In San Francisco (CA), we meet Co-Founder & CEO of Quantcast, Konrad Feldman. Konrad talks about his story how he came up with the idea and founded Quantcast, how the current business model works, as well as he provides some advice for young entrepreneurs.


Martin: Hi! Today, we are in San Francisco in the Quantcast office. Hi Konrad.

Konrad: Hi.

Martin: Who are you and what do you do?

Konrad: So, I’m Konrad Feldman. I’m co-founder and CEO of Quantcast.

Martin: What is Quantcast?

Konrad: So, Quantcast is a data and technology company that helps the companies that produce all the content that we consume on the web, get a better understanding of those audiences so they can better serve them and we also help the advertisers who effectively pay for most of the free content that we as consumers get access to, to make that advertising experience more relevant for consumers.

Martin: How did you come up with this business idea and what did you do before that?

Konrad: So, my background – actually my career started in research in Machine Learning in London. While I was there studying for a PhD in Machine Learning, which I actually didn’t finish, because I started a company with colleagues of mine called Searchspace and we built systems that use machine learning, looking at transactions mostly in financial services, predominantly for the banks to help those banks spot patterns of transactions that they couldn’t find – that the people couldn’t find themselves because there were so many transactions. These systems were used mostly to help major banks to take money laundering and terrorist financing which sounds a long way away from helping understand digital audiences and making advertising more relevant. But fundamentally it comes down to how we can get computer systems to understand the massive amounts of data that are created by people in sort of – in everyday practice and it’s about the means for processing that data, finding those patterns and putting those patterns systematically to use. So, that what was the link there.

In terms of specifically for Quantcast in the space that we’re in, actually, it didn’t start. It didn’t start specifically with the, sort of, the business focus. It started from a fascination of the internet. So, my last company once we worked predominantly with banks. Back in 2000, we got to do some work and fault protection for Alta Vista and it was the first time I’ve been exposed to Internet Data. I find that I’ve always understood the world and businesses through data, because that’s what I’ve always done. My way of understanding different businesses I worked with is by looking at the data they generate and here was the internet which was data that didn’t just relate to a business, it related to the world, it related to commerce and culture and habits and behaviours – it’s really fascinating data set.

And so in 2005, my last company Searchspace was acquired and I was living in New York at the time at Financial Services Center, and I decided I wanted to do something internet-related, didn’t know what, and so, but I knew that to do it I should you come to the Bay Area, come to San Francisco because that’s where everything internet was happening. So  I moved out here and I started looking at a number of different areas involving sort of advanced analytics from Internet data. And one of the areas that interested me was advertising. Initially, just because there’s a lot of talk about advertising in Silicon Valley – this was 2005 (end of 2005). Google just gone public and lot of people here just assumed that advertising was search and search was advertising.

I knew there was more to advertising the just search and what fascinated me was, sort of, a curiosity was that: it was really clear why search was so effective. With search, a computer system was able to use data to make the experience much more relevant and useful for the consumer and so my question was: “Why is that process not happening outside of search?” If I went to a particular website or my friend went to website or my mother went to website, everyone was seeing the same ad. That just seemed strange and it literally started from there. I wanted to know “why was that?” And I didn’t know anything about advertising myself and my co-founder neither, who really knew much about advertising. So we actually put a classified ad on Craigslist asking to speak to people who could teach us about advertising. And the ad was something like: “Do you know about advertising? Can you spare an hour? We’ll give you $20 and buy you a coffee. Teach us about advertising”.

And as we met with people in the advertising industry here in San Francisco, we started to realize that the problem related to data. So, there are lots and lots of searches every day. So a computer system has to make the decisions about what results will be relevant for the consumer and need the data to make those decisions and so the data and the algorithm uses that data. Well with search, the data that the computer system needs is presented by the consumer. Whenever we do a search, we say: “Hey, here’s the data to use to create relevance”. The types of data that were necessary outside of search advertising, the types of things that advertisers use to say: “This is the type of person that I believe will be interested in my product”. It’s much less to do with intent or context keywords, there’s much more to do with who the people are: their demographics, their interests their aspirations, their habits, their lifestyle. And none of that information is presented by the consumer when they make a request to the web server. And so the sort of, the idea that eventually became Quantcast was: “If we could somehow manufacturer consistently that sort of data, we will be able to make the experiences more relevant for consumers.”

