Voices » Jeff Stibel » Web Marketing's Darwinian Secrets
11:28 AM Tuesday August 12, 2008
For the last 50 years or so, marketing has been a one-trick pony. Olgilvy, Mather, Young, Rubicam, Saatchi (and Saatchi), and all the rest built their reputations on having great advertising campaigns.
But that model breaks down on the Internet. The Internet provides a greater wealth of feedback, more inventory than other media, and response times measured in milliseconds, as opposed to days.
Marketing online has more to do with economics (Schumpeter would be proud) than it does with advertising. Creative destruction--where one thing is replaced by another, often at a frenzied pace--is how online advertising works best.
The brain, of course, also creatively destructs. You see this best with short-term memories, as new ideas pop into your head only to wipe away the old.
In economics, online advertising and the brain, only the fittest survive.
One of my companies--The Search Agency--has become one of the largest online advertising firms by following the lessons of the brain. With over 100 people, we have few--if any--traditional marketers, not even a Chief Marketing Officer.
The reason for this is because a marketing background can often be counterproductive in the online world of fast media. With the ability to track changes in real-time, you no longer need to think through the best way to advertise; instead, you throw everything you've got into the universe and see what works. It is here that you see creative destruction at its fastest, most frenzied pace.
For a typical search marketing campaign on Google--for say a Fortune 500 company or a top 10 website--a Web marketing firm may be actively testing upwards of 10,000 keywords, 1,000 different pieces of copy, hundreds of pictures and countless variations of websites. On top of that, they may have two dozen price points and a baker's dozen of products and packages in rotation. Campaigns are measured in performance (determined by sales, visits or other metrics the client determines as important) across multiple dimensions: time (to the second), geography, demographics, user profiles, and whatever additional data we can gather.
The complexities of the campaigns are immense but the model is simple. Imagine a campaign for Ford, testing every variation of search words (car, cars, auto, Ford Truck, GM Trucks--yes, competitor's keywords tend to perform well), every different combination of creative (buy Ford, buy Ford Trucks, Ford Trucks are cool, buy Ford Truckes--yes, misspellings often perform well). And then you see what works best and when: Ford Trucks performs best on the keywords car and auto between 8-10 PM for men in Green Bay; worse than Ford Truckes between 4-4:30 PM in Green Bay but equally well in Seattle for women during that time.
Imagine hundreds of thousands of variables, thousands of ad campaigns, all competing with one another to survive and flourish---just for one Ford Truck. Each set of campaigns is a new generation, pushing forward one step at a time. Yet it goes so fast, we often don't even know what or why it works.
The Internet ad campaign is fast becoming the fruit fly's closest evolutionary cousin.
How do we know what campaign is best? The same way that the fruit fly evolves: we don't know and we don't care (well, we find out using the wisdom of hindsight). Instead, we let the creative destruction algorithms create a model of survival of the fittest, where the best campaigns live to fight another day and the worst are killed.
Sound a bit maddening? It is. But that's creative destruction for you.
No brands, no positioning, no marketing--just creative destruction at the speed of the Internet.
TrackBack URL for this entry:
http://blogs.harvardbusiness.org/cgi-bin/mt/mt-tb.cgi/2647
No trackbacks have been made to this entry.
Posting Guidelines
We hope the conversations that take place on HarvardBusiness.org will be energetic, constructive, free-wheeling, and provocative. To make sure we all stay on-topic, all posts will be reviewed by our editors and may be edited for clarity, length, and relevance.
We ask that you adhere to the following guidelines.

Jeffrey M. Stibel is an entrepreneur and brain scientist. He studied business and brain science at MIT Sloan and Brown University, where he was a brain and behavior fellow. Stibel has authored numerous academic and business articles on a variety of subjects and is the named inventor on the US patent for search engine interfaces. He is currently President of Web.com (NASDAQ: WWWW) and serves on academic Boards for Tufts and Brown University, as well as the Board of Directors for a number of public and private companies. Stibel is the author of Wired for Thought: How the Brain Is Shaping the Future of the Internet, being published by Harvard Business Press in September 2009.
ADVERTISEMENT
Michael Jackson and the Zombieconomy Umair Haque
How Michael Jackson Became a Brand Icon John Quelch
Debunking Social Media Myths David Armano
A Good Way to Change a Corporate Culture Peter Bregman
Great Communicators Are Great Explainers John Baldoni
Debunking Social Media Myths David Armano
Michael Jackson and the Zombieconomy Umair Haque
How Michael Jackson Became a Brand Icon John Quelch
How to Identify Your Employees' Hidden Talents Steven DeMaio
Why Microsoft Had to Destroy Word Peter Merholz
This simulation will help you learn how to craft conversations that are fact based, minimize defensiveness, and draw out the best thinking from everyone involved.
In many organizations, marketing exists far from the executive suite and the boardroom. Learn how to improve the link between high level corporate strategy and the marketing function.
ADVERTISEMENT
Comments
Well Jeffrey,
I do not really with you say.
