Simon Chen interviews Avinash Kaushik

Simon Chen: Welcome, everybody. Today it gives me great pleasure to welcome to the call Avinash Kaushik. Avinash is an author, a blogger, and an analytics evangelist. I met Avinash at the recent Web 2.0 expo in San Francisco. My name is Simon Chen, from the Eight Black Group. So Avinash, welcome. It has just gone after lunch in the West Coast, yes?
Avinash Kaushik: Yes, exactly. I had some of the famous Google food a few minutes ago. [laughs]
Simon: So Avinash, let's sort of start from the beginning. I sat through your presentation with significant amount of interest, given that a couple of years ago, I met someone by the name of Perry Marshall, and Perry Marshall is quite well-known to the Google world, and certainly in the pay-per-click market--he essentially dominates the Google AdWords segment.
Avinash: Mm-hmm.
Simon: And Perry has been no stranger to testing and analytics, and recently, in about the last 12 months, has become obviously very well-aligned with another gentleman called David Bullock. And David has been really the pioneer of Taguchi Testing, which is obviously an advanced format of multivariate testing, which you referred to.
Avinash: Mm-hmm.
Simon: So talk about how you actually got started in analytics and why the need for it.
Avinash: Yeah. In the last three or four years, I have been focused exclusively on doing web analytics, just a pure form of all kinds of data and analysis and decision-making around the web. But my first job out of MBA school was working for Silicon Graphics.


And I have an MBA in finance, and I was going to do marketing as well, and I went to a class called MIS--management information systems. The web was just getting famous, and I took a class in database, and I said, "Wow! It's going to be pretty cool if somebody puts a database back-end with a web front-end. And it's going to change the world."


So I got a great job offer from Silicon Graphics, to say, "Hey, we're thinking of building a web-based application that will query data from a data warehouse, and would you be interested?" And I said, "Sure." And that's basically how I got into the world of business intelligence and decision-making, and most of my life I've spent staying both on the marketing side, as well as corporate IT--information technology--and building large-scale data warehouses and business intelligence systems.


And that was a lot of fun--multi-channel data, and how do you traverse tons and tons of data to figure out trends and patterns that you can exploit to make money or understand customers? I got an offer from Intuit a few years back, and they said, "We need somebody to build up sort of a world-class web decision-making organization." And I thought it would be very cool, because the web's everywhere, and it's just so complex, and just the challenge of it was probably the most appealing thing to me.


And I've been focused on web analytics since then and been expanding my horizons gradually, starting from Clickstream to going into this data warehouse-business intelligence stuff, with web data, bringing data from AdWords or Overture, mixing it with the CRM stuff or orders, and then competitive intelligence.


And then, more recently, in the last couple of years, doing lots of hardcore experimentation and testing on the web as the primary way of improving customer experience. So that's a quick snapshot of my journey.
Simon: OK. We'll jump right into a more advanced question, perhaps. At Web 2.0, John Battelle sat there with Eric Schmidt, the CEO from Google.
Avinash: Mm-hmm.
Simon: And asked Eric point blank, basically, that does Google track and interpret the data, right?
Avinash: Mm-hmm.
Simon: Because, John Battelle being, obviously, part of Federated Media and has a large publishing organization, is both a customer and then a user of analytics.
Avinash: Yeah.
Simon: So he's a customer of Google via AdWords, and he's then a user of Google Analytics as the owner of the site. So this has got a lot of people worried, and I want your take on why you think people are worried and what people can do to be calmed about that worry. You also wrote to your blog recently about this very point. And the fact that you made the point that the mountain is just way too high for even someone like Google, which perhaps I might disagree with you, but let's talk about this because I know this is a very important subject for a lot of people.
Avinash: Yeah, it definitely is. And I have to say that I am now a consultant to Google and my role here is as the Analytics Evangelist. And that means that I can see things from a slightly different perspective. I've also been on the outside as a user of Google data, and tools, and applications, and now inside Google to help Google do various different things that we can talk about later.


But overall I have to say that Google treats data privacy extremely sensitively. There're just so many eyes on the company and every single thing that happens. And they treat our privacy with the utmost and uber concern and care. And I think often people miss it. I'll get some things around Google Analytics specifically and say, "Oh you were watching my data, and increasing my AdWord spend, and things like that."


And, in fact, I encourage people to read the Terms and Conditions of, for example, Google Analytics or any of the Google products, and you'll see specific and explicit language that says that the data is collected and stored in a walled off quarter for GA -- cannot be merged with other things. It will not be merged with other things. And there are very severe penalties, as you can imagine, in an environment that we live in for any breach there.


Overall Google has extremely clear terms and conditions. It has extremely clear language about how we treat data. I definitely encourage people to read those.


But overall I believe that there is too much riding on the various services that Google offers for it to go around making mischief. That was the word sort of I used in my blog post.


And the other thing is that Google I think has often tries to go the extra mile.


Simon, I'm sure you noticed that it released the web history tool recently whereby you can have access to your own history. And Google is very transparent about the fact that as you use its services using cookie based technology, or your login, it's keeping track of the history. And recently it opened up the application where you can go in and have full control over your history that Google has a record of. And you can go in and delete URLs, you can keep them if you think they are helping improve your search quality.


And there's no other company out there that's actually doing... So Yahoo tracks so much stuff, for example, about users, but it hasn't provided the functionality in it. I'm sure it's thinking about it as well.


So I think each person will come to their own conclusion, but I always encourage people to sort of read the Terms and Conditions of the service and make sure that they're comfortable with the language that's in there.
Simon: I see where you're coming from, but I know this is probably the biggest point that's got Perry Marshall worried and his entire community. I was with Perry and Dave over the weekend in Chicago.


But here's the thing, Google can say all of this, but we don't know that entirely that it's true. I mean, just because you say that there's terms and conditions and there's a lot at stake, I agree that there is. And perhaps not mischief, if Google was caught doing this, it would be a significant breach, forget about mischief.


