Jim Novo wrote a Web Analytics Association Review on Firm Created Word Of Mouth.
I strongly recommend the read.
Although the paper was published in the most recent edition of Marketing Science, it was based on findings that span four decades.
The first finding reaffirms the 'strength of weak links' hypothesis. Let me explain:
Like people tend to clump, alike.
Among my friends, more than half own iPhones with occupations centering on technology and the Internet and most have roles that are heavily steeped in data. Three quarters would be classified by Forrester as being Tech Optimists and Creators. A majority live in the inner city.
Not everybody in my circle are uniformly this way: I used the word 'more than half' for a reason, but you get the picture: like people clump alike.
I'm linked to a number of other communities by way of acquaintances. Lots and lots of acquaintances! Everybody is.
Very few in my circle of friends would ask me if I would recommend Google Analytics over Omniture. They already know enough to have their own opinion.
However: an acquaintance who runs a company might ask me how they would know if their money is well spent - and my response: that recommendation - one that travels along a weak link, would carry more weight.
A weak link is one between acquaintances that span two heterogeneous communities. A strong link is one that spans between friends in a homogeneous community. A reccomendation made over a weak link is stronger than one carried over a strong link.
We use the words 'forward to a friend' all the time in social. In all reality, forward to a friend isn't what you want to drive product adoption. To drive frequency of purchase - yes. But in most cases (there are exceptions!) not product adoption. The 'forward to an acquaintance' action is the one you really want to happen.
In spite of this, we as marketers continue to use the term 'forward to a friend' because that's the base call to action. This entire weak ties is a footnote.
In all, Jim Novo's review is a good read. It's valuable.
Sunday, September 27, 2009
Sunday, September 20, 2009
Conference Ecosystem
I had the good enough fortune to talk with Stephane Hamel, a director of the Web Analytics Association, and Andrea Hadley of eMetrics last Friday while at IMC in Vancouver.
As usual with any conference - the really interesting conversations happen in the lobby during the day.
Andrea, being the super-networker she is, got me into talking with Stephane about the Research Committee, and fast tracking was to be had. We also talked about the diverse audiences involved in any industry, and how to try to serve each group really well at a conference.
There are experts, newcommers, and vendors/consultants. Vendors want to sell, newcommers want to learn, and experts want to talk to each other and recruit talent. eMetrics is experimenting with different formats to serve all three groups. In general, vendors don't really care if the experts screw off of to the lobby. They know Vendor X really well and they're not the target. Newcommers are the target. Most newcommers are looking for solutions to the new problems they've just been handed (typically a web analytics login and password!).
In the end, I think all three audiences need to be served by the same set of conferences. There should always be a vendor component and a welcoming environment for the newcommers. eMetrics might want to consider inviting cross-over experts in other fields to share practices and try to get the water as brackish as possible.
As usual with any conference - the really interesting conversations happen in the lobby during the day.
Andrea, being the super-networker she is, got me into talking with Stephane about the Research Committee, and fast tracking was to be had. We also talked about the diverse audiences involved in any industry, and how to try to serve each group really well at a conference.
There are experts, newcommers, and vendors/consultants. Vendors want to sell, newcommers want to learn, and experts want to talk to each other and recruit talent. eMetrics is experimenting with different formats to serve all three groups. In general, vendors don't really care if the experts screw off of to the lobby. They know Vendor X really well and they're not the target. Newcommers are the target. Most newcommers are looking for solutions to the new problems they've just been handed (typically a web analytics login and password!).
In the end, I think all three audiences need to be served by the same set of conferences. There should always be a vendor component and a welcoming environment for the newcommers. eMetrics might want to consider inviting cross-over experts in other fields to share practices and try to get the water as brackish as possible.
Wednesday, September 9, 2009
Fraught with Skepticism
I'll confess that one of my favourite lolcats is Skeptical Cat is Fraught with Skepticism.
Look deep into that expression. The cat really does look skeptical, doesn't he? He's not believing a single word you're thinking right now.
There's also something about that orange background that makes the expression and the entire image that much funnier. I don't know what it is about it. But I'm aware of the effect. I think anthropologists have a term for the tendency of humans to superimpose human emotions onto animals - which there is no evidence that an animal actually feels. I can't remember the term, but it's funny as hell that we all do it.