You cannot just throw everything in the air and see what flyes...That costs a lot.
I think that like for traditionnal Marketing you have to study what you want to do with your product/service, define goals, mesure performance, organise work.
If not you are just waiting for random success.
Monsieur J
- Posted by Monsieur J
August 13, 2008 7:56 AM
Gee Jeffery, I wonder what causes the person to be searching for information on a Ford Truck to begin with? Oh, let's see, could it be all that other communication in all kinds of media outside your limited example of search?
I think it all works together and you need a full range of approaches; including, but not limited to, search. We've seen activity on search and other online media increase significantly when complimentary schedules of traditional media were running simultaneously.
Maybe if you had some people with traditional marketing experience on your staff, you'd understand how communication and persuasion actually do work.
Bill
- Posted by bill
August 13, 2008 2:00 PM
Great post and right on the money. Consider that for less than the cost of single minute of prime time TV advertising, any one of the major hotel operators (or their high paid agencies) could have bought the domain name hotels.com. Instead, IAC owns it and for the next 50 years Hilton, Marriott and the rest will be paying IAC for leads and bookings when they could have gotten them for almost nothing.
Instead they are still focused on branding and have entirely missed out on the true potential of the Net; having as many front doors to your business as possible and testing a million variables until the control is as close to a perfect ROI machine as possible.
Truly its time for marketers to retrain and rethink if they want to help clients succeed in what may be the best media for marketing every invented.
Sadly it seems that Academia and the big brands are mostly missing out on this massive opportunity due to lack of creative intelligence and an ability to figure out how Internet Marketing, Search Engine Optimization, CPM, PPC and CPA really work.
- Posted by domainpubber
August 13, 2008 2:54 PM
Well said Jeff. Touch and feel media will never go away completely, but digital media will continue to consume market share over the next 10-years in particular.
Search is not just about Google or Yahoo! It's a way of life for most Americans. It shapes brands, acts as the primary funneling device on the Internet, and is responsible for the development and support of hundreds of thousands of businesses across the nation. No other medium has made such a profound impact on our great country. God bless Search.
- Posted by Joe Griffin
August 14, 2008 12:17 PM
Great comments and a healthy debate. My thoughts:
To be sure, there is a place for traditional media, brand building, market research and focus groups. But on the Internet there is huge value in the experimental method of quickly testing a null hypothesis, proving it wrong, and moving on to bigger and better campaigns.
- Posted by Jeff Stibel
August 14, 2008 1:14 PM
So, Jeff, you believe in the "braille System" of prospecting?
That is, as you said, throwing as much S* against the wall as you can, and seeing what sticks, and what smells.
What a system!
You are correct, you don't need a marketing person, because they might have brains to formulate objections to this hare-brained idea. It may work because of the nature of the Internet today, and the way Google is structured, and others who are looking for "eyeball" payments.
That may even be a good thing.
But in the next evolution of the Internet, coming soon to a screen near you, in your bedroom, car, office, home office, mobil screen,plane, train and in Africa, Sahara Desert, Antarctic, wherever and whenever, is a whole new world.
Anything, Anytime, Anywhere-remember that. That is what will drive the Content eyeballs; that is what will drive the next generation of "networks."
Oh, and in that context, do you really think the "Braille" system will work?
Can it be you really believe you are doing your clients a service by divorcing intelligent planning from the branding process?
What about the 20% of the market that you can't reach that way? Are they worthless?
Or is it the money?
- Posted by Mediaman
August 15, 2008 2:44 PM
Your premise is flawed. The firms you list (Ogilvy, etc) are not marketing firms but rather advertising agencies. The marketers that hired them would never concede that marketing is only about great advertising. Advertising is only one (rather small) piece of the marketing equation.
Replace the word "marketing" in your article with "online advertising" and your argument is at least coherent, if not naive.
- Posted by Nick
August 18, 2008 3:46 PM
So, this reminds me of the story of the two Samurais, the young novice and old pro Samurai. The novice goes into a fight full of enthusiasm, lashing out at anything and everything that moves. The old pro Samurai selects his target carefully and figures out the minimum cuts it will take to get the job done. Then the old Samurai takes the money and goes back to his farm to play with his grand children.
- Posted by Vladymir Rogov
August 19, 2008 6:12 PM
Great post Jeff andyou couldn't be more accurate.
And to the critics, haven't you ever heard of the scientific method and proving a null hypothesis. This is no different.
And whomever thinks it is too expensive is nuts. This is the cheap way to market online. You can get a large sample for a few hundred dollars and let the market determine the best campaign.
That is good market research, and always better than a moron who calls himself a marketer, sitting in his corner office saying things like "I know best" and "no need to test" and all the other BS you here whenever marketers are challenged.
Keep up the great posts.
- Posted by Sam
August 20, 2008 10:40 AM
Great one Jeff. It is dead on point but few marketers really get it yet.
To the Samurai: This reminds me of Indiana Jones watