The “Minor Orgasm” Of Using Analytics Well – with Avinash Kaushik

Avinash Kaushik is the man many of the entrepreneurs I interviewed turn to when they want to use analytics to uncover ways to increase conversions, decide what to build next or figure out why sales aren’t growing. He’s Google’s Analytics Evangelist and the author of two best-selling books on analytics.

I know this will be one of your favorite programs because Avinash will open your eyes to new ways of making data useful, and he’ll illustrate each idea with specific techniques and real-world examples.

The FULL program


video and analytics powered by Wistia

About Avinash Kaushik

Avinash Kaushik is Google’s Analytics evangelist and the author of best-selling books, Web Analytics: An Hour a Day and Web Analytics 2.0. He is also the co-founder of Market Motive, which offers Internet marketing certification courses.

He also writes a popular blog on analytics.

Raw transcript


Mixergy’s audio transcription is done by Speechpad

Three messages before we get started.

If you’re a tech entrepreneur, don’t you have unique legal needs that the average lawyer can’t help you with? That’s why you need Scott Edward Walker of Walker Corporate Law. If you read his articles on VentureBeat, you know that he can help you with issues like raising money or issuing stock options or even deciding whether to form a corporation. Scott Edward Walker is the entrepreneur’s lawyer. See him at WalkerCorporateLaw.com.

And do you remember when I interviewed Sarah Sutton Fell about how thousands of people pay for her jobsite? Look at the biggest point that she made. She said that she has a phone number on every page of her site because, and here’s the stat, 95% of the people who call end up buying. Most people though don’t call her, but seeing a real number increases their confidence in her and they buy. So try this. Go to Grasshopper.com and get a phone number that will make your company sound professional, add it to your site and see what happens. Grasshopper.com.

And remember Patrick Buckley who I interviewed? He came up with an idea for an iPad case. He built a store to sell it and in a few months he generated about $1 million in sales. Well, the platform he used is Shopify. If you have an idea to sell anything set up your store on Shopify.com because Shopify stores are designed to increase sales. Plus Shopify makes it easy to set up a beautiful store and manage it. Shopify.com.

Here’s the program.

Andrew Andrew: Hey, everyone. My name is Andrew Warner. I’m the founder of Mixergy.com, home of the ambitious upstart. And what I do here is interviews with entrepreneurs and other people in business to find out how they built their companies and then bring back their best lessons to you so you can go out there and build your business and hopefully do what today’s guest is doing which is an interview here on Mixergy where you get to teach others.

So the big question for today is how do you use analytics to measurably improve your business instead of drowning in data all day? To help me answer that question I invited Avinash Kaushik. He is Google’s Analytics evangelist and the author of best-selling books, “Web Analytics: An Hour a Day” and “Web Analytics 2.0.” Avinash is also the co-founder of Market Motive which offers Internet marketing certification courses. Welcome.

Avinash Kaushik: Hi, Andrew. Good to be here.

Andrew: All right. Now we’re having a little bit of Skype issues today. We’re going to try it without the headset, but guys if you’re hearing audio problems, then I’ll ask him to put on his headset. But for now we seem to be okay. So the first question that I had for you is I’d like to see a before and after. Do you have an example of company that didn’t do analytics so well, what they did and how the company improved as a result of it?

Avinash: Yes. Absolutely. I’ll give you a very odd example which is a website by a hospital. And this is for a very big hospital. I wish I could name them but anyway, it’s one of the big hospitals and one of their big, ambitious drive is to go from having a lot of the simple tasks being called into the phone center and into the hospital and get people to use the website to be much more effective. And in order to do that they eventually pumped in literally hundreds of thousands of dollars to put a lot of content on the site and a lot of quizzes and forums and videos and really all kinds of appointment scheduling and cancellations, all kinds of things on the hospital website. And at the end of all this process they weren’t sure if they were having any impact even though they knew that the page views on the site were up and the visits were up but they really did not know if their intended outcome was done.

So we did a very simple analysis for this website and we did, it’s analysis using Google’s analytics tool called Google Analytics. But you could do this with any analytics tool; you don’t have to use Google Analytics. The analysis that they ended up doing is created a very simple graph. On the X axis were all the pieces of content that were available on the website: videos, self-help, quizzes, forum, e-chats so and all this stuff. So the kinds of things that they had worked on. And on the Y axis what we plotted is the per cent of content available in each area and the number of people that were consuming that content. So get some data from the CMS system, get some data from Google Analytics, put it right next to each other and this very simple graph showed them very, very quickly that the three main areas where they had invested the most amount of money, creating new content and syndicating a lot of content, was being consumed by a grand total of 20% of people who came to the site. And 80% of the people were actually in the areas where they had invested no money: the forums, the e-chart and the areas about how do you figure out . . . there were a few areas where they could figure out what is the medication for something or what is the best way to find a doctor.

nd so these “what if” scenarios those were the three areas where people wanted the most amount of content. And clearly this particular hospital had invested in what it thought people wanted, and the reality is what people wanted was a completely different thing. So they literally, Andrew, reorganized the team and the number of people that were working on the website, and they started to focus more in areas where people were looking for content. And now the user satisfaction rates are higher. Some of the specific things that they were tracking in certain areas, “Am I having a heart attack now?” I’m kidding. But things like that. They were actually lower. And so they were able to move by even this very simple report into doing things very differently, using people differently and net-net spending less money on the website and yet having a higher impact on the website. So that is one simple scenario.

And the other one great example is an e-commerce website where the answer didn’t come from using a tool like Google Analytics but by running a survey. There’s a company in Canada called iPerceptions, and it provides a free survey to anybody in the world in 40 different languages. The survey asks four simple questions when people leave the website. The first question is, “Why are you here?” The second question is, “Were you able to complete your task?” Third question, “If you were not able to complete your task why not?” And the fourth question is, “Were you satisfied?” But that e-commerce site that put out the survey, and one of the things they found was only around 32% of the people coming to the site had anything to do with either considering possibly to buy the product some day or here to purchase now. Just 32%. The other 68% of the people were coming there for support, to write reviews or other scenarios, or to buy from some other retailer online or offline, and the website wasn’t doing literally anything to address the needs of these 68% of the site.

They would never have known that if they simply used a web analytics tool like Omniture or Google Analytics. This massive stream of people who are not satisfied on the website because the web analytics clicks are only recorded for the content you have. How do you know what you don’t have? And in that case, that qualitative analysis using a free survey allowed them to understand there’s a 68% of needs that they could meet and create a longer term loyalty with their brand and create happier customers. Those are two examples with completely different kinds of analysis that got companies to make decisions in fairly rapid order.

Andrew: I see. You mentioned a few tools here. And on your website I see lots and lots of tools. One thing that I’m concerned about is, and that’s why I used it in my introduction, is it feels like we’re drowning in tools. You helped break down all the available tools into five categories. Can you give that for people? And even that feels sometimes to me like it’s overwhelming so I’ll ask you a follow-up question after that of how do you know where to get started? But what are the five categories first?

Avinash: For the longest time web analytics was entrenched in just doing click stream analysis and this was slapping a quick Java script tag on Mixergy.com and understanding metrics like visits and page views and time on site and things like that. And one of the things I found early in my career in analytics is that data is not very rich in insights even if you’re capturing every single click on the website. So I created a new mental model, a new way of thinking about analytics on the web. And that’s the one that you’re referring to. And I call it Analytics 2.0, a very cheeky reference. But just to say, “Let’s kick up our game.”

And Web Analytics 2.0 it starts with analyzing click stream data using any tool that you have, free or paid. And that I say only answers the question “what?” What pages were seen, what sources people came from, what products did people purchase and things like that. So that answers the question, “What?” But then the second layer of questions that you need answer is how much. What was the impact of your website? And I find that usually we get so enamoured in this massive amount of click data, we rarely bother to think of all the sophisticated outcomes that we could have from our website. Of course, increased revenue from any activity we’re doing but also more number of leads we’re getting, increased RSS subscriptions, increased customer satisfaction, increased retention rate, net promotive scores, increased number of job applications to the website. Right?

So we don’t think of the impact of the website in a very sophisticated way. So I’m a very big fan of that second question: how much? What was the impact of my website in a holistic way? So that’s the second question. The third question, the next two layers are more interesting which is that even if we have every click that the person makes, the web is full of people who are individuals and weird in their own unique little nuanced ways. And it is very hard from the clicks to infer what people are thinking. As I said in my second example, if you don’t have the content your clicks [inaudible] is not telling you that so using surveys, usability and online remote testings. These things have become cheaper with every single day. And I say use those methodologies to answer the question, “Why?” Why do people behave the way they do on your website because the clicks normally don’t tell you that. So that’s the third question to answer.

Using testing and experimentation methodologies are really great ways to say, “Don’t let decisions on our website be made by Hippos.” Hippo is an acronym for the Highest Paid Person’s Opinion. So typically we’ll go into a room with all our great ideas and the Hippo is sitting there. The man or the woman who makes the highest salary in the room and the Hippo says, “I want a dancing monkey on the home page.” And you know what? You have to put the dancing monkey on the homepage because that’s the Hippo, right? The senior creative director of our agency, the VP of marketing or whatever and usually they may or may not have the right sort of understanding of our consumer base. And what I say is rather than going with the Hippo’s idea let’s take the idea from Andrew, take the idea from Avinash, take the idea from some other people, create a test on the website and let the customers tell us which versions work. So that’s the fourth layer of understanding which is experimentation to involve our customers actively, creating the right kind of content using the right kind of images, even setting price structures correctly. Some of the most beautiful, sophisticated tests I have run are trying to figure out what is the upper limit of what a customer will pay before conversions tank out. And it’s always a surprising answer. It is amazing that people will pay now, will convert higher on a rebate when you get a 10% that you get later rather than get $7 off right now. And so many people don’t send in the rebates so answering those kinds of questions and creating better websites.

The last part of where Web Analytics 2.0 is competitive intelligence. One of the great things, let’s say Andrew, I got a request from Mixergy to do an interview. “I wonder if I should do an interview Mixergy?” You know what I do? I quickly go into Compete or Google Trends for Websites and type in Mixergy.com and I’m going to see how much traffic he gets. I’m going to get to see what websites refer traffic to you, what keywords people use to come up with Mixergy.com. And without ever asking you for your data, I have access to all your data in order to make a much smarter decision.

So when I met a CEO from a big car company yesterday I said to him, “Look. You’ve got all the data you want for BMW and Lexus. Why don’t we go get that data for free, understand what they’re doing right or wrong on the web and use that to crush them? Or better still, to become better ourselves?” So the fifth layer of analyzing web analytics is about using competitive intelligence data which is freely available on the web order to understand what our competitors are doing that is better and using that data to create better website and better website marketing strategies. So those are the five layers: the what, the how much, the two whys and the what else. And it’s the best way to get a holistic understanding and create some really interesting, intelligent, sophisticated decisions when it comes to online marketing.

Andrew: Avinash, I read that on your site. I understand all those points. I knew the answer going into it but I just wanted my audience to hear it. But even as I hear the answer after having known it, it still feels overwhelming. If somebody’s coming into this either within a company where they have to convince the Hippo to take action or maybe on their own project where they have to take action themselves, where do they get started? Of all of these, what’s the first thing they should do?

Avinash: Typically I say there are two factors into deciding where you should start or what you should you do. And the first factor is the size of the company. I have this metrics in the second book, and I don’t know if you’re going to put this video in some kind of blog post, but I can send you that image and you can post it along the blog post if you write it, but it’s a simple matrix. And it says if you’re a small business, a medium sized business or a Fortune 100 company, here are the things that I recommend you do and I set a priority order.

So for a small business I’ll say, “You start with a click stream analysis tool, then you should move into a simple survey based tool and the last thing you should move into outcomes.” And worry about the other two things. Just do these three things in that priority order because you’re a small company, you don’t have resources, so I prioritize the order in which you will get the highest possible ROI. If you’re a medium sized company, there’s a different number of things you should do, not all of them, and the different priority order. And a different one for the large sized business. I’ll send you the image and you can include it if you want. So the size of the company sometimes dictates how deep you want to go; not everybody has to go all deep.