And that was really the hypothesis, and we tested it out. In 2006, we found a way that we could start to generate some of that, sort of, that data or that signal and when we saw the initial results, we decided we put more efforts behind it and Quantcast launched in September 2006.

Martin: Cool. Once you’ve tested your hypothesis, what was the next step then? So, did you choose only a subset of potential customers and then work only for them and try to optimize with them? So what, some people would say these as some better customers? Or did you just try this out for only one customer? Or did you try to acquire tons of new customers?

Konrad: So, actually the approach we took, we saw that the opportunity was that the world was moving away from a broadcast model. So in a broadcast model, the content and the advertising – and good advertising should be content — is tied directly to the media. You open the page of a magazine or you look – you know everyone – the signal goes out in the airways and everyone tunes in. And the opportunity with digital media was to create experiences for individuals and so we felt that the long-term value is to actually help make those individual decisions, but we realized that to me able to make those individual decisions, to be able to truly enable real-time advertising you had to have real-time measurement and historically, media measurement was not real-time at all, it was actually based on samples that were done, sometimes continuously, sometimes periodically, and will be updated monthly and those sampling approaches, which were necessary in broadcast media. You know, if you put a signal in the airwaves, you don’t know who is tuning in, you have to use a sample, but they didn’t work very well as the number of media choices grew.

When there were three TV stations, they worked fine. When there were 300, they didn’t work at all. And here was the internet with millions of different channels, millions of different destinations. And so we saw an opportunity in realizing that we would need real-time measurement. We saw an opportunity to essentially disrupt the way the audience measurement had worked and move from a model that was based upon, sort of, out-dated sampling, the one that was based on direct measurement of all the media consumption. And so that was the idea how did we get to the first set of customers.

So, for us, it was I think it was really quite, sort of, classic, disruptive innovation where we produced a solution that initially was really useful to a subset of the market instead of a market that couldn’t find a solution. They couldn’t find a solution, there wasn’t a solution that worked for them, maybe their site was too small to show up in one of the samples or the site was big enough that they couldn’t afford the budget that has cost them to buy that research service.

So we had two things that worked different:

  • One was, we made the system based on direct instrumentation, we measured directly every media consumption event and we built the computer infrastructure that processed all of that data.
  • The second one was, we made the audience measurement free.

So, initially there was a set of customers where this new service was accessible to them. It could work and it was affordable, it was free and they started using it. As they used it, they talked about it and more people discovered it.

But another thing happened as well, which was: as we started to gain visibility into not just individual websites but behaviours across websites that’s where our machine learning approach can start to learn more about consumer behaviours and understand more characteristics and we can then introduce new ways of understanding audiences, maybe based on their demographics or the other interests that they have. And as we started introducing these new features, the service became attractive to a new set of publishers. As more publishers started to participate, a sort of, a network effect from data – more data coming in means we can produce better results for everyone. So, everyone got back out more of what they put in, so then more people started to participate.

This sounds like “Wow, just you know, just happens like that” – it wasn’t like that. It’s always a lot of growth hacking in terms of discovering what works well in helping people discover the service and use the service and come back, and recommend it even. But over a period of a couple of years, we built up very, very significant adoption for Quantcast Measure. Today it’s used by millions of web destinations. That could be sites, it could be apps but also many other platforms that are used to empower the web, use Quantcast Measure.

Martin: Great. So, you have at least two products, one is Measure and the other one is Advertising.

Konrad: Yes.

Martin: So, it sounds to me that you started with Measure because maybe it was easier and the basic building block for Advertising. And then you started with acquiring the first kind of standalone websites trying to bundle them and say “Okay, we know at least we can understand some patterns based on all the people that are interacting with these 50 websites. Do you want to have optimized advertising based on behaviour analysis?”

Konrad: So, it’s similar to that. When we started off, we didn’t know precisely – we decided that Quantcast Measure should be free. And in the future, we would have a commercial product by helping not just giving someone an aggregate report saying: “Hey, hope this is useful”, we would make money when we help make our customers businesses better and we get aligned with their sort of ultimate goals. When we launched Quantcast Measure in 2006, we did not know exactly how that would work, we just knew that, basically we realize that if we’re able to measure, if we had enough data we could measure not just whole sites but individual impressions. And we knew that that would give us data that could be used to influence outcomes. We did not know precisely how to do it.