The reason I bring this is all up is that I deal with skepticism all the time in my line of work as a web analyst and marketing scientist. Sometimes the data disproves long held assumptions about aggregate user behavior and attitudes - which causes skepticism. This rejection of your own hypothesis is something a scientist isn't troubled about. For non-scientists it can cause all sorts of personal problems. Yet, even with empirical proof, skepticism can persist. They don't say it out loud, but you can see it in their faces. They're a skeptical cat. And they're fraught. With skepticism.
I'm very lucky in my career so far. I manage to manage a team of people and multiple programs while still being a practitioner. I still use statistics to make sense of data. When I'm talking to the consumers of the insights and findings that come out of this work, I don't talk about regressions and the origins - I translate the results. I make the results actionable and the insights relevant to their interests.
The consumers of these insights don't need to understand the nitty gritty of sampling error or SPSS syntax. Could you imagine trying to explain all that? "It all started off with a gambler in the 1600's.... and then Laplace....and then just before prohibition they wanted to put statistics on the table and scientific advertising....". Ha!
So it goes with this NextStage skepticism.
Some of the greatest inventions in the world are discovered by combining two unusual technologies and deleting anything that is unnecessary. Invention typically requires strong lateral thinking.
Two years ago I was deeply troubled by the "optimization cycle problem" of web analytics, and reckoned that if physical technology got us into this, it could get us out. It led down a number of wonderful research rat holes and inquiries.
I got introduced to Joseph Carrabis by way of June Li at eMetrics Toronto in March 2008. What prompted the introduction? I was inquiring about behavioral economics (there's a sub-field called "evolutionary programming or collective intelligence" - to be precise) as applied to website morphing (which, at the time I was referring to a vague Dutch test of the technology to a medical website. The test failed because they were morphing the nav bar. Website Morphing has since been updated by Hauser et al the next year.). I was trying to learn how to combine these two technologies. June Li figured she had to introduce me to Joseph.
Joseph announced his patent the next day.
Our second conversation pretty much extinguished any lingering skepticism I had. (And the third conversation was probably more unpleasant for everybody involved as I expressed concern about the usage of the technology.). So what happened between the first conversation, the second, and the third?
First, a herd of intelligent people were not skeptical. He has social validity.
Second, if he couldn't think laterally - odds were that he was a phony. He was a strong lateral thinker and smart to boot. He has personal validity.
Third, he knew things about EP that nobody who was faking it could have known off the cuff. Only somebody who had experienced the data and thought very hard about the problems could have known about it. He answered these needling questions in a manner that had very, very strong face validity.
Fourth, the patent. This is very strong external validity.
NextStage technology incorporates multiple areas and technologies while deleting the extraneous bits. There's neuroscience, anthropology, statistics, language psychology, behavioral economics (EP/CI) and computer science (specifically: computability). It blends together nicely because it is informed by a central theory that is explanatory and powerful.
Since then, much of what Joseph has demonstrated over the past 18 months meshes with what hasn't been published (yet - papers have been presented) in the field of decision neuroscience - an emerging sub-discipline in Marketing Science. I have no reason to challenge the veracity of NextStage's claims because every esoteric challenge has always been met. Their claims have been independently verified and I consider the matter closed. They really invented a new technology. Yes, it can happen.
Back to skepticism.
I don't believe that 90% of the market researchers and 70% of the data miners I've met have ever read or understand Laplace - and yet that doesn't keep any of those people from using statistics to make judgments about vast numbers of people. Most practitioners don't have degrees in mathematical science. Are any of these people who practice linear regression on a daily basis ever skeptical of the science that underlines it?
There's a baseline amount of verification that people need to do on their own. A smell check or a sniff test. Everybody will need their own source of validity. I have mine.
In sum:
You don't have to understand how the cow turns grass into milk to drink the milk.
You don't have to understand how the cat makes that purring sound to enjoy it.
You don't have to understand the mechanism in your mind that makes skeptical cat look skeptical to laugh at the joke or to understand NextStage. Skeptical cat looks skeptical and NextStage technology works. It's as simple as that.
Look deep into that expression. The cat really does look skeptical, doesn't he? He's not believing a single word you're thinking right now.
There's also something about that orange background that makes the expression and the entire image that much funnier. I don't know what it is about it. But I'm aware of the effect. I think anthropologists have a term for the tendency of humans to superimpose human emotions onto animals - which there is no evidence that an animal actually feels. I can't remember the term, but it's funny as hell that we all do it.