The second variable I use in making recommendations to say, “What stage of evolution are you?” And let the sophistication of your evolution dictate what you should be doing. So let’s say…I’m working with a start-up right now in San Francisco. We have more than we want here in the Bay Area. So I’m working with a start-up and these are so smart, sharp people, really analytically driven, solving a very unique problem in the realistic business. And in that case I said to them, “You know guys, we’re going to start you on some really cool competitive intelligence stuff and some very deep click stream analysis.” Because the stage at which they are as a company means that’s the way in which they’re going to find the best bang for your buck. So I don’t recommend, unless you have an army of two or three analysts in your company, to do the entire web analytics 2.0 spectrum all at the same time but to use the size of your company, i.e., the resources you have available and the stage of evolution, to pick the right thing.

At the end of the day I tell people, “Meet the person who is paying your salary and figure out what are the question that person wants to answer for the next three to six months. Don’t look beyond that. Just three to six months.” Then don’t use the tool you think is God’s gift to humanity. Let the question you want to answer, the what, why, how much, what else, let the question you want to answer dictate the one tool or two tools you should be using and go all out. Make love to those tools so that you’ll find the insights you need to provide to the person who’s going to make a decision about what salary you be making. But I think that people shouldn’t feel overwhelmed. What they should be educated on is we have these options and let our business strategy dictate which tools we end up using.

I want to tell you another really quick thing connected to this, Andrew which is my last job was at Intuit. It is about a $2.7 billion. So it’s a tiny little company; it does accounting software and personal finance software for those of your listeners don’t know what Intuit is. And when I was at Intuit I had access to all kinds of tools, lots of money invested in all these tools. And the company at that early stage was not very data driver in spite of being data rich.  They were not data driven. And at that time I postulated a new rule that I highly recommend to all people engaged in analytics, especially on the web, and the rule is called the 10-90 rule. And the 10-90 rule says if you have $100 to invest in making smart decisions on your website or your web based business you should invest $10 in tools and consulting for implementation of those tools. And you should invest $90 in the people who will play with that data and provide insights. Because you see it turns out, my bruising lesson was a lot of data you need and a lot of tools any you can get for free. Literally any tool you can imagine for any of the five components of web analytics 2.0 you can get free. Very good tools. But the thing that is missing in most companies is smart people who can take the data, understand the business, understand the strategy and go back and make sense from it. So I highly recommend Fortune 100 companies or small start-ups to have the 10-90 Rule because it’s not the tools that will make a difference in a company. It’s having smart people who understand the business and data who will truly revolutionize your existence.

Andrew: Let’s try this. First of all, you and I were introduced by Noah Kagan who came here, did an interview and said that you just helped him re-think the way he was presenting his business and I think you were the instigator of his number one selling product on App Sumo. Let’s go back in time and assume that you and Noah’s are sitting having coffee and deciding we’re going to go into business together. The idea is App Sumo; we’re going to sell deeply discounted software, web apps for a limited time. Noah’s going to concentrate on the relationships because he loves to talk to people and get those great deals. You’re going to concentrate on analytics. What’s the first thing that you do on App Sumo before there really is even an AppSumo.com?

Avinash: There are two tools I use a lot to understand market interest. And the two tools I use a lot to understand market interest is the first one is called Google Insights for Search. And what Google Insights for Search is essentially all of Google’s organic search data across the entire world available for you to analyze for free. So it’s extremely easy for you to go into Google Insights for Search, type in any terms that you want and try and gauge interest in that across any time dimensions, across any geographic dimensions and across any other category dimension you want. So a simple query I could run for Noah, Noah and I could run together is to say, “I wonder if there’s a lot of interest in applications that are helping people do online usability better?” So you can literally drill down into the category of applications and usability and run queries to see if the interest is growing or decreasing. The cool thing about using Google Insights for Search is at the very bottom you get two tables. The first table tells you that if people search in this category what are the top 100 terms that people type in when they search for applications connected to usability.

The second table on the bottom right tells you what are the fastest rising queries, the most statistically significant queries. So Noah and I could easily run queries in that tool that would help us understand market interest and that would help us understand associative queries to know which applications or problems people are most interested in and which ones are fastest rising by a statistically significant . . . that one table on the right is really important Andrew, because those trends are hidden below the surface. So most people don’t know these things. And we can get to them first and say, “Wow. In the last 90 days we’ve seen spike in interest in these categories. They’re not yet in the top ten but you know what? We’re going to create a bundle for that because we know there’s a rising interest.” We’ll create a bundle for that before anyone else can figure it out because we’re using, not research data collected by Noah’s gut feel, not because he had a drink in a bar, not by doing a survey by standing outside a supermarket and asking people questions. We’re going to use what people are actually typing into Google. I call this I made a grand recall, right?

The Database of Intentions. We use the Database of Intentions to figure out if there is any interest and if there is an interest specifically in which categories there is interest. So that’s the number one place of starting. And I have…I’ll tell you this. Last year we working with a very large CPG company. So let’s switch from Noah, a start-up entrepreneur to a very large CPG company and this company they . . . I’ll start with the people who run the shampoo business. And I say, “You know what? I did exactly this thing.” I wrote shampoo, mid-Western United States and I tried to understand what is the actual audience interest. And I found the fastest rising queries in this shampoo category in the last six months at the time was hemp shampoo. And I said to these guys, “Do you have a hemp shampoo?” And they’re like, “No. Never even thought of hemp shampoo.” And yet in that category the fastest rising thing over the last six months was hemp shampoo. It prodded them . . . never even considered remotely to think about and yet there is this enormous consumer interest. Now it turns out there were a few mom and pops making hemp shampoo, fueling this interest but this company was completely blind to that. Not launch a shampoo tomorrow but an idea that has an unmet need to a product that we haven’t thought about that could possibly be really big. So it’s a very direct, number one place to start is to tap into the Database of Intentions.