So, in 2009 we launched Quantcast Advertise and our timing was good, it was just fortuitous because I think, we launched in June of 2009 and in September, I believe it was, the DoubleClick Ad Exchange by Google launched which was a new way of being able to buy and sell digital advertising where rather than doing individual deals with the whole site or a set of sites, you could actually evaluate in real-time, opportunities to place advertising in an auction-based market. And we looked at that early on and we thought it was really interesting. We thought it was a really nice way of aligning the interest in the ecosystem, it would work for publishers, it would work for advertisers, and importantly it would work for consumers because if you had bidding model to where someone is going to pay more, they have to be creating more value, which means they have to be creating more relevance, so we like the alignment.

And the other piece that was I think fortuitous for us was that: the technology necessary to integrate, you have to have interface to these marketplaces very, very quickly and the technology necessary to do that was quite similar to the technology we already built worldwide to power Quantcast Measure. So we decided to invest heavily behind the growth to what was called RTB or real-time bidding, which is programmatic advertisement, making advertising not just much more efficient but more effective. And that turned out to be a good bet, that market has grown dramatically since then and programmatic will be the, will over time, be the bulk of all advertising.

Martin: Konrad, how hard was it to acquire the first like, five to ten websites? And just tell us how was it like pitching in front of those media publishers and saying: “Hey, I’m a young start-up and I’ve got this awesome measurement tool, please use it.”?

Konrad: So, that’s a really great question. So, we started off with a good understanding of what we thought would be useful. So lots of our friends here in the Bay Area had consumer internet start-ups. They all were frustrated about the available measurement services and so we had an understanding, there was a need. People would like a better service if one existed. We thought, we so awaited to delivering it. So we were able to sort of speak with a few people that we knew in the industry and some of the people we met through research and have them look at what we’re doing and test it out, and then we just launched it.

We went and launched it and we did, you know, what promotion we could around it and we worked on things like SEO and making sure there was good visibility into the services that we can provide, but it wasn’t an overnight thing. You know, maybe one day, a site would sign up. But because it was all online and we could look at the traffic that was coming to our site, and we could understand whether they were signing up for a service or not, whether they were dropping off, we could constantly iterate and optimize it and it went from one day there’ll be a site sign up, the next day, no sites, there would be a service working, there’ll be another site, a few sites. And then you know, again this wasn’t overnight but within weeks maybe, there were ten sites a day, within months, hundreds, and within a year, thousands of sites a day that were signing up.

Martin: And you only acquired them via SEO or email marketing or did you also have some acquisition?

Konrad: No, we also did limited acquisition. Early on, we felt that we still have a lot to learn, lots to develop, to have a service that would really resonate with the biggest media companies and so we gradually worked our way up. Now, it turns out that some of the sites that joined us early on when, you know, when they were small sites, happened to be really huge sites today, you know like Buzzfeed and things like that. But over time, it started attracting attention from the big media companies and that was more, sort of, classic business development because these are really big companies with a lots of established processes, and we spent a lot of time, we realized it would be important to our credibility to have many of the, sort of, established big media companies using our service as well. And we spent quite a bit of time in the first ones, they took time. They took many, many months, some of them to get on board, but that wouldn’t be where I would start. When you starting up, you’re trying to discover: do I have a service that is valuable to someone? Iteration is really important. You’re social experimenting and trying to discover the right mix of features and functionality that is attractive and useful to a customer set and if it takes months and months to run an experiment that really limits the rate at which you can learn, so it was a sequencing thing for us for sure.

Martin: How long did it take to develop up the first iteration of your product with which you didn’t feel super embarrassed to go to the smaller websites?

Konrad: Actually, the very first iteration where we didn’t have a product to show people, where we had sort of concepts based on data, probably took us a couple of months, a pretty good couple of months. The first version of the product that we could really put out, that took about probably six months.

That’s one of the interesting things – that we are building the sort of service, we knew initially there’ll be a very small number of customers but that could grow at a rate that might be difficult to predict and it was very, very important that any site that signed up for our service – two things:

  • First of all, that by measure – by basically by measuring the tag that’s based on their website, we couldn’t slow down their experience. So when big sites signed up, it had to work seamlessly every time, we wouldn’t get second chances.
  • And also if we started collecting all the data, we have to process that data. So we actually spent quite a bit of time building out the infrastructure – both the real-time systems and the, sort of, the back-end cluster computing, sort of the big data platforms based on MapReduce. We spent quite a bit of time getting ready for that, sort of, a plan to success, that wasn’t an easy way back then. It’s sort of before Amazon EC2 and you know those sorts of things so we had to do quite a bit of work to build out that infrastructure anticipation of getting traction and fortunately we got that.