The reason I bring this is all up is that I deal with skepticism all the time in my line of work as a web analyst and marketing scientist. Sometimes the data disproves long held assumptions about aggregate user behavior and attitudes - which causes skepticism. This rejection of your own hypothesis is something a scientist isn't troubled about. For non-scientists it can cause all sorts of personal problems. Yet, even with empirical proof, skepticism can persist. They don't say it out loud, but you can see it in their faces. They're a skeptical cat. And they're fraught. With skepticism.
I'm very lucky in my career so far. I manage to manage a team of people and multiple programs while still being a practitioner. I still use statistics to make sense of data. When I'm talking to the consumers of the insights and findings that come out of this work, I don't talk about regressions and the origins - I translate the results. I make the results actionable and the insights relevant to their interests.
The consumers of these insights don't need to understand the nitty gritty of sampling error or SPSS syntax. Could you imagine trying to explain all that? "It all started off with a gambler in the 1600's.... and then Laplace....and then just before prohibition they wanted to put statistics on the table and scientific advertising....". Ha!
So it goes with this NextStage skepticism.
Some of the greatest inventions in the world are discovered by combining two unusual technologies and deleting anything that is unnecessary. Invention typically requires strong lateral thinking.
Two years ago I was deeply troubled by the "optimization cycle problem" of web analytics, and reckoned that if physical technology got us into this, it could get us out. It led down a number of wonderful research rat holes and inquiries.
I got introduced to Joseph Carrabis by way of June Li at eMetrics Toronto in March 2008. What prompted the introduction? I was inquiring about behavioral economics (there's a sub-field called "evolutionary programming or collective intelligence" - to be precise) as applied to website morphing (which, at the time I was referring to a vague Dutch test of the technology to a medical website. The test failed because they were morphing the nav bar. Website Morphing has since been updated by Hauser et al the next year.). I was trying to learn how to combine these two technologies. June Li figured she had to introduce me to Joseph.
Joseph announced his patent the next day.
Our second conversation pretty much extinguished any lingering skepticism I had. (And the third conversation was probably more unpleasant for everybody involved as I expressed concern about the usage of the technology.). So what happened between the first conversation, the second, and the third?
First, a herd of intelligent people were not skeptical. He has social validity.
Second, if he couldn't think laterally - odds were that he was a phony. He was a strong lateral thinker and smart to boot. He has personal validity.
Third, he knew things about EP that nobody who was faking it could have known off the cuff. Only somebody who had experienced the data and thought very hard about the problems could have known about it. He answered these needling questions in a manner that had very, very strong face validity.
Fourth, the patent. This is very strong external validity.
NextStage technology incorporates multiple areas and technologies while deleting the extraneous bits. There's neuroscience, anthropology, statistics, language psychology, behavioral economics (EP/CI) and computer science (specifically: computability). It blends together nicely because it is informed by a central theory that is explanatory and powerful.
Since then, much of what Joseph has demonstrated over the past 18 months meshes with what hasn't been published (yet - papers have been presented) in the field of decision neuroscience - an emerging sub-discipline in Marketing Science. I have no reason to challenge the veracity of NextStage's claims because every esoteric challenge has always been met. Their claims have been independently verified and I consider the matter closed. They really invented a new technology. Yes, it can happen.
Back to skepticism.
I don't believe that 90% of the market researchers and 70% of the data miners I've met have ever read or understand Laplace - and yet that doesn't keep any of those people from using statistics to make judgments about vast numbers of people. Most practitioners don't have degrees in mathematical science. Are any of these people who practice linear regression on a daily basis ever skeptical of the science that underlines it?
There's a baseline amount of verification that people need to do on their own. A smell check or a sniff test. Everybody will need their own source of validity. I have mine.
In sum:
You don't have to understand how the cow turns grass into milk to drink the milk.
You don't have to understand how the cat makes that purring sound to enjoy it.
You don't have to understand the mechanism in your mind that makes skeptical cat look skeptical to laugh at the joke or to understand NextStage. Skeptical cat looks skeptical and NextStage technology works. It's as simple as that.
Tuesday, September 1, 2009
You don't need to understand it to use it
At the Marketing Science Conference earlier in the summer, Shaina and I took in the Neuromarketing session. The session was very good, with 3 really great presenters out of the 4.