But why don't you think that Google, because obviously this is a growing concern and it's got a lot of people worried, why don't you think Google wouldn't get someone to verify that you don't analyze data? Like a third party, like a Price Waterhouse Coopers, or a KPMG. Somebody that would give it validity and credibility that "Yup, we've audited the Google systems and we can tell you categorically that they don't analyze the data."
Avinash: Yeah, and I think that's a great suggestion. And there are suggestions like that here in the U.S. market as well, where different stakeholders are suggesting options that can be used in order to reassure the market. And my understanding is that Google is very seriously evaluating all of the options that are on the table to see how it can... And what are the things it can do to ensure that its customers trust Google to keep and store the data in a very safe environment and for it not to be used in any manner that users are not comfortable with. It's something that the company has talked about very, very frequently.


At the end of the day, real people use Google's offerings and services. If the people will not trust Google to keep data in a very secure way, then it's to a great detriment to Google. So I do know that they are evaluating a mix of options that people are suggesting. I'm not sure what the official response is at the moment at. "OK, so this is exactly what we're going to do." Perhaps it's the suggestion that you just provided, Simon.


I'm not sure that the decision has been made, but all options are on the table and very seriously being thought through because there's nothing more important to Google. I can say this both as an independent consultant, as a part of Google, as somebody who's been outside and inside - there is nothing more important to Google than the trust that people place in the company. I would hope to hear more in the near future, for sure.
Simon: I have a six year old son, right? Now, if I said to him, "Look, your mother and I are leaving, and there's a kitchen full of candy, " right? And I said, "We're going out for two hours. Don't touch the candy." The temptation is just too great. I kind of think that you've got this massive amount of data on us, on every consumer, on every transaction. Probably there's no more significant repository of data anywhere on the planet. "Yes, it's a rock. We won't interpret the data, " but surely the temptation is there. Surely engineers have had a look under the hood, so to speak. It's Adam and Eve stuff, Avinash. I mean, surely there are people tempted to say, "Look over our shoulder, look over our shoulder. Let's have a quick look under the hood."
Avinash: The grading between two year olds and adults is that hopefully the adults can think better and have a better coping mechanism against temptation. But the interesting thing, Simon, though is that - let me reply to your question with a concrete example. So, I love using the MSN adCenter Labs. I don't know if you use the MSN adCenter.
Simon: Yeah, it's one of the few things that MSN is good at, right?
Avinash: Exactly! I love the MSN adCenter Lab. I have talked about it in my blog everywhere. It's amazing how you can go in and punch in any site, and you can get demographic information from a site, you can get some cycle-graphic information, keyword predictions. It's amazing. All the data comes from MSN and Microsoft collecting data from the people who sign up for it's services and then overlaying it with the visit behavior on their websites.


Now, I think, as you mentioned, Google perhaps has more data than that available about a user. Yet, if you go into Google Trends the tool, you will notice a wonderful application that gives you generic trends for websites of code search key terms and overlays it with press releases. But it has not yet provided all of this other data that some of the competitors are providing because, at least as I understand it, Google thinks that, "You know, that's separate data, and this is separate data, and this what we're going to provide."


You're absolutely right. I'm sure that there are people who are thinking, "Oh my God, it would be great, " but there are controls and restrictions within the company itself, when it comes to customer privacy and where the data is physically and logically located, that would be a huge barrier to anybody who might want to go on a fishing expedition, because again, the company treats it with a great degree of concern and importance.


I think this is a debate that's going to evolve. I fully anticipate that Google will do all it can to put more meat behind it's words, to ensure that it's users are comfortable, that the data is kept privately.


I just say, I was very gratified to see the example that not only is Google going to expose to you what is your web history, but it gives you the ability to go and delete your web history. And I think perhaps this is a small tiny thing.
Simon: Delete it off the Google servers itself, or just delete the essential cased record?
Avinash: My understanding is that when you go in and look at your web history you can completely delete it so nobody will have access to that.


And this is again, I completely realize this is a tiny small example in a world where people are worried about AdWords, and people are worried about massive numbers of more important things, but it's just a small tiny tangible example of things that Google will do to ensure that you trust that Google is going to keep your data safe.


And I fully anticipate that we'll see more of that.
Simon: OK, well, Avinash you've been well trained by Google's PR team.
Avinash: No, I've only been here a few weeks. [laughter]
Simon: But that's OK, we're not going to break your chops anymore. Give me another five minutes and I'll think of something else.
Avinash: [laughing] OK.
Simon: So tell me, you know, how you got to work for the hottest company on the planet and what is a Google evangelist? I mean, do you go around the Googleplex hugging people or do you chant stuff?
Avinash: [laughter]
Simon: I mean, what's a evangelist? Because when you presented at web 2.0, I wasn't quite sure whether you work for them or you work for yourself.
Avinash: Yeah. So let me clarify that. I am an independent consultant for Google, so not a permanent employee of Google, we have these two classifications, at least in the U.S.
Simon: Do they want you to be?
Avinash: Yes, they definitely had mentioned that, that they would want me to be. And I think that we came up to this really great arrangement that allows me to have a little bit more flexibility of doing things on my side.


And Google was very kind enough to say, it's OK, you can work with us as an independent consultant and we'll provide you with the flexibility that you want for the near term.
Simon: So you're the only guy I know that's ever knocked back a job with Google.
Avinash: [laughter] No, no, no, I'm sure there's tons more.
Simon: OK.
Avinash: But it is an amazing place to work. I was just observing to somebody that I had obviously been interacting with Google from the outside for such a long time as a user, and I've visited the campus so many times, because it's only half a mile from where I used to work at Ensured. Literally it's just down the road.


And when you come and work in Google from the inside it's such a difference, at being in the company and observing the company working from the inside. It's just such an amazingly different experience.