Martin: Konrad, alright, so let’s talk about the business model of Quantcast. Besides publishers, are there any other customer subsegments that you were targeting?

Konrad: Yes, we actually work with advertisers and advertising agencies. We worked with thousands of the world’s leading brands to make their advertising more relevant for consumers and therefore perform better for them. So, our revenue comes predominantly from advertisers and advertising agencies in relation to their advertising campaigns.

Martin: And what is the value proposition that you are providing to them? Okay, obviously due to programmatic advertising more relevance and therefore potentially higher CPMs or so?

Konrad: It’s actually a very simple value proposition. It’s that, we will make any advertisers’ budget work much harder for them so that will enable them to more effectively reach the audiences that will respond and that translates to measurable results. So we can deliver much higher return of advertising spend and much more scale against their goals because advertising isn’t just about: “Can I reduce the cost of acquiring new customers”, it’s actually about “How much scale can I get at a given targeted cost of customer acquisition” and we can do that better than anyone else.

Martin: When you look at your publisher customers for example, are you serving them differently? For example, having a key account management for the larger ones and maybe just customer support for the smaller ones or..?

Konrad: So, the answer is yes. But in fact the vast majority of customers use all of our self-serve interfaces. We were very fortunate to have huge number of participating publishers and they range from people with their individual blog through to the biggest multinational media companies. And absolutely, we have different levels of service depending upon what products they’re using and what their needs are but we’re available for any of our publishers. They can get in touch, they can call us, they come through email, they can use Twitter, they can use our website. But actually for most publishers usage, we’ve got great online documentation and the product is actually very, very easy to use and get set up.

Martin: Okay, Konrad. So when I look at Quantcast, I see okay, you have lots of publishers. You know a lot about the users and therefore you can have some kind of real-time bidding kind of retail advertising platform. How does this compare to Google model? Because when they have something like Google Analytics, they also know lots of what the people are doing and what they’re searching for and also they have this kind of advertising platform.

Konrad: Well, Google Analytics is a little bit different from Quantcast Measure and that its predominately focused on helping people understand how their site is used. So we use Google Analytics ourselves on our sites to find out if we change a landing page, or we change a navigation flow, does it improve – does it make it easier for our customers to find the things they’re looking for? Do we drive more of the outcomes that we’re looking for? Quantcast Measure focuses very much on understanding who the audience is, so it’s designed to help with decisions about sort of editorial and content curation but also for research and sales associated with advertising so then the difference is there.

Then in terms of Google’s advertising business, absolutely Google is a big player in digital advertising and always will be. This is a very large market and it’s a market that is growing very quickly.

We’re only 6 years now into, sort of, real-time bidding which is what truly enables real-time advertising. And that’s quite early and people forget that web search was already five years old when Google launched so there’s actually a lot of innovation. And in a space where you need to have, we need to make a lot of investment, have rapid innovation, certainly having sort of scale and expertise that is similar to what Google has, that’s a great plus, but also the speed at which you can iterate. Back to my earlier point, your ability to run experiments quickly and I think that’s easier for a company like Quantcast with a real focus on this particular space.

Martin: And how are you managing the business, because I assume you have something like a core business where you get your revenue and you can pay your bills and then as you said, you do want to do some experiments in order to disrupt the market. How are you doing this organizationally?

Konrad: So, I think there’s two pieces to that actually.

One is, you know, fundamentally we believe that constant experimentation is key to the success of our business, not just in terms of new products but actually in terms of enhancements to our existing products. We make our products better all the time. And actually one of the largest investments we’ve made from a technology perspective is designing our entire system to support constant experimentation. We are always experimenting on new approaches that will make advertising perform better and doesn’t matter how good the idea is until you really try it out, you don’t really know how it works. So you want to make the cost of running experiments as low as possible and run as many as you can and recognize that most of them won’t work and that’s okay because you can actually learn more from the ones that don’t work than at the ones that do. As long as you have a way of making those experiments relatively low cost and you can sort of communicate the findings easily as the organization, you have a mechanism for experimenting and learning all the time.

So that’s fundamental to our core business – we always experimenting and enhancing our core products to make advertising more relevant, make our reports more insightful, and so on. And that sort of factored into our business, how we do planning. Well, the first thing we do each year is, say we got to invest to make everything we’ve already got better.