I learned several important reasons why people make the choices they do. For instance, I saw it empirically proven that self-control is like a muscle: you can hold a certain pose for so long, and then that muscle gets fatiqued and weak. Then you can't hold it for any longer and you break that pose. It's a very attractive causal variable for periodic consumption and lapses in self-control. Prospection - the ability to think of the future - when combined with anchor-and-adjust tendencies, cause discount-rate curves to deviate from what classical-economists would predict. This has important implications for trial-bonus marketing.
So many breakthroughs in behavioral economics and marketing science are originating from neuroscience. It's exciting.
Meanwhile, the great people over at Next Stage Evolution are making advancements in making neuromarketing accessible. You don't need a degree in neuroscience to use their technology to make things better and get benefits.
Take for instance, web analytics. Right now, in a majority of fortune 500 and fortune 5,000,000 companies, managers and web analysts are making decisions on how to improve their site based on numbers most of them assume they really understand. In fact, it's a relatively recent development that the Web Analytics Association has worked with vendors to really define what most of those numbers really mean. Among the most misunderstood include "unique visitor", "time spent on site", and the actual definition of what a "bounce rate" really is. It doesn't matter. They don't have to understand to know that more unique visitors is 'good', time spent on site is 'interesting', and a high 'bounce rate' is generally 'bad' - so long as there's some baseline education and comfort.
You don't have to understand how an airplane flies to be a passenger on one, no more than you need to know how a light bulb really works to derive benefit from it.
There's always going to be skepticism and fear whenever a new technology comes about and it starts to be adopted. Electricity and soap were once feared. Flying was too. We're starting to see some of that around web analytics this year.
That's not to say that people who aren't curious about how neuromarketing is done shouldn't explore and ask. The curious should.
So, when I assert that you should offer visitors with browsing pattern X an offer of $50 paid out in 3 months if they sign up now and visitors with browsing pattern Y an offer of $20 instantly if they sign up now, you might challenge that. Good. Then I'll explain for 15 minutes about the hyperbolic discount rate curve and prospection tendencies of different browse paths. Then, if I've done my job correctly, you'll trust the science just as you would trust a pilot or trust a light bulb. Abstraction can be a wonderful thing. Getting there is a longer process.
You don't need to understand it all to use it.
I learned several important reasons why people make the choices they do. For instance, I saw it empirically proven that self-control is like a muscle: you can hold a certain pose for so long, and then that muscle gets fatiqued and weak. Then you can't hold it for any longer and you break that pose. It's a very attractive causal variable for periodic consumption and lapses in self-control. Prospection - the ability to think of the future - when combined with anchor-and-adjust tendencies, cause discount-rate curves to deviate from what classical-economists would predict. This has important implications for trial-bonus marketing.
So many breakthroughs in behavioral economics and marketing science are originating from neuroscience. It's exciting.
Meanwhile, the great people over at Next Stage Evolution are making advancements in making neuromarketing accessible. You don't need a degree in neuroscience to use their technology to make things better and get benefits.
Take for instance, web analytics. Right now, in a majority of fortune 500 and fortune 5,000,000 companies, managers and web analysts are making decisions on how to improve their site based on numbers most of them assume they really understand. In fact, it's a relatively recent development that the Web Analytics Association has worked with vendors to really define what most of those numbers really mean. Among the most misunderstood include "unique visitor", "time spent on site", and the actual definition of what a "bounce rate" really is. It doesn't matter. They don't have to understand to know that more unique visitors is 'good', time spent on site is 'interesting', and a high 'bounce rate' is generally 'bad' - so long as there's some baseline education and comfort.
You don't have to understand how an airplane flies to be a passenger on one, no more than you need to know how a light bulb really works to derive benefit from it.
There's always going to be skepticism and fear whenever a new technology comes about and it starts to be adopted. Electricity and soap were once feared. Flying was too. We're starting to see some of that around web analytics this year.
That's not to say that people who aren't curious about how neuromarketing is done shouldn't explore and ask. The curious should.
So, when I assert that you should offer visitors with browsing pattern X an offer of $50 paid out in 3 months if they sign up now and visitors with browsing pattern Y an offer of $20 instantly if they sign up now, you might challenge that. Good. Then I'll explain for 15 minutes about the hyperbolic discount rate curve and prospection tendencies of different browse paths. Then, if I've done my job correctly, you'll trust the science just as you would trust a pilot or trust a light bulb. Abstraction can be a wonderful thing. Getting there is a longer process.
You don't need to understand it all to use it.
Subscribe to:
Posts (Atom)