The secondary item, and I watched Noah’s interview with you and he mentioned some of his “competitors” who were in this app space but not solving that particular problem that he is. So another great place to go is either Compete.com for United States or Google Trends for Websites for the entire world’s data. And those two tools are really great Andrew, at trying to understand if the competitors in our space what are their trends? What are their traffic trends? What are their keyword trends? What are the destination trends? Who refers traffic to them and once people go to our competitors’ site where do you they go next? And if you do have direct competitors before you get into this space analyzing their data can be a great source of insight to see if the trend in interest is going up. You know what? There was this big spike. The guy got on Mixergy but the thing has all fall flat. After the Andrew effect wore off it’s all dead. So I’m getting into this far too late. So that’s the second thing I would do with Noah in terms of understanding the market in competitive dynamics before I did anything else even launch a business or have a website.

Andrew: First of all, thank you for doing this interview after seeing my interview with Noah. That was one of the most outrageous interviews he did. He showed his underwear in that. You still have respect for my work after that? He’s a character.

Avinash: Noah’s a fun guy and he goes on all kinds of really interesting tangents. He has such an interesting experience behind him. It’s always fun to watch him, even when he takes his underwear off.

Andrew: this is a sidetrack here, but I’m curious so many people helped Noah. Is he a customer of yours? Is he paying you?

Avinash: No, no, no.

Andrew: So why? So why do you help Noah? Help me understand the magic that certain people have that others help them. He’s one of these guys and I don’t know how to capture it in an interview. Why do you do it?

Avinash: I think for me people who populate blogs and spend time writing all this content with no advertisements or why connected to it because there is this tiny little gene somewhere that says, “Well, I have benefited a lot from other people’s wisdom in this sphere where content and knowledge seems to be free. I have learned so much from watching videos on Mixergy and how can I pay it forward in a very small way?” And you cannot help, especially if you live in the Silicon Valley, but to be surrounded by these passionate young people who want to go out and change the world. Right?

I met these two guys who just moved here from Iowa and they have this crazy, insane passion about crushing one of the largest companies in their space. “Man, you dream big.” And Noah’s one of those guys. He’s passionate about trying new things and solving problems in new ways and so for me it’s being able to take what little knowledge I have in domain expertise and saying, “You know? I can help you in this small, unique way and I’m going to try and help you.” Because you want to be a part of that passion. You want to be part of trying to change the world in a tiny little way. You want to be a part of failing faster.

One of the things that I say to people is that I passionately believe that God created the Internet so that we could fail faster. We could learn and . . . that’s the reason God invented the Internet. And then so when I see people like Noah he is the living embodiment of failing faster. Trying things, failing, learning, trying something else and keeping the lessons going on so that you fail less the next time around. So for me it’s the passion, it’s wanting to contribute something to the eco-system and just trying to make the world slightly better. And I really like doing that. There’s this high. When I’m sitting at Google, most of the products Google gives out are free. And most people will never know the cool things, features that I help build Google. But there’s something of a minor orgasm when something cool you built is out there and used by millions of people. You feel good. And there’s no money that can buy that happiness. And that’s the answer to your question.

Andrew: I think I might have my headline there, “The minor orgasm of doing something good.” The other thing that I wrote down here to ask you before we continue is Compete. You’ve brought them up a few times as a tool for seeing what your competitors are doing. Did you see the post that Reddit did a while back that showed Compete’s analysis of their traffic and Quant Cast and I think one other site, Alexa, and then they compared it to Google Analytics and they said essentially, ‘These guys are frigging nuts. Google Analytics is showing us completely different data and they’re all over the place.’ So with the understanding that’s not perfect, and far from perfect, how can we use these tools properly?

Avinash: And this again is more for people who really want to get into this to make some really big expensive decisions but it’s important to understand how data gets collected by these competitive intelligence tools. And that explains a lot of it. So I’m going to start with an analogy. When you use tools like Google Analytics, Yahoo’s web analytics tool, another wonderful free tool, Web Trends, Omniture, yada, yada, yada it’s like you’re standing at the door of a supermarket and you have that little counter thing that people hold on at Wal-Mart. And as each person walks in you press a little counter. It’s hard to get that wrong; you’ll get it wrong sometimes but that’s a pretty accurate count of how many people are walking into the supermarket. Using these competitive intelligence tools is like standing on the moon with a telescope and the telescope is pointed to the door of the supermarket. It’s as good a telescope as you could find on the moon. And you’re using that telescope to figure out how many people are walking into the supermarket. You’re not even going to be close.

So it’s important to remember that you’re comparing those two different counts. The guy on the moon with a telescope versus the guy standing, or the girl standing, at the entrance to the market. So what you want to do is you want to use tools that can help you find the best possible telescope. That’s the way to think about it. Because it’s the only way you will get at your competitor’s data. The competitor is not going to give you the data collected at the door. He’s just not going to give you that. So when it comes to Alexa, the way Alexa used to create was a long time ago Alexa was the dominant toolbar on the web. And when people used the tool bar to visit websites the toolbar collected the data and Alexa mashed it into competitive intelligence data and shared it out. Now the problem is really hard for you to find anybody using Alexa anymore. So the data that Alexa collects is extremely sparse and I actively advocate people not to use that as a source of data because it is one of the sparsest sources of data you could possibly consider using. It is so rarely used by such a small, particular type of people that it’s just not good enough.

If you think about Compete, the way that they collect data it’s a mash. So they buy data from ISPs in the United States. So Andrew, whether you want to give your data or not your ISP is collecting your clicks. And it doesn’t know that it’s you, Andrew. It doesn’t even know who you are or anything about you but it does know that somebody on my network clicked to go Google.com. Clicked to go to Occam’s Razor, my blog, right? So they know that. So Compete buys the data. Compete also has a small panel. On their website they say how many, I forget how many, but several tens of thousands. They have a panel; they monitor their behaviour. Compete also buys data from some toolbar providers. So what Compete has done is they’ve taken these two or three sources of data, combined it together to get a much better signal than anything Alexa could get. And if I remember right on Compete’s website they say their sample of people is around 3 million people of 5 million people in the US. I forget the exact number. Something like that.

