that we are working with, we have collected their data for the last one year or last 2 years, right. So we know a lot about what’s going on. Our starting point is much better than anyone else. The network effect now that we are working with so many retailers as an example, our performance with the next retailer is so much better than it was with the first retailer, the first time. So big data advantage, network effect advantage.

And third is, we have a large customer base that gives us the responsibility so that becomes a huge advantage. So you just have a huge footprint and now you can upsell more profitable thing to your customer base.


Martin: In terms of market development, one part of your business is related to SEO and SEM. Can you tell us a little bit more, what implications that your business model have on the total market of SEO and SEM because there are so many SEO agencies out there consulting websites how they should optimize the website and you’re doing this somehow automized?

Ashutosh: So what people are doing is manually looking at small pieces of content and seeing what can be done over there. Now in that case, most of the time, because like how we can get around and get more traffic. As against as looking at like, why something needs to be done for SEO as an example. So if you have a website, but if you build it in a way that it can not be crawled, I mean that is a fundamental limitation of the way the content is consumed by different web search. So it is not really about SEO, it’s more around like the basic structure of the website. A simple way, say you have 5 pages, say you have 10 pages on your website. I can come with a link structure, and it will take 9 jumps to get to the last page. Or if you have a website of a million pages, I can build a link structure with just empty links on our page so that we can get from any page to any page within like 5 jumps.

Now, in a crawler based world, you need to make sure that the content is organized properly. Actually fundamental thing that has happened over the last 15 years is, if you go with any standard text book on marketing, it would be about how do you differentiate yourself, how do you stand out in the crowd. Make sense. But in a search based world, you need to me more part of the crowd than the stand out of the crowd. You can define everything that you’re doing in a way that is different from how everyone else is calling. But guess what, people are searching based on how everyone else is calling and not how you are calling it. So you want to show up over there. So you need to be as similar to the crowd as possible as against as different from the crowd.

So that’s the opposite how marketing was in the past. So we are trying to solve those basic problems, which require a lot more data analysis and everything. So it’s neither competing with the agencies, so I mean, they have their place and they are doing a lot and they are helping. So if you are like, Okay this is the 5 categories that I’m worried about, share it with the agencies, but if it’s more around like I want a website and make sure that is working well right over a million pages, no one can do it manually. That is where you need tools. And we have, that’s where BloomReach come into place.

Martin: Understood. So in terms of market development, you work at Google, what happens if Google enters a similar market like you do, because in addition to your data set, they have all the SEO or search engine data? Is this some kind of viable threat that you are feeling or not?

Ashutosh: I mean, Google has like 20 thousand engineers, sure they can do everything. And so, it is the case with any other company. But the customers that we involved are with, one is we know a lot more about them than anyone else would ever.

So, this is a simple thing about, you don’t know what you don’t know. If there’s a page that is not getting crawled, you just don’t know that, there’s nothing you can do about it. I mean, in our case, because we work very closely with these enterprises, we know about that and we can work surfacing it. So the amount of data and understanding that there is so much more the than anyone else.

But then, if someone wants to get in this space, sure they can and then. But I think the competitive advantage that we have built over them, now we know, for example my customer. We know what has been going on on their website for the last 1.5 years, no one else does. So we will have that big advantage. So the starting point for anyone else would be much lower than ours.


Martin: Imagine your little brother comes to you and says, Ashu I want to start a company. Maybe he lives in India. What would you recommend him? What should he do and what he shouldn’t do?

Ashutosh: Do it because you’re passionate about something, don’t do it for money or don’t do it for a quick outcome. I mean you are an entrepreneur yourself right? I’m sure you guys have heard of a startup is a roller coaster ride.

Martin: We also have written a post about that.

Ashutosh: A startup being a roller coaster ride? And actually a startup is anything but a roller coaster ride. It’s not a roller coaster ride. Working in a big company is a roller coaster ride. And let me explain what I mean by that.

When you are in a roller coaster, you might be going at 50 miles an hour, 70 miles an hour in that roller coaster for 5 minutes. You know that at the end of 5 minutes the roller caster will stop. You’ll be exactly where you sat, you have a harness around you, nothing is going to happen to you, you will get all the jerks, but you’ll be where you are. Nothing will change. The world will still be the same for you, right. Become me, you’ll be thrown around from one project to another, by the end of the month you will get a fat paycheck, that’s it. Your check won’t go by more than 10 – 20 % next year. That’s it.

A startup is like driving, you’re in the Bay area, like driving on a 101 and 150 miles an hour. Either you’ll be the first to reach your destination or you will die, nothing in between. There’s no harness, you don’t control anything else on the road. If you succeed, you are the fastest one, you’re just not there yet. That is what a startup is. So if you are ready for it.

The other thing is that most people, I mean, I meet so many entrepreneurs, they are like,

– So what are you trying to do?

– Here’s a good idea so I’m working on it.

– Why?

– Because I want to build WhatsApp for this thing, I want to build Uber for this thing.

– Are you passionate about it?

– No, no, but this will have a huge market and customers want this.

– What are you passionate about?

– No, but I can make money through this thing and the VCs are saying that this is a good idea.

– But have you give it a thought to what you want to do in life?

And I think people missed out on that aspect. So you should have a reason why you want to do it, you should have passion behind it. That’s the key thing.

Martin: When did you know that your passion is big data, machine learning and maybe search?

Ashutosh: I have been doing this for the last 15 – 18 years. My PhD was in that, my undergrad thesis was in that, my master was in that. So it has been just over time. But the key passion is, how you make life easier for people. That’s starting way to go. And all these are different ways and baby steps you can take.

Martin: What other advice would you give your little brother?

Ashutosh: I mean this is tough work. You have to be ready to do everything. Like we were discussing earlier right, it’s not only about like, I like this piece of work or I don’t like this piece of work.

  • You will have to do whatever is needed for you. You are the master in the office but you are also the janitor of the office. You just do everything. So that’s one thing.
  • It’s a long journey. You may have an exit in a year, you maybe successful right. You will die with that. And for the next 25 years right. So make sure that you’re ready to make that commitment.
  • Other thing is you need to make sure that you have enough support from your family, because in this race, for doing things and achieving our vision sometimes our family has to pay the price. And are they ready for it or not? Because once you are in, there’s no looking back. You will have good days, you will have bad days. Are you ready for that?

Martin: How did your family support you?

Ashutosh: I was on vacation actually last weekend, which was probably the first vacation I took in 2 years. So that is one part of how they have to suffer from this.

Martin: Not seeing you.

Ashutosh: Not seeing you and you’re not being there. But more importantly even when you are there, my wife kept asking me, so what are you thinking right now on the beach and not enjoying it. Do you mind if I quickly go and respond to that email or answer this thing. I mean, it’s a huge sacrifice. You are just not there for them.

Martin: Ashu, thank you very much for your time! And next time you are thinking about starting a company, pick some idea that you are very passionate about. Thank you very much. Great! Thank you very much, Ashu.

Comments are closed.