But there’s also the opportunities to explore the development of new products. We just launched a new product called Audience Grid, which is a way of leveraging the infrastructure and modelling and data that we develop with Quantcast Measure and Quantcast Advertise to everyone else’s data. So we launched this program with – working with data companies like Oracle’s Datalogic and Experion and research companies like IRI and Kantar Shopcom, a whole bunch of different organizations where they can leverage our data modelling as well. So you’ve always got to make room for new products, new product development and again sort of, experimentation around that.

That’s sort of part of our budgeting processes, to make sure that we can do that but I think also we try to encourage it within the work that we’re doing so you know. One classic way of doing that is Hackathons, we do those regularly. We also have a program here we do annually called the Quantathon, which is conceptually similar to Hackathon, but Hackathons tend to be limited to people who are writing code and the Quantathon is designed to be inclusive for everyone in the organization.

So we actually encourage cross-functional teams but the general idea is “Can you produce something that improves the company?” Maybe it improves our products, maybe it’s a new product, and maybe it’s a new way of doing business, maybe it’s a new fun feature or function for employees here. And over the years, we’ve had some tremendous contributions.

In fact we launched a product – last year’s Quantathon winner was a product, it manifested itself in a product we launched about two months ago and it’s a suite of solutions called Search powered audiences which enables advertisers that already invested a lot in understanding search advertising and they’ve developed and curated sets of keywords that they understand, represent their core audiences. We provide away for them to use that taxonomy to find new customers before they search. So we can look at the types of people got, we’ve got a set of people that have, we see lots of search activity, we can look at the set of people that have searched a particular product and rather than saying “Hey, let’s show ads to those people that have already searched”, we say “What is it about those people? What are the patterns of people that ultimately do that, that could predict that other people are likely to be interested in that as well?” And we’ve never had a product launch that’s moved so quickly, I think we are, say, two months in. Now, over 300 global brands are using this product.

Martin: Konrad, let’s go back to the advertisers. So, imagine I’m an advertiser, I want to place my ads and you were trying to optimize it and I’m maybe I’m totally newby and I’m only give a content type text ads. Would you give your recommendations and say “If you would provide a picture or optimize your text like that then we would predict, you will get, I don’t know, lower cost per leads or cost per link or something like that”?

Konrad: Generally, that’s the case. One of the things that’s very attractive about search based advertising is that text ads are easier to create than graphical video ads, you know, things like that. We are generally focused on the larger advertisers; I think it’s certainly the case that in search, if you look at the number of advertisers that Google works with, which is in the millions now. There’s a very, very long tail. And once there’s a fairly long tail outside of search, the bulk of the – there is about six hundred billion dollars a year spent on advertising – the bulk of that, the vast majority of it spent by the top few thousand advertisers. So we’re generally working with large advertisers that already have these assets and work with creative agencies and so on, to produce them.

There are a bunch of interesting companies around, start-ups around them that are helping those that perhaps don’t have to creative assets, start to create adverts other than other than text ads and I think that’s an interesting area for the future, is working with companies like that, partner the companies like that.

Martin: And what is your perspective on the future of this real-time bidding in the advertising space? Is it more like, okay, there will be like 3, 4, 5 major players who are part, who cover the major understanding on what the biggest dataset of what the people are doing and therefore can provide this kind of island to advertisers to penetrate? And the second question, regarding the perspective would be: What is the role of agencies? Because in the old work, we had this kind of, I don’t know, the big brand, the agency and the publisher. When you look at the Google Advertising or so, the agency is diminishing.

Konrad: Let’s start with your first question. So, it’s an interesting space, so the market that we live in which is sort of, often referred to as advertising technology has probably thousands of companies in it, so which is a function in part of how large the opportunity is. I think that the sort of the venture investment model is probably fairly efficient in aggregate – there’s a multi-billion dollar opportunity so billions of dollars gone after it but of course it’s not evenly distributed. And as in most fields, most of those companies won’t make it in the long term.

I think that in terms of the dynamics of what will make the companies that are successful, successful, well, one of them will be technology, which you think: “Well of course, it’s technology, it’s advertising technology!” But actually many of the companies in that area, in the space and so of in the ad text space fundamentally aren’t really technology companies. They haven’t invested that much on technology and maybe, historically that wasn’t necessary but it really is today. To be able to make, sort of differentiated outcomes on a continual basis, you got to do it with, sort of, the necessary operating efficiencies, they have to deliver profitability long-term, doing this is and so on, it takes real technology.