Now if you move to HitWise, HitWise is one of the largest purchasers of ISP data in the US. Their sample in the US is around 15 million but they don’t have other data sources but you could use that for something else. QuantCast also uses an amalgamation of sources; their sample size is much sparser than Compete or much, much sparser than HitWise. But QuantCast has another source of data whereby they will give you a tag and say, “Andrew, if you put this tag on Mixergy.com it’s like having the person at the entrance to the supermarket. We’ll count your data accurately.” So if you want to use QuantCast data only use it for websites that are quantified. You’ll see the logo. That’s basically accurate data. And then when you use the tools I mentioned, like Google Trends for Websites essentially it uses Google’s toolbar data, it uses Google’s search data, it uses a whole bunch of data and you can imagine how many people use those sources. It’s one of the largest possible sources of data you can get. All anonymous, non-PAI and if you use Insights for Search it’s essentially basically Google is saying, “We’re collecting all the data for searches on all sources of Google.com. It’s the greatest possible source you could ever imagine getting. Damned near perfect source of data, right?”

So what I tell people who want to be serious about this or who read that post on Reddit is that you have to pick the right source of data for the right question to answer. So when I’m analyzing search data as I did in the case of Noah earlier, our hypothetical example of Noah, I’ll not go to any other source than Google Insights for Search because that’s the direct source directly from Google. That’s where I’m going to go. If I want US data for websites that get more than 10,000 visits, Compete is a really good source to go. If I had money I would go to HitWise because there’s no free data there. But if I want to analyse data for the rest of the world Compete is a very poor source because they only report US visitors.

So if you wanted to analyze a website in Argentina, typing into Compete would only show you visitors from the United States that went to that site in Argentina. It wouldn’t show you the Argentina data. So then I would say, “Andrew, you should use Google Trends for Websites.” It’s important that you get educated about what is in each source. It’s important that you get educated on what is the strength of each source because of how it collects data. If you get educated of these two things you’ll end up using the right tool for the right question to answer. There isn’t one tool that is God’s gift to humanity yet. Maybe there’ll be one day. But smart people will do these two things and use the right tool in order to answer their questions and then use broad based trends. I’m never going to use . . . when I analyze Mixergy’s data I actually don’t care about the actual number of people who came to your site in Compete.com last month. I care about the last six, 12 months worth of trends. I look at the trends. And the trends normally will give you close enough information about your competitors that it is quite useful.

Andrew: All right. So we understand now how Noah, or anyone else, would figure out what to launch. First thing you do when you launch is you fail fast. Product doesn’t work out so well or maybe you didn’t pick the right product because you didn’t analyse the data right. Maybe you didn’t put the button explaining where to buy clearly enough. Maybe you made some other mistake. How do you know what to fix and what to improve. Let’s not say “what to fix.” Let’s not assume that my audience is going to make a mistake right away. Say ‘what to improve’.

Avinash: How do you learn better? How do you learn? So here’s a great example of Noah and how he’s quite smart in many of these analytical approaches. When he asked me to promote one of his bundles, the optimisation bundles, he sent me a unique url, even in bit.ly. And when I typed in the bit url I noticed he was smart enough to type in [inaudible].com /?parameter=Avinash. So here’s this smart guy who, unbeknownst to me, is tracking my ability to drive traffic. He didn’t tell me that. Of course, I could see he’s doing that. Now when he sends you that link and me that link he sent different URL’s to both of us. And what Noah is doing is saying, ‘Well, all these guys think they’re ‘influencers’. But you know what? I’m not going to take their word for it. I’m not going to take their Twitter follow count as their word for it. I’m going to send them all different URLs and use Google Analytics to figure out how much traffic each one of them sent to me and how many of those converted because it’s possible Andrew sent 100 people, Avanir [sp] sent 15 people. I got 10 conversions, he got two.”Well, now you know.”

I want to use this very simple, specific example to say that when entrepreneurs launch these things and you invest amount of effort, if not money, in acquisition, this is an example of acquisition, do this simple little trick because it’ll help you understand who is able to bring you audience and who is able to bring you the right kind of audience. And you fail fast because it turns out the top ten sources of traffic you thought would work out don’t work out. But it does mean having the forethought, creating those short links, making life for you and me easier because they didn’t send us a big-ass, hard-to-type link. He shortened it and sent to us, unbeknownst to us with a tracking code in it. Do that, right, so then you’re much smarter about what to do the next time you launch a bundle, the next time you launch something. That’s one thing.

Second thing to do is to experiment. And I want to give you a very specific example that I bumped into the other day. One of the qualitative tools that answers the “why” question in my model is doing surveys and one of the new survey tools is called KISSinsights. It’s built by Hiten Shah. You might have bumped into it at places. And Hiten wants to know how to price his product. He doesn’t know what is the market’s ability. So his price, the “free version,” to be able to collect 30 responses. Now he has no idea if giving away 30 responses is good or bad, right? Next to it he has a premium offering and it’s $29 a month and you get 1,000 responses in that. And then, he’s such a wily character, he has the last column is an unlimited offering, Andrew, unlimited number of surveys but for 24.99. Less than premium. Remember there’s a limit of 1,000 in premium. Unlimited. It’s unlimited and it’s cheaper.

Now what cleverly Hiten is trying to figure out is can he encourage you to pay up front for a year and get the unlimited deal rather than getting a more expensive monthly deal. You know how he’s figuring out which one of these works? If you stay on that page for 15 seconds a little survey pops up from the bottom; it only has three questions in it. The first question is, ‘Is our pricing clear?’ Because actually it’s not clear. It’s deliberately confusing. The second question is and a box pops up where you can tell him why it’s not clear. Clever little guy, collecting good data. And the last radio button if you click says, “I’m really interested. Here’s my email address.” Now what he’s doing is he’s using qualitative analytics data at launch to figure out if he can con you into the right bucket. What’s your propensity to pay? And he’s done this from day one. Now he’s eating his own dog food because KISSinsights is the survey that powers the thing I described but using it on site such a clever little entrepreneur. Using analytics to figure out the right price point. So that’s another example of how an entrepreneur can start day one. You don’t have to make everything free. Collect qualitative data. And of course, day one if you tag your site correctively with Google Analytics or Yahoo analytics you’re of course measuring conversions and clicks on that page. But this qualitative aspect is the thing that gives him insight into the person’s head and what they’re thinking when they land on the page.

