Google AdSense, AdWords and a brave new approach to marketing
Google's AdWords/AdSense program represents a great leap forward in ad delivery, but they're not the real pioneers, and they're not even in the best position to deliver targeted advertising to consumers.
This CS Monitor article profiles people in India and the Phillipenes making up to $1000/month with Google's AdSense on their blogs. That's a full-time programmer's salary for one guy in India, and allowed one article subject to return to school. It's a compelling and heart-warming story, and it underscores the billions of dollars Google is raking in with their ad program.
Google has a great product with AdWords/AdSense, and they're slightly ahead of the marketing industry as a whole. They're able to deliver ads people want to see on a much more consistent level than a broadcaster. Google also delivers analytics which give a detailed breakdown of how your ads are peforming individually, and against various keywords, as well as the ability to track ad clicks from the point at which the ad is selected to the point at which users complete a transaction.
It's a killer combination: I was recently emailing a friend about a problem I was having with JBoss, a finicky open source Java application server. When I read a reply to my email in GMail, an AdWords ad appeared to the right of the message telling me about a JBoss configuration tool which might help with my problem.
I haven't bought the product yet, but for that advertiser, it's a great success: for a few dollars they were able to get the ad in front of someone who was deeply interested in what they sell, to the exclusion of others. Try that on TV. Google has opened web advertising to small players -- someone with $50/month to spend or less, a non-existant budgets by industry standards.
An old marketing saw goes something like "I'm wasting half of my budget, but I don't know which half". When you can track an ad from click to checkout, it's a brave new world of advertising, and you can know exactly which half goes to waste. You can perform statistical analysis on ads, who the ads are shown to, when, and develop an excellent idea of how many dollars it costs to acquire a paying customer. In short, it's a scientific approach.
The marketing industry as a whole has been taking a radical turn towards the scientific in the past 10 years or so, and online advertising hasn't been the sole market-driver.
CapitolOne, the credit card issuer, is frequenly cited as one of the earliest and most successful adopters of this approach. They perform A/B tests on things as detailed as envelope color for mailed solicitations, and leverage artificial intelligence in their call centers to determine not only why you're calling, but what they want to do about it: if they suspect you're about to cancel your card, computers decide whether to route you to a customer retention specialist, or an automated cancellation routine.
The ability to test envelope colors makes Google look rinkydink. Google's targeting is largely limited to keywords now: their newly-released demographic targeting is limited to a small number of categories and deployed on low-traffic sites. But other companies are about to take this much further.
During my day job I work with a cellphone carrier who is applying the same approach to people who browse the web on mobile phones. Your cell carrier knows a lot more about you than Google: birthdate, ssn, credit rating, household income, number of kilobytes downloaded per month, among other things. When they combine the data you provide when buying a cellphone with commercially available profiles, what they know becomes creepy.
Carrier X is taking all this data, and combining it with your browsing and purchase history to figure out what ads to show you. They're sensitive to the 'creepiness factor', as they call it, but at the same time it's potentially even more effective than Google's demographic targeting due to the detail and richness of the data they're able to feed into the targeting system.
In any information processing system the quality of input data correlates directly to the quality of output: the success of ad targeting so far will effectively suck in more and more detailed data, simply because it works so well. Advertisers like it, and I as a consumer like seeing ads I'm interested in.
Will people freak out about how much is known about them? Once the ads start crossing a 'creepy' threshold, some people will for sure. For those people, there's a suite of privacy tools available, and hopefully marketed through a highly-targeted ad campaign.
(link via robotwisdom)
This CS Monitor article profiles people in India and the Phillipenes making up to $1000/month with Google's AdSense on their blogs. That's a full-time programmer's salary for one guy in India, and allowed one article subject to return to school. It's a compelling and heart-warming story, and it underscores the billions of dollars Google is raking in with their ad program.
Google has a great product with AdWords/AdSense, and they're slightly ahead of the marketing industry as a whole. They're able to deliver ads people want to see on a much more consistent level than a broadcaster. Google also delivers analytics which give a detailed breakdown of how your ads are peforming individually, and against various keywords, as well as the ability to track ad clicks from the point at which the ad is selected to the point at which users complete a transaction.
It's a killer combination: I was recently emailing a friend about a problem I was having with JBoss, a finicky open source Java application server. When I read a reply to my email in GMail, an AdWords ad appeared to the right of the message telling me about a JBoss configuration tool which might help with my problem.
I haven't bought the product yet, but for that advertiser, it's a great success: for a few dollars they were able to get the ad in front of someone who was deeply interested in what they sell, to the exclusion of others. Try that on TV. Google has opened web advertising to small players -- someone with $50/month to spend or less, a non-existant budgets by industry standards.
An old marketing saw goes something like "I'm wasting half of my budget, but I don't know which half". When you can track an ad from click to checkout, it's a brave new world of advertising, and you can know exactly which half goes to waste. You can perform statistical analysis on ads, who the ads are shown to, when, and develop an excellent idea of how many dollars it costs to acquire a paying customer. In short, it's a scientific approach.
The marketing industry as a whole has been taking a radical turn towards the scientific in the past 10 years or so, and online advertising hasn't been the sole market-driver.
CapitolOne, the credit card issuer, is frequenly cited as one of the earliest and most successful adopters of this approach. They perform A/B tests on things as detailed as envelope color for mailed solicitations, and leverage artificial intelligence in their call centers to determine not only why you're calling, but what they want to do about it: if they suspect you're about to cancel your card, computers decide whether to route you to a customer retention specialist, or an automated cancellation routine.
The ability to test envelope colors makes Google look rinkydink. Google's targeting is largely limited to keywords now: their newly-released demographic targeting is limited to a small number of categories and deployed on low-traffic sites. But other companies are about to take this much further.
During my day job I work with a cellphone carrier who is applying the same approach to people who browse the web on mobile phones. Your cell carrier knows a lot more about you than Google: birthdate, ssn, credit rating, household income, number of kilobytes downloaded per month, among other things. When they combine the data you provide when buying a cellphone with commercially available profiles, what they know becomes creepy.
Carrier X is taking all this data, and combining it with your browsing and purchase history to figure out what ads to show you. They're sensitive to the 'creepiness factor', as they call it, but at the same time it's potentially even more effective than Google's demographic targeting due to the detail and richness of the data they're able to feed into the targeting system.
In any information processing system the quality of input data correlates directly to the quality of output: the success of ad targeting so far will effectively suck in more and more detailed data, simply because it works so well. Advertisers like it, and I as a consumer like seeing ads I'm interested in.
Will people freak out about how much is known about them? Once the ads start crossing a 'creepy' threshold, some people will for sure. For those people, there's a suite of privacy tools available, and hopefully marketed through a highly-targeted ad campaign.
(link via robotwisdom)
posted by $!# at 4/03/2006 08:24:00 PM
0 Comments:
Post a Comment
<< Home