And I think that people, even me, who from the outside cannot appreciate sort of the kind of working environment that Google provides. It's such a contrast between looking from the outside and being inside in the company.
Simon: And what's your role as an evangelist? What does that entail?
Avinash: Right, so I actually work in the marketing team for Google. I work very closely doing three important things. And the first one is that I am working with the internal Google teams, such as the AdWords teams, which has potentially the Gmail team and other teams who offer products and services, Google, to help them do better analytics on the marketing efforts that Google undertakes.


I think a lot of people don't realize that Google actually itself does a lot of marketing. How can we actually create world class analytic systems so Google itself can measure efficiently how it's spending it's money. It's various marketing services, Google takes out press ads, it does things in newspapers, it does lots of AdWords, Google is on the AdWords using it to drive traffic to it's services.


So how do you optimize trade marketing analytics for inside Google for internal purposes. The second role is to work closely with the different analytics teams in Google to figure out what should be the product strategy and road map for Google, for the various tools that it owns.


For example there's measure map, there's reporting that you find in the Google AdWords interface, of the report center where you can get access to your data, there's Google analytics obviously and there's several other tools that Google has that does do analytics.


We have six or so now, so what should be the strategy, should it be one, should it be many different tools. How can we create applications that will be more easy for our customers. And that's one...
Simon: How big are these...
Avinash: Sorry, go ahead.
Simon: No, I was just going to say, how big is the analytics team itself?
Avinash: The specific number, I don't think that I can provide, sorry. But it's a good size. I was surprised at how small it was. It's very lean and mean. But, as I was saying, one of my big jobs is having been a customer and a practitioner of analytics in general, a web or otherwise.


I hope to bring a lot of customer-centricities to Google's applications and services. To bring the hardcore customer point of view, somebody who will actually use this to say, "How can we actually meet real world needs and create applications that are easy to use, provide a robust insights from our data." Because some of our applications are further along and some are not further along in terms of usage and the insights that they can provide. So that's the second big part, working with the products that Google has.


And the third one, I think, is more outward facing, is to be out there, and essentially help evangelize and outline the use of analytics. And I think that was the sort of session of Web 2.0. You noticed that I didn't talk about Google a lot. I mentioned that Google has a web optimizer product. I talked about Opramatic and Optima as competitors. Mostly that speech was all about teaching people the value, the hardcore value that you can get from doing experimentation and testing to fundamentally the change the customer experience on your website, which benefits your customers and it also benefits your bottom line if you're creating the relevant experience and offers and all these things.


But it was very much to go out and teach people the value of testing, the value of analysis, so that you can benefit from them. And if you want to use Google's applications, that's fine, but you notice I had screenshots from Google's competitors as well in that space.


So I'm very happy to do it. It's very outward facing role and truly it sounds very altruistic but really my charter is to help truly evangelize some of these methodologies so we can help create a better website for our customers and just a better web experience. You noticed I was wearing a Google shirt. They do allow me the freedom to go out and preach the, if I may say so, gospel.
Simon: And I agree, we certainly got the fact that you are very independent in running under your consulting entity when you presented it. So lets move on a little bit, you've written a book that comes out at the end of this month and which a lot of people which obviously are looking forward to, me included. It's called "Web Analytics: An Hour a Day." The strange thing is that you are donating all the profits to charities. So one is why write a book? And the other is your donation to charity. Let's talk about that for a second.
Avinash: Actually, I didn't want to write a book, that's not how it all got started. I just used to speak at conferences and people always used to tell me, "You should have a blog," and I just kept on resisting it for such a long time. Finally at the start of the year, one of my close friends, he convinced me that I should start a blog as a way of sharing some of my learnings and some of my insights with others at a more scalable level.


So I started to write a blog, but I actually watched many other blogs and how they behaved, so I participated in the system, before I started my blog. One of the most cogent pieces of advice that I got I read was from a Guy Kawasaki's blog -- I'm a huge fan of guy Kawasaki in general -- but one of the things the guy said, his principle is "when you write a blog, you should not write a personal diary because nobody gives a crap." [laughs] "What you should do is you should write a blog as if you're writing a book. It should have a start, it should have a finish, it should have a value that it adds to the blogosphere and then everybody will read your blog."


And I really took that to heart, and so when I write a blog, I write infrequently, I don't write everyday, I write twice a week at most and I have a start and an end and it was truly a way to communicate and teach people about analytics and I think that people misunderstand analytics. And so I said I'm going to tell people a new way of thinking about it, a different way of thinking about it. And that's how the blog got rolling, and after about six months, someone from Wiley called and said, "Hey, you should make this into a book." I said who would want to buy a book, it's already there on the blog, but they convinced me. And because I was hard pressed for time I had been initially resistant to write a book, but one of these days I was just reading something and I thought, you know what, I love doing things for non-profit type stuff.


And I wanted, this would be a great way, I know that I can write a good book that people will buy and that would be a great way for me to raise money for charity. And that was my proposal to Wiley, is give me a decent amount of money, don't give me little money, and they still give me little, but anyway. [laughter]


I'm going to donate it all to charity and they were very supportive of that. And I'm very grateful to Wiley that I'm going to take all of my proceeds from the book and donate them to two charities.


The Smile Train, which does cleft pallet surgery in the third world, in Asia and Africa where it's truly a taboo and can destroy a child's life to have that condition. And it's totally fixable for $50 which is amazing.


And then the other 50% we donate to Medecins Sans Frontieres, which is Doctors Without Borders. Again a charity that's very close to my heart. And in the end the decision Simon was very simple.


Every year we donate money to these two charities, they're very close to myself and my wife, but we could only donate a few hundred dollars because we're not really wealthy in any way.