The third thing I would say is increasingly we’re creating application inspired websites. And they’re not straight web pages, right? And I think that what that means is that you’ve got to do some advance tagging using web analytics tools. And most entrepreneurs don’t do that. They’ll say, “It looks like it’s like it’ll take half a day. I’m not going to do it, I’m going to launch.” Now the problem is you’re going to launch and be blind. But if you use sophisticated methodologies like custom variables and event tracking, these are two literally gifts to an entrepreneur that is creating fluid experiences on the websites on day one when you launch you’ll be able to use web analytics tools to measure the experience inside the application; flash, RIA, html5 based websites we are creating. And it can be phenomenally fantastic to try and understand where are people dropping off? How are they using it? Are they finding the right buttons? Are they making the right clicks? How long do they stay? How many times do they come back? But most people don’t do that.

But you really want to spend half a day that’s required to understand these sophisticated methods. You can use them for free. You have to pay no money to track this, right? But it is an amazingly fantastic way of failing faster. And most start-ups that I work with, that I advise with, I strongly say [inaudible] or loyalists or evangelists or first wave or whatever you want to call it and use that data not just to create better buttons and better content and better forums but to truly understand if our perception of the problem we’re trying to solve is accurate. In a start-up that I’m working with as soon as you get in you don’t know this, but when you log in Andrew, you see a button that takes you down A, one path. And I’ll see another one dynamically created using the Google Website Optimizer [inaudible] and it’ll take me down another path. And then very easily using a free tool we’re able to understand if A experience is better or experience B.  When it comes to customizing the radio stations, the streaming radio station that we have created, in such a clever little way most entrepreneurs would say, “You know what? I know what I’m doing.” And I think many of them are smart, young, intelligent, they know what they’re doing. But using data as a component of getting additional insight by just investing in a day or two, I think is phenomenal. Spend half a day in understanding, spend a day in coding. Off you go.

Andrew: What was the name of the company with the radio station? We lost the connection for a little bit.

Avinash: Oh, I didn’t say it because they’re just still stealthy. We can say there are tons of radio stations so it’s easy to imagine what they might be doing. But the unique problem they’re solving is the hardest thing to do for you to get you to stick with one of these offerings is that you customize. If you customize you’re there for life. Literally. And it’s really hard to get people to go from the default or the first artist you type to just saying, “I’m going to create a custom radio station.” And so what they’re saying is let’s figure out how to solve that unique problem because the qualitative analysis for or competitor suggests that is the point of failure. Not the fact that you have a good stream, not the fact that you have the artists’ portfolio, not any of the other things. It’s that you get somebody to customize, that’s the hook. And how do you do that? And they’re testing various ways of doing that.

Andrew: I’m glad you brought up Hiten Shah earlier. He was on Mixergy twice. That second interview where he talked about lean start-up methodology at his current company and how he should have used it at the past companies that didn’t go so well. One of my best interviews ever. One of the things he brought up about KISSinsights, his survey application, was that people don’t know how to come up with the right questions which is why they don’t ask questions. And that’s also why he offers a batch of suggested questions to new users. The batch of suggested questions is not very useful. How would you suggest that someone come up with the right questions? A new company or maybe an old company, like the one you mentioned with the shampoo. How do you come up with them?

Avinash: Actually the one core challenge is what people say is that we’re going to go out and we’re going to ask all the people in the company, whether you have two or 2,000, what are all the possible things we could ask and then we’re going to create this big assed 50-question survey and show our users we have absolutely no respect for them or their time. And that’s the worst possible thing to do. I think that the method that we’ve found to work the most is to get the initial set of questions to be extremely short and to think through the core fundamental questions that anybody would want to ask. So when we created the 4Q survey with iPerceptions, the on-exit site survey, what Hiten has built is a fantastic page level survey. So those are the two kinds of surveys: page level, site level. And KISSinsights is a page level survey. Wonderful, wonderful survey. But when we created the 4Q survey, also free, I had written a post saying these are the three greatest survey questions in the world and nobody should ask any more questions. And I said any website wants to know primary purpose, task completion rate and why not. And that was the basis. So starting with questions that give you a peek into the customers’ head is the best way to go and those can be standardized. And then what you do is give people an ability to customize after the first month.

The thing that I tell people and Hiten is doing is saying, “Look. What we’re going to do is we’ll give you the first blush of question. You start with this. You start with these three things initially for the first month. Don’t question us. Trust us. We’re experts. We spent a lot of time on this. Take one month only, use our questions. And then use the information you’re getting from those three questions to decide how to customize.” What people end up doing is they bring their own limited experience to the table and say, “I’m going to customize from day one.” And because they have very limited experience they are . . . what’s the technical word? Screwed. From the beginning. Because they’re not benefiting from the knowledge that is available outside.

There are there things in the world that make decision making harder. The known knowns. This is Donald Rumsfeld I’m quoting here. Most obscure quote ever. But Donald Rumsfeld said about Iraq, there are the known knowns. The things you know you know. That’s great. Anybody can solve, any entrepreneur. Things you don’t know you know. And that’s also an easy problem because you know that you don’t know something and you’re knowledgeable about it. But life gets screwed in the third category: the unknown unknowns. Things you don’t know you don’t know. And I love that quote. I’m not a big fan of Mr. Rumsfeld but I love that quote.

So what I tell people is open your mind initially to being exposed to the known unknowns, things you don’t know you don’t know and when you start with the first recommended questionnaire you’re exposing yourself to the unknown unknowns. Take that data then use it to customize. I recently had the privilege of seeing some of the new questions that Hiten is coming up with as a suggestion. And one of the things he’ll end up doing perhaps, I’m not revealing anything secret, is to say, “We’re going to create a series of suggested questions by purpose type.” So you could come in and say, “I’m an entrepreneur selling applications.” Okay, well, here are the first initial survey that will expose unknown unknowns to you. Or you could say, “I’m an e-commerce site.” Boom. “I’m a blog.” Boom. “I’m a . . .” Boom. So by having a small portfolio of questions the purpose of the site I think will be a great first step to expose yourself to the unknown unknowns. And then Hiten and his team are building functionality that will then allow you in the second month to go in and say, “Ah. I understand how people react. I understand how they speak. I understand how they refer to my products. What a surprise. They don’t call it what I call it. Now I’m going to go in and customize.” And I recommend that process to listeners of our video as a great way to then kick-ass with this kind of an approach.