And the other piece that it requires is data and not small data, but big, massive, massive data. And so to give you some perspective on the scale of data that we operate with – our daily processing can exceed 40,000 terabytes, so 40 petabytes a day. So the data that’s involved are really significant so I do think that technology and data processing and proprietary data are going to be absolutely key to if the company’s successful. It won’t just be one or two but it won’t be 10 or 20 either. It’ll be a limited number of companies that really have true scale, billions of dollars in revenues scale.

Martin: Konrad, you said before: the first company that you started was related to machine learning, the second company was into machinery learning. What type of industries, or business models do you think are currently super ripe for disruption based on machine learning but that might still not be tackled by much?

Konrad: So..

Martin: Except advertising.

Konrad: So, what industries won’t be transformed by big data and machine learning? Every industry. Every aspect of society will be completely transformed by big data which I would sort of, include machine learning in that over the next call it, two decades. It’ll be a bigger impact over the next 20 years than the internet was in the last twenty years.

Now, just so happens that the epicentre of this change is online media and advertising and one of the reasons for that is where the data is. That data exists here, is this a data rich environment but it’s also a world in which we’re able to measure the results quickly. We have to make an investment in technology and data processing and algorithms, and we can see that we get a return from that investment. And when we get a return, we can justify investing more. So, you have these cycles. And the benefits from that are felt much more widely than in digital advertising.

A lot of the fundamental technology that powers big data in all sorts of organizations comes out of online advertising. Google created the processing paradigm called Map Reduce, Yahoo’s open source version of Hadoop is sort of the standard platform the people use for big data processing. We initially started with Hadoop very early on and we gradually outgrew aspects of it and built our own stack. But one piece of that, we got our file system and we open sourced that file system of years ago: the Quantcast file system for dealing with very, very large datasets and that’s now used by many organizations outside of digital media and advertising. But the same sort of, transition and the rise of the use of big data that we have, which is central to success in programmatic, we’re going to see that in many other industries with tremendous benefit and opportunity.

Martin: Do you think that the companies in whatever industry it might be will be winning: Who has the most proprietary data or is it more like really having the best algorithms?

Konrad: I think it’s actually it’s a combination of both and it’ll probably vary a little bit by industry. But having lots, having large datasets really helps. So, in trying to solve the problem, a simple algorithm on a really large data set will normally vary from sophisticated algorithm on a small dataset, okay? But if you can have a really large dataset, preferably, one that is proprietary and you got really sophisticated algorithms then you get the best of both worlds.


Martin: Konrad, you started two companies. What have been the major learnings over the years that you can share with other people interested in becoming an entrepreneur?

Konrad: That’s a great question, I’m trying to sort of, distil it. It goes with a saying that you should enjoy what you’re doing because you’re going to spend a lot of time doing – it won’t be joyful every day. But it’s, I think it’s not just that you enjoy what you doing, you got pursue something significant, something that’s meaningful in terms of the problems you’re going to solve, something that can be commercially very successful because no one builds anything on their own. And you need to be able, you want to be able to attract the very best people in the organization which means having a real mission and ambition in terms of what you’re pursuing.

So, I think finding that combination of things that you are good at doing, the things that you enjoy doing and the things and the things that can be commercially very successful is really important. Taking time to think about what those might be and having to know how to satisfy those criteria, I think it’d be really useful and can pay you back a lot in the future.

And the other thing I would say is: “Don’t be afraid to experiment” and recognize that things won’t go well every day. In fact, most days there’ll be – there are days when things will be like the world is falling apart and days were you’re going to take over the world. The truth is always somewhere in between and so, look back at the pattern and hopefully, you know, it could be ups and downs but hopefully in the ups and downs, there’ll be a pattern going up.

Martin: How did you manage the times when you think: “Oh, I’m at the bottom right now.”

Konrad: Well, hopefully, once after little while, you will realize, you learn from the fact that those days happen but you have friends and colleagues around that you can discuss those things with. There are tough days and there are good days and hopefully the balance on those is used towards the good.

Martin: Konrad, thank you so much for sharing.

Konrad: My pleasure.

Martin: And if you have a very big website and you really want to understand: who’s my audience? Check out Quantcast. Thanks.

Konrad: Thank you.

Comments are closed.