This was just a way in which I could raise a lot more money than I ever have and so some good, a small good in exchange. And that was the story.
Simon: Well then well done, and I whole heartedly support it. For those people listening to this call or the podcast which it will be turned into, Avinash just give everybody your blog address and talk about where they can find details about the book.
Avinash: Yes, my blog's URL is www.kaushik.net/avinash or you can just do a Google search for Avinash, A-V-I-N-A-S-H, in Google and you'll see my blog come up pretty high. And there's on the left nav there's a little promo for the book, there's an image you can click through. And it would be great if you at least check it out, it should be a great book.
Simon: And you ca pre-order it on Amazon which I know a lot of people have done, so well done on that. I take my hat off to you for your commitment to those two charities, which I think are worthwhile.
Avinash: Thank you.
Simon: So let's keep going. People are laughing, you know Avinash, when you said you're not a wealthy person but you work for Google, so maybe we can talk about that.
Avinash: [laughter] Oh, You know there was, Guy Kawasaki had a joke on his blog a while back, you know it would have been great to be a receptionist at the building of Google ten years ago because now she's driving her own Ferrari, but the rest of us, no.
Simon: The rest of you are just humble beaten workers, right?
Avinash: [laughter]
Simon: OK, I find that hard to believe. I remember Guy Kawasaki from his garage.com days.
Avinash: Yeah.
Simon: And I went to a conference where one dodo was alive and well. And I think he's terrific too. Listen, let's talk about why you think traditional analytics is dead, and I want you to explain everybody what the ten ninety rule of analytics is.
Avinash: Absolutely. Let me talk about why traditional web analytics is dead. And actually way back when the first log file must have gotten generated from a web server, the Internet was starting, some very smart IT person must have noticed there are numbers in here that we can crunch.


And I think that fact, that there was just data spewing out from the web server was the genesis of the world of web analytics. And we had log parsers, then we moved to script tagging.


The one thing that never got lost, in fact it never evolved is that we kept staying sucked into the world of click scheme analysis. Purely trying to, taking all of these clicks that the web servers captures and processing them to infinity to try and find some insights.


And for a while perhaps it was OK to say, just the fact that my visitor count is going up is a great thing. Or maybe average time on sight is a good metric, or engagement is going to be defined by number of pages somebody has seen. And for a while on the web that was OK.


The more pages you saw, perhaps you liked the sight more, maybe you come back again and again. But the problem is that most web analytics still stop there. And what we need now is the web is such a complex ecosystem, and people come to websites for multiple different purposes, and trying to accomplish everything that they need on that site whether the company likes it or not.


And so what's happened is that metrics that were relevant in the past, and somehow are still buzzwords like hits. I think Jim Stern said "HITS" stands for How Idiots Track Success, H-I-T-S. I think it's very true, because it doesn't say anything. The world needs us to evolve beyond metrics that don't give us any insights.


And so the old world of traditional web analytics is dead, and we're living in this new world where what you care about a lot more is primary purpose. You care about multi channel impact. You care about knowing task completion rates of why people come to your website. You care about what's happening in an ecosystem of websites.


And all of these are much more evolved metrics, like customer satisfaction or net promoter. And that is the new world of web analytics, where you measure things that give you true insights into what's going on in your websites and your customer's head.


Because so far, click stream analysis simply tells you the what. What is happening on your website. But it doesn't give you the why.
Simon: Sorry to interrupt. But when you talk about click stream analysis, are you talking about the upstream, the downstream, or both?
Avinash: Both. But mostly, click stream analysis people are focusing purely on what is going on in the silo of the website. And that's all they focus on.


And just taking you back to our presentation of Web 2.0, where we talked about testing. And that's a great way, right? How do you know if somebody is seeing four pages on your website because they're frustrated and can't find what they want, or if it's the four pages that they wanted to see?


And it's very hard for you to understand customer intent on your website purely from clicks. You need a much more advanced way of doing analysis that helps you understand why people are coming. Why are you looking at this page, and why are you not able to complete your task? And how can we improve your customer experience?


And that is what I think is sort of the next layer of doing analysis. And a lot of companies are evolving to that already.


And that's what takes me into the 10/90 rule. One of the things is, our websites are spewing out so much data. And it's very complicated data, and what we do is, we'll typically go out and buy an expensive tool, and throw it on top of that data, and we wait for answers. We wait for somebody to come and precook the meal and feed it to us.


And that really, sadly, is not true. Standard reports do not provide you with as much insights as you possibly could need. So I came up with this 10/90 rule which simply states that if you spend $10 on web analytics tools, you need to spend $90 on the people, the brainpower. You need to analyze the data in order to get you the kinds of insight that will add value to your bottom line.


If you're spending $90 on tools and $10 on the analysts or the brainpower, then you are probably not going to get any level of insight that would move the dial for your profits or improve your customer satisfaction.


And it's simply because the web is a very complex interaction mechanism. Lots of different kinds of customers are interacting with you. And there is a lot of data to figure out before you can understand what is actually happening, what is success, and what is failure.


So that's the 10/90 rule. It encourages you to have a disproportionate impact in human beings who can apply intelligence to the data that you are collecting.
Simon: Do you think Google will ever roll out a consulting practice? One of the criticisms I've always had of Google, and people might be surprised to hear this. But I've always thought that Google hasn't done a very good job of teaching its customers to consume its own content. And by that I mean, it's got all these wonderful tools and applications.


But because of the massive trajectory that the company's been on, and the shortage of people and the lack of resource, that it's not had enough people in the field.


And I think, by the way, that Google has far too much focus on hiring engineers, and not enough focus on customer service people and customer facing people. But that's another point.


But do you think, going back to my initial statement, that Google will roll out a consulting environment? Where I think they could make a small fortune. Because, to the multinational companies, if they could hire an Avinash Kaushik, right?


To say well, look, we'll use the platform, which is now Urchin, but now Google Analytics. Now we're going to put you through to our consulting group that will now teach you how to interpret the data. Right?
Avinash: Yeah. I have to admit that at some level I definitely agree with you that Google is creating and providing lots of really fantastic tools and services that they're offering, and there is something to be said about putting an adequate amount of investment as well into teaching people.


Because unlike the search engine, which you really don't need to teach, Google is getting into creating these more advanced things that actually do need a certain amount of teaching.


And a small sign of the change is that there are people such as myself that Google is hiring, and many others more intelligent and important than I am, with these kinds of titles, to go out and truly try and change that.


And I think you will see things happening in the next so many months, where you will see solid signs that there will be a lot more of that.


In terms of going back to analytics, that's also another great example, Simon, as well, because unlike many other Google programs, Google Analytics from day one went out with a crack partner program, where there are these agencies and consultants who are certified after rigorous training here, by the analytics team at Google, to go out and offer value added consulting services to its customers.


I believe there are 25 GAAC partners, Google Authorized Analytics Consultants, around the world. Nine of them are in the U.S. And Google will refer you to a consultant, and you can pick any one you want, whatever you're comfortable with, after meeting with them, to find a partner that works best for you.


And this partner will come in and do exactly what you said, Simon. They will work very collaboratively. They know the tool in and out. And they're probably going to be geographically closely located to where you are and provide you a great level of service.


So with analytics, there's certainly something in place. I happen to believe we can do better. We can improve and we can do more things. We can continue to improve the program that already exists.


And so for the optimizer tool, as well, Simon, the tool that I talked about at Web 2.0, there is already a group of authorized consultants. So I think that slowly Google is moving towards the direction of addressing the concern that you express, Simon. I think it's just an evolution of the company and the fact that its tools and service is becoming more mature and complex, beyond the search engine.
Simon: Let's talk about your competitor for a moment. I'm going to discount Microsoft for a second, because I don't think that they've got a lot of traction in this space. But let's talk about Panama and the release of Panama and how that's impacting the folks at Google.


Because I think that perhaps one of the advantages that Panama and the folks at Yahoo could take advantage of is actually teaching people to get ahead of the analytics curve. And perhaps take advantage of a window that might be open for a little while, right?
Avinash: Yahoo traditionally has had very close relationships with agencies, wherever they might be, who use their platform. And have provided sort of custom analytics back to the agency, which has been a great thing.


Yahoo does not have anything that's competing with the Google Analytics platform, which is mass analytics. Let's just go out and let everybody have the power of going and analyzing their websites.


Also, based on my review, with Panama they've come up with some good features in terms of, if you're somebody who's monitoring and doing all of your ads spent for a company, you can go into Panama and get some good reporting.


So I think that's great. I definitely think that Yahoo's done a great job there. I'm not sure what their plans are in terms of coming into the global web analytics market space.
Simon: Do you think they should buy a company, like Omniture or one of the others?
Avinash: Actually, I happen to know, just from having publicly interacted with the executives at Yahoo, that Yahoo has a very, very strong team, an SDS team, Strategic Data Services, who have spoken at various conferences that I have spoken at as well.


So that Yahoo has a very strong data platform in the company. So there are definitely two schools of thought, in terms of would you want to take that same group who's doing analytics at a smaller level and do something globally or go out and buy somebody. And I think it depends on each company. Yahoo! Might choose to leverage the strength they already have and build over time. We'll have to see. They've been acquiring a lot of companies recently.
Simon: Well, they just announced their full purchase of Right Media, given it's hot on the heels of obviously your owner's purchase of DoubleClick. I suppose they didn't have much option. They had to actually execute a deal like that quickly.
Avinash: Yeah. The difference, I think Simon, for now - if I might be objective for a second and that I have been all through the podcast hopefully [laughter] - but the difference is, who are you solving for? Yahoo! Traditionally has been strong in the banner advertisement area. It's been very close to agencies and large media players. Yahoo!'s focus has been on providing and solving the analytics for that group of people.


Google's focus has been a lot more diverse. It definitely works with the biggest media players in the world as well, but Google's focus has been, "How can we let anybody on the street have access to the most sophisticated bidding system in the world? If you want to spend $5, that's fine. We're going to make it easy for you to do it." It's been more of what I tend to quote as "more of a democratic approach." Let's have lots more people. Let's have lots more people have access to analytics. Lots more people have ad words and make it easy for them to use it. It's more of a dispersed focus, and Yahoo! Is more of a concentrated focus.


I think this sort of cultural environment will drive sort of what Yahoo! Will do specifically in terms of web analytics. There's some really great people and great platforms, and I think that whatever they do, I'm sure it will be great.


One of the things actually - I should probably be bad-mouthing Yahoo! Or something - but in all honesty, as a matter of fact, I am in that sense a true evangelist. I think that it would be great if Yahoo! Comes out with a kick-ass web analytics platform that more people can use because a rising tide will rise all the boats. I think it would be a great thing if they do something like that.


That's why I very much welcome Microsoft's entry. In a few months, my dear friend, Ian Thomas, who works at Microsoft, he's going to launch Microsoft's own entry into the web analytics space. I heartily welcome Ian, too, because it's just going to be great for all of these people who want to have options to use.
Simon: I hope he has a plan B, Avinash. I hope he has a plan B.
Avinash: No, no, no! Simon, if you met Ian, you wouldn't say - he's just one of the smartest people I've met in my life. He's going to come out with something great.
Simon: all right, OK. Well, we may get him on the phone. We'll have a chat together then. Let's see what happens.
Avinash: Sure.
Simon: Avinash, let's talk about some things that you would do tomorrow if you were to walk into the marketing folks at - let's pick any company. Let's pick General Motors, for example. Let's pick a big, lumbering dinosaur of a company, right? So if you were to walk into GM tomorrow, what would be the three things you would do to the website? You'd sit the web folks down. What would you do? What would you insist happen?
Avinash: Right. The first thing I would do is definitely not take the website down, but the first thing I would want to do is try and understand why people come to GM's website, and are they able to do what they want.


I am absolutely flabbergasted at how few people in the world actually know why people use their website. It's just amazing. You might think that people want to come to Joe Morris' website at chevrolet.com to buy cars, but it could turn out that they want a part number. It's just amazing how few people actually design websites knowing fully well why people are trying to use their website and what are the problems they're trying to solve. That's the very first thing I would want to do, is do a hardcore, robust understanding of what do the people want - not what the company wants to push on the people, which is sort of the standard approach no matter where you go. That's sort of the first thing I would want to do.