Andrew: And by the way, he does have that on now where you can, from a drop-down menu, say you’re a blogger, you’re e-commerce. But I didn’t know whether he was showing it to everyone or not. One of the issues we’re talking about with any of Hiten’s sites is I can’t tell if I’m seeing an experiment or you’re seeing an experiment and is I’m seeing what everyone else is seeing because he’s constantly testing.

Avinash: I know. It’s the price of one of these fast moving, lean entrepreneurs. But over time I think Hiten has done a great job of evolving by this rapid iteration initially. And now when you go into one of his established products, like Crazy Egg, things are pretty settled and all of us are seeing the same thing and he knows exactly what works from perspective. So regardless of whether you use Hiten’s products or not, the other people who are listening to us, and regardless of whether you are in the market, some of the ideas he’s talked about to you and to other people I think Hiten personifies failing faster. I have yet to meet people who are as passionate and driven as him to leveraging these free technologies on the web and inventing some of these free technologies on the web and finding the right niche and the right market. And he’s also very good when he’s passionately launched things that nobody wants and he’s moved on quickly. This is such an important thing for people to understand. You’ve got to be able to give up and move on.

Andrew: That’s one of the big questions that I ask. How do some people know that it’s the right time to give up without being quitters? And there’s some magical solution there that I haven’t figured out but you can see some people they’ve just passed on past projects that others would have spent their whole lives trying to prove out. One last example here. We talked about a brand new company, we talked about a company after it launches. What I’m wondering now about is a company that’s struggling. The product is build, it’s out there. Now we need to find a way to get customers, your users to come back. But for some reason the ones that are coming in actually aren’t reacting well, aren’t buying, aren’t interacting. I’m trying to come up with a failure scenario but the problem is there’s so many different goals that a company would have and with each goal there’s a failure that goes along with it. But maybe you can pick one. A struggling company, what can it do?

Avinash: I will advocate two things. One perhaps not surprising; another one surprising The first one thing I would recommend is using things like on-exit surveys and qualitative analysis to try and answer those two questions that I mentioned: why are you here and were you able to complete your task? So you start to get a really good understanding . . . the primary purpose of people coming to your website is lined up with the reason for which you have created the website. And it is astonishing how rarely entrepreneurs have a good handle on this. I think they’re very passionate and go out and build things but it turns out that just like in the case of Ideal that became Twitter, it turns out people come for a certain purpose and really you’re not building sites for that purpose. Or you think you’re solving one problem but you don’t. So I encourage people to figure out how to get into the heads of people who are already there to understand what might be happening. Also if you use things like…I wrote a post at the beginning of the year and the post was geared towards…it started with these two circles in a Venn diagram. And I said one of the two metrics I recommend people look as a basement metrics, a foundational metrics, not the ultimate nirvana metric, but a foundational metric, is a metric called bounce rate. And what the bounce rate metric measures is this phenomenon from a customer perspective: I came, I puked, I left. So it measures the number of people who come to your site and puke and leave right away without giving you a single click. And so the post was geared, Andrew, towards helping you understand what is the purpose of a page, one circle, and what was the intent of the person? And what I advocated is if there isn’t overlap between these two things, what is the purpose of the page and what is the intent of the purpose, you are going to die. It is only when those two circles overlap, not hugely, but at least a little bit that you’re going to win. So that post outlined six different ways in which you could understand the intent of the people coming to your website using tools like Google and Yahoo analytics. So I’ll send you a link to the post and you can post it and people can see there are six methods to figure out the intent of the person who comes to your site. And that’s the way to win. That’s the way to figure out why you’re struggling. It’s possible that your circles have no overlap; there’s no overlap between purpose and intent. But I don’t want to go deep into it because we don’t have a lot of time. So I’ll send you a link.

Second is a very clever way. There are a lot of services now like Loop11, UserTesting.com and a number of these things. And what I have done in some entrepreneurs, Andrew . . . and each usability you initiate on the site is around $15 or $20. I forget how much it is but something like that. What I have done in some companies that I have been advising I go to UserTesting.com, I pick the target audience group, age, male, female, sophistication web, whatever. I’ll buy 10 of these sessions and I will get 10 of those people to complete the exact same tasks that I want them to do but on my competitors’ websites. Not mine, my competitor’s, my competitor that is successful.  So I am going to use $200, just $200, to find the audience that I think is interested in the thing that I am selling or my website. I’m going to get them to use it for my competitor who I know is working out really well, and then I will watch those video sessions. And I will watch what people perceive are working right for them. And this is a great way, fantastic way of opening your eyes to outside your own obsession of your own website. It’s trying to understand, “Wow, the dumb-assed thing I thought my competitor is doing is the killer feature. That’s the thing people want.” And it’s been so revealing for just $200. I can pay from my own pocket; it’s not a lot of money. For $200 to get these videos that I could watch in half a day, and really hear people talking, watch how they’re clicking and get insights from what they’re doing on my competitors’ websites. Normally what we’ll do is we’ll go and do that on our site. But I haven’t done anything really new, big or earth shattering. What I want to know is I am already struggling. I know that. And I know some things about it. I don’t know why the other guy is successful.  And why not use $200 to figure that out so that’s the method I would use.

Andrew: So many people have brought up UserTesting over and over in my interviews, you’re the first person to say use it that way. To have the people who are on UserTesting use your competitors’ sites and then copy their best ideas. I didn’t even get a chance to ask you about Market Motive which I’m so frigging fascinated about that I included it in my intro. Want to just quickly want to tell people what that is?  And I’ve got a ton of questions. Maybe I’ll save them for a second interview.

Avinash: Absolutely. Market Motive came out of . . . I have the privilege of doing guest lectures at Stanford and UCLA and other universities and one of the things I found is that the curriculums in these universities when it comes to internet is a bit older. No textbooks about the stuff we’re inventing tomorrow, right? So my friends John, Michael and I started Market Motive as a way to get people certified and educated in seven different web disciplines using the latest, greatest most current knowledge. And videos, lectures. We have a three-month course that people can take in web analytics, online PR, SEO, SEM and social media and things like that. And our hope is that when you come to Market Motives and take this three-month course you’ll get the kind of education you might get at a university but it will be the latest, greatest most leading edge. And when you take the course you watch the videos, you have to take tests. At the end  you have to write a PhD-like dissertation which goes to a large percentage of your grade because we’re trying to grade it not only that you understand the technical things but that you have absorbed the mental model of what it takes to be a kick-ass marketer out there in the real world. And at the end of it if you pass, you get certified. A bunch of people unfortunately don’t pass but our hope is to give people a place where they can learn about the latest and greatest internet marketing. And it’s sort of a fun hobby. And we’re excited about the people we get from all around the world at many different levels who go through the curriculum.

Andrew: I’m the kind of person now, I’ve got an hour’s worth of questions. I’m not going to hit you with them, because I can see we’re at the top of the hour.  I want to know how you get your customers. I want to know how you failed when you launched and what you adjusted.  I want to know why the page looks the way it does. And it is a very, very well designed web page. I was able to see instantly what goes on. I want to know how many people are willing to go through . . . this is not a simple online course. When I say “online course” people are thinking one of these get-rich-quick courses that have these quickie solutions. No. This is frigging intense. I don’t know how you convince people to sit through all that work.

I’ve got so many questions. I hope you come back. Before I even invite you back . . . I’ll invite you back now and let me just say this. Sometimes people say, “Andrew, why do you want to deal with the Skype issues and the scheduling issues and all the research that goes into an interview?” Anyone who made it this far into this interview has to understand why I love doing this. This kind of conversation you just can’t have every day. I’m going to go home here and I’m going to have interesting conversations over the weekend with people, but they’re not going to be this densely interesting and this densely packed with useful information. And this was my first time meeting you. I’m so grateful to you for doing this interview, for giving me this interview. And if anyone out there is wondering why I do it I’m sure at the end of this interview you’ll understand. So, thank you.

Avinash: Thank you, Andrew. It was a pleasure. I have to tell you that when I watch many of the Mixergy interviews I think the difference is between…a lot of people do short videos that probably add some value but not a lot. But being on Mixergy and being on Charlie Rose in the sense that you get an hour, you get in depth, really intelligent conversation.  What you’re doing here is really amazing. An hour is a big commitment in time, but it is undeniable that at the end of the hour the loyal people who listen to you and stick with you get an incredible amount of value. So thank you for what you do.

Andrew: Thank you. The website. Let’s give people your personal website. I want to make sure I give it right and I don’t have it here. It’s dot net, right?

Avinash: Yeah. It’s Avinash.net/Avinash. Or just type Avinash into Google.

Andrew: Actually type A-V-I-N-A-SH. Avinash, the way that it’s pronounced. You just type it into Google. I am telling you Noah Kagan said that once he got serious about selling online, he started following your blog and pulling information from it. I spent all morning going through it and now I understand why. Check out the website, read it and I’m sure you’ll understand why too. Avinash, thank you again for doing the interview.

Avinash: Thank you Andrew, it was a pleasure.

Andrew: Cool. Thank you all for watching.

This transcript brought to you by www.SpeechPad.com.

Related links from Avinash

1. The image he referenced in the interview when I asked him how we can choose what tools to focus on.

2. Six Tips For Improving High Bounce / Low Conversion Web Pages

3. The Definitive Guide To (8) Competitive Intelligence Data Sources

Sponsors I mentioned

Shopify – Remember the interview I did about how the founder of DODOCase sold about $1 mil worth of iPad cases in a few months? He used Shopify. It’s dead simple and very effective.

Grasshopper – Don’t make the mistake of comparing Grasshopper with other phone services. Check out their features and you’ll see why Grasshopper isn’t just a phone number, it’s the virtual phone system that entrepreneurs (like me) love.

Walker Corporate Law – Scott Edward Walker is the lawyer entrepreneurs turn to when they want to raise money or sell their companies, but if you’re just getting started, his firm will help you launch properly. Watch this video to learn about him.

Share

  • Neowebstudios

    the video is extremly slow. It stops all the time. My Internet connection is good. You should change your hoster. The video is great btw!

  • http://dailybrite.com Madhav SBSS

    Very insightful interview on using Analytics, thanks Andrew and Avinash.

    I would love to hear any viewpoints on Analytics specific to initial Customer Discovery, Customer Development phases. Thanks.

  • http://mixergy.com Andrew Warner

    Sorry about that. I’ll check it out.

  • http://mixergy.com Andrew Warner

    Good reason to have him back on. Thanks for the suggestion.

  • http://twitter.com/buzzdan Dan

    Killer guest and content :>

  • http://twitter.com/daysonp Dayson

    I’ve only reached half way through this interview but I couldn’t resist my urge to comment.
    Very interesting information delivered by a very enthusiastic person through a very good interviewer.

  • http://mixergy.com Andrew Warner

    That’s what I was thinking too as I heard him talk.

  • http://mixergy.com Andrew Warner

    Thanks for the compliment Dayson!

    When a guest is this good, it makes interviewing so much easier.

  • Kris

    Great guest! Great content! Thank you.

  • Nath

    Interesting topic & mindset..

  • http://www.wesedholm.com Wes

    I like how he uses the videos on his competition… a very out of the box idea.

    I also checked out kiss insights several times but never tried it cause was turned off by the questions. Nice to know they have sub options.

    I did finally sign up for grasshopper after we had some problems with our other voicemail to email solutions. Just shows you how long a sale is disconnected from a purchase. First heard of grasshopper here on an interview… then ended up purchasing several months later from a different computer through appsumo. Try to measure that with a click stream!

  • Robyn MacDonald

    Such clear and intelligent insights. Thanks Avinash !

  • Fabian

    This was one great interview. Thanks for doing this Andrew!

  • http://twitter.com/melvinram melvinram

    Lots of very useful information. Thank you Andrew and Avinash.

  • Pingback: You’ve installed Google Analytics: So what? :: James Kennedy()

  • Pingback: Trends Erkennen – Seba's Blog()