The second thing I would want to do is actually take a hardcore look at GM's acquisition strategy for online because most of the time I find that companies, big and small, buy too much into the hype of the various, whatever the hype of the moment is. The hype of the moment is big pop-ups. The hype of the moment is banner ads. The hype of the moment is page search. The hype of the moment is something else.


We all get caught into hype of the moments, and probably don't do a very good job of spending our precious marketing dollars in making sure that they're optimizing our spending.


And there are a lot of really excellent things that you can do in understanding the value of all your acquisition vehicles, no matter what they are. And what they're doing end to end, not simply by bringing people to the website. And are they doing that efficiently? Are they the right people? But, also what's coming out of the door? Were they more satisfied? Were they able to complete their tasks? Did they convert? Did they give you money? Did they buy a car?


And most companies, big and small, are very poor at tying this entire picture together. So I would definitely spend... My second thing would be to spend a lot of time understanding what the current acquisition strategy is in terms of online marketing. Is it optimized? Are we using it appropriately? And is it actually solving end to end for what we want from our website?


I'll give you a quick example. I went to this company and they were very happy spending all their AdWords dollars, virtual dollars, buying up a particular keyword that was very standard for the industry. And turns out it was costing them 50 cents a click. No big deal.


And I said, "Oh, I wonder, let's do this little thing in the tag that's coming through. And let's start measuring cost per unique visitor." And it turns out the cost per unique visitor when we measured it, not the cost per click, for a unique visitor was at $5 a click. I said, "Oh my god. That's insane. You were just spending ten times over." But it's not a big deal because the company sells very expensive products. And they cost $300, $400. So, you know, $5 is not a bad idea.


And then we did true integration all the way to the end. It turns out it was costing the company $1,200 an order for an average order size of 250.


And this is just a tiny humble example of the kinds of things you can do it if you stick your data end to end, and optimize your acquisition strategy. It doesn't not only work for paid, but it works for everything else.


And the third thing I would do is I've satisfied the hardcore marketers by helping them figure out how to optimize their dollars and make more money. They've gotten a good understanding of our customers. The third thing I would do is roll out an aggressive experimentation and testing strategy at General Motors or any other company.


I think that we are often wrong about what our customers want and expect from our website experience, whether it's the color of buttons, the pages, the navigation structure, the location of the search box, the smiling picture of the CEO, we're just wrong about what our customers want. We're too close to our companies, and products, and services.


And the only want to be truly hardcore customer relevant is to ensure that you're creating experiences that your customers get to tell you what they want. And experimentation testing, whether you do AB, you do multivariate, you do experience testing, they're all ways in which you can create optimal experiences that your customers want, not the company wants to shove on top of them.


And I have seen companies who've done this improve conversionary by 10%, 15%, 20%, 30% in a span of months, not even years. The payback is so immediate.


Multivariate testing is so accessible in the immediate now you could start it, experiment on your blog, Simon, in less than a week. And it will improve conversion rate for you. Then that would be the third thing that I would focus on. So, understanding customers, optimizing acquisition strategies, and experimentation and testing.
Simon: I think they're all sound advice. I think the community I know that gets a lot of this is the Internet marketing community, which is I know... And it's where people like Perry Marshall and Dave Bullock play very well.


And Perry, as I said at the start of the call, he's probably one of the most respected global authorities on Google AdWords, but he talks a lot about people paying the stupid tax at Google. And what that is, I mean, is that 80% of people, let's say it's 80%, it could be higher, really don't know what they're doing with AdWords. They buy a bunch of keywords, Google changes its relevancy score and its quality score, and it's continually pushing the boundaries of relevance because of the user's experience.


But a lot of people, and I'd be interested to hear your thoughts on this, but a lot of people really don't know what they're doing, so they have a porter average experience with Google AdWords. And where Perry's become literally famous is in teaching people how to become much better at this.


And Dave Bullock comes along and then ads the whole Taguchi sort of dimension. But do you think that it is that hard? Do you think that's one of the reasons that search companies don't want to teach too many people out there how to be better AdWords Advertisers?
Avinash: I am not sure of the real number, how high the number is? But I definitely think that there are a lot of people who don't understand and able to use and optimize their experience of using AdWords or Overture for that matter or msn AdCenter and use a dollar as effectively as it should be.


And they actually, ironically one of the reasons, why Google bought Urchin to become Google Analytics because they said 'You know we think that there are a lot of you could be doing a lot more with the data and your spending with AdWords. And the first thing that was done with Urchin, the only feature those worked for the longest time was the seamless integration between AdWords and Google Analytics. So that people could understand lot better which keywords are working for them and which are not working for them.


So that, I think, is happening not as much as we want but I think what Perry's work is perhaps very important, bidding and buying keywords is as much an art and a science as you can imagine, its not a non-trivial process. It takes a good understanding, it takes a good understanding of your customers your keywords and brand terms and category terms the long tail the head and all of these wonderful things, in order to make sure you have an optimal bidding strategy, even optima ional acquisitioned strategy. And that I think there's needed to be a lot more elevation in the education and sophistication of the users because there is really at the moment -- and God only knows what the world will be five years from now, but at the moment and the near future in next few years there is really no better channel for any customer out there for general items or my personal books publisher.


There is no better vehicle to acquire most relevant traffic and use Google and Overture and msn and the search engines because there was a study published a few months back that 80% of the people in US, 80% of the people in US start there browsing at the search engines. I found it unbelievable, you see it has become people has to go to home page people start. The home page is the search engine, is the most effective way for you to acquire relevant customers you are looking for but you do have to invest. You have to invest and making sure that you and your company are getting smart about how AdWords work, what are the new answers of Overture what are the benefits that AdCenter provides and look you can do the demographic targeting this way, and I don't think that people understand that as much as would be optimal.
Simon: I agree, and I think most people don't even get the basics lets you take any competitive segment so if you taught into the stub in a Google Guinea pigs, right, you get over pay-per click ads come up. But if you hit refresh I could guarantee you that 90% of those advertisers aren't split testing that ad, right. That features such a basic element of AdWords, right.


So they are not running a split test to say let's test ad A and test ad B, and then we're are not testing display URLs versus destination URLs and that obviously is a minuscule product of what Perry talks about.


But when you got back into your AdWord account there are no messages, pop-ups, hints or tips saying, "Want a better result try split testing these two ads, right. It always says create another ad, but for novice they wouldn't know what that meant, right.
Avinash: Yes, and I have to tell you, Simon, I recently had a lot of experience with AdWords because I took a few dollars from myself and opened an AdWord account a few months back, and I created an ad on which I wanted to bid. It was just for a sort of playing with it so I knew the experience was like and sending people over to my blog as a destination URL.


And I started with the starter product, I had to make a choice: you want a starter or advanced? My feeling was Google probably thought that you know not everybody is being able to get... you know they might be get confused and not be able to use the product efficiently, so they created starter version.


When I was on the starter version I got some help I could get suggested keywords and things but it was very basic -- very, very basic. And then when I switched over to advanced version I saw a lot more options, I saw a lot more ways in which I could do testing, it allows creatives or destination URLs that you mentioned. So I think its an evolution process, but even in the AdWords product I think that there could be a lot more sophistication applied and created so that more people are able to leverage from some of the ways in which they can optimize their campaigns, and I think it's an evolution and certainly a challenge.


But I can not underscore enough that, no matter who you are, no matter what search engine you use, you should really invest small amount of time, if not a large amount of time, familiarizing yourself to what it means to do paid search marketing and what are all of the great ways in which you can benefit. And some of them I think you already mentioned, Simon, that Perry is teaching already but it's worth that sort of an education and understanding -- whether you get it from Perry you get it from some one else -- is worth it's weight in gold.
Simon: I agree, now wait a minute. Let's go back a bit. You created your AdWords account to drive traffic to your blog.
Avinash: Yes! I got $50.
Simon: You put $50?
Avinash: Yes. [giggles]
Simon: So why don't you just walk down the corridor to the head of AdWords team, and say send a buckle load of traffic to me in 10 minutes.
Avinash: [loudly] No it doesn't work that way. [laughing]
Simon: Sure does, why not?
Avinash: No, no, no Google doesn't work that, way it does not. I have to sign up with every body else and I have to top to my bids and creative. And you know for the longest time I couldn't get any traffic because I was not bidding high enough. [laughs]
Simon: OK well they must brain wash you really well when you start. Like it's day one and day two they just sort of shop the curtains and this neuron-linguistic programming comes out or something.
Avinash: [laughs]
Simon: You are such a good ambassador, Avinash, for Google. Let me...
Avinash: I have to tell you I have not attended my mandatory PR training yet.
Simon: Yeah, they probably hook up electrodes to your head.
Avinash: [laughs] No, no.
Simon: Even You better have children before they do that. So let's go on because I know you got some time contraints. I want you ask a couple more questions that are important. One of the things that hit me like a bolt of lightning during your presentation at Web 2.0 was the slide about true opportunity, right. Because we have a lot corporate clients here and a lot of people get carried away specially the marketing folks saying our website got four million unique this month. And it gets always its sort of the golf club talk. I could use obscenity but I better not. It's the bragging rights -- how much traffic did we get? And blah, blah, blah. But what really hit handful with this slide that you had a true opportunity, just talk about that a little bit about what that is, right.
Avinash: Yes, definitely essentially after having analyzed so many websites literally hindered of web sites over the years and having gotten pretty good understanding many different kinds of businesses, one of my seminal learnings was the fact that we always over-estimate the size of the opportunity pie, that's what I call it, literally a simple pie graph.


And what we think is the kind of conversion for any website is 2% and will be sale, we can execute a web strategy thinking that the size of the opportunity is 98%. The reality, in fact, is that not even remotely close to the truth. And on the slide I showed sort of this gradually shrinking pie. If you take out the people who bounced out your website in less than five seconds or after seeing one page, well, you loose 30% of that pie instantly. It's probably greater for your website.


If you actually collect data and compute it correctly and remove all the robot traffic and other crap might be in your data, you loose another 10-15% of your traffic. Now if you take away all of the people who are not your website to remotely buy. They're there for research, they want a picture, they want to do other things, they ended up there by mistake, you take all of these people who are not there for the primary purpose of why you exist that state of sale, then what happens is your pie went from 98 to probably closer to 30.


So essentially, your size of the opportunity to convert people, you were thinking you have to go out and it's so easy, I only need to convert 25 people out of the 98, and I get a whole huge bonus. Well, the reality is, your challenge is to convert 25 of the 30 that are left. That's a dramatically different challenge than what you think that you're trying to accomplish.


[coughing]


Pardon me. So I'm a huge fan of people understanding the true opportunity of what exists on their websites. One of the learnings that they will have, as have I, is that the size of the pie that they need to go attack is extremely small, and it's very, very competitive.


There's a post on my blog, if you just search for "true conversion rate" or "what's your real conversion rate, " you'll see these pies, and you can read the post. There's a lot more detail in there about how specifically you can figure this out for your website.
Simon: Forget the individual marketer, because I think the individual marketer that's invests in any sort of learning online is so far ahead of most corporates. Why don't you think corporates get analytics and testing and essentially just even how to build a website, in this day and age?
Avinash: You know, Simon, it sounds kind of ironic, but the bigger the company, the less sophisticated it is about the web. It's as true for tons and tons of companies. It doesn't mean there aren't companies out there that are giants and who truly kick major butt on the web. There might be a few, but for the part most companies who are trying to do business on the web - and we talked about a few during this conversation today - don't truly get the power of the web. They really don't get the web!


In fact, I just got this number the other day. There's a huge retailer in the US who dominates the US market. The marketing spend each year is $2 billion. So you can imagine they're somebody really big, right? Now, this is a huge corporation. It has a website, tons of websites. Their total spending on online marketing is $100 million, out of $2 billion. Imagine the kind of traffic you can get, the kind of accountability you can get for your marketing dollars, how much can you do with the same amount of money on a channel that is so accountable like the web online.


In any country, compared to the phone channel or the newspapers or retail box stores or anything, I think that the fact of the matter is most big companies don't truly get the web. That comes down from the top, and it permeates the organizations. Then I find that they're not actually sophisticated about the web. They're still trying to think of the web as they might have thought about their traditional businesses. Let's say their traditional business is phone channel or catalogs or retail stores, and they try and take those metrics and those mindsets and try to apply them to the web.


The reality is the web is nothing like any of those channels. It's not like a retail store. It's not like a phone channel. It's not like a catalog. Stop trying to create websites that are online versions of the catalog! If I wanted to shop in a catalog, I'll use a catalog. That's not the point.


There is a generation that is coming behind the generation that exists right now, and my hopes are being that this generation that's taking over from some of the earlier generation is going to truly move the web to maximize its potential. The web is a baby right now, and the generation that's coming now - and Simon, maybe that's your generation, that's coming and taking over the world.
Simon: No, no, it's not mine. I'm older than you, my friend. It's your kids' and my kids' generation perhaps.
Avinash: And they're going to truly unleash the power of the web. It's unbelievable, but even when one of the most advanced countries, when it comes to the web, in the world like the US, many big corporations use it in a very primitive way and don't truly understand how they can unlock it's potential.
Simon: I agree. I totally agree. That company that you talked about, that spends $2 billion and a 100 million online, I'll bet of that 100 million, that 98 million's on banner ads, right?
Avinash: No comment, but you're close.
Simon: All right, so let's wrap it up, Avinash, because I really enjoyed having this conversation with you, but I'm going to ask you some of these questions. You can blow me off, or you can say yes or no. Does the database of intentions exist?
Avinash: No.
Simon: Does Google interpret our data?
Avinash: Absolutely not.
Simon: Come on.
Avinash: Except, no, no. Google, for example, take web history. Google uses your web history and only your search results and nothing before it and nothing after it. It uses your search results and your search keyword history in order to serve the most relevant search results to you. Not even ads, search results to you so that you will find the right result at the right time.
Simon: OK.
Avinash: And that's all there is to it. But I think if your question was eluding to different Google services and products and the various islands of data, and they truly are islands. There's a very strong wall between them.
Simon: Is Docs and Spreadsheets a competitor to Microsoft Office?
Avinash: Not right now.
Simon: Not right now?
Avinash: It's an alternative to people who want online collaboration and just the set of most used features. If all you want is your most frequently used features and you want to have online collaboration then Docs and Spreadsheets is the perfect right tool for you.


If all you do is make text bold and frame it here and there and format it a little bit, you know, Docs and Spreadsheets are for you. If you're doing mail merge, you're doing more complex things with Microsoft Word you should use Microsoft Word.
Simon: Shame on you, Avinash, shame on you.
Avinash: Sorry.
Simon: Should Microsoft be in the search space?
Avinash: Oh absolutely, the more the merrier. I went originally, Live.com came out last year, I was pretty excited that they had some very innovative features in the Live.com data program.


I really liked the unlimited scrolling, you could keep scrolling on a page and more and more results appeared, you never had to go to the next page, and that's a great feature. But when it took over MSN search and it became Live search that feature disappeared.


So I think there is a lot of innovation still left in the search space, tons more. A lot of people think it's over but it's not. And I think that competition is great. I think it means that there will be a better experience for the customer and it's going to push all the main principles forward.
Simon: A lot of people, including the founders of Google, and I think even John Patel mentions this, believe that search is only 10% solved, agree or disagree.
Avinash: No, I completely agree, I completely agree. I think that it's an infant, it's a little baby. And we have so much farther to go and it's going to get so much better. I am an optimist. I'm a huge optimist.
Simon: Who's Google going to buy next?
Avinash: [laughter] Now you're really going to get me into trouble. I have no idea.
Simon: OK, shut the door, just shut the door to the office and...
Avinash: [laughter]
Simon: Roll down the window blind and you know, come on, pull out that little internal memo, you can just do a search on your desktop right now, right?
Avinash: I know, I wish it were that easy.
Simon: OK, all right, so folks we've been talking to Avinash Kaushik the world famous author, blogger, and analytics evangelist. I certainly recommend that you check out "Occam's Razor" which is Avinash's blog, and certainly order "Web Analytics: An Hour a Day" on Amazon.


It comes out at the end of the month. Avinash you've been an absolute pleasure to speak to, I hope I haven't given you too much of a hard time, I think you actually got off light today but we'll do this again. Any final sort of words of wisdom about testing and analytics?
Avinash: You know it has been a great pleasure to talk to you Simon. I enjoy your blog tremendously and I hope more people sign up to it, it's a great piece, keep up your style, I simply love it.


And it's been a good pleasure talking to you today. In terms of testing all I want to say is that just try it, you know. You want a chance to create great customer experiences online and now there are so many tools by which you can do it quickly, you can do it efficiently, and you can do it cheaply.


It used to be a huge hurdle, all of those are gone. Just experiment, don't you want to figure out how wrong you are? That's a great way to figure it out.
Simon: I couldn't agree more. So folks, the website again for Avinash Kaushik, his blog is kaushik.net/avinash and you can find his book "Web Analytics" on Amazon.com. Avinash, you're a good man, and one final question, does Google use our data?
Avinash: [laughter] No.
Simon: Come on. I was trying to get you there at the end.
Avinash: [laughter] It was a pleasure to talk to you, Simon, thank you for the opportunity.
Simon: I'll talk to you soon. See you, Avinash.
Avinash: Bye.


Transcription by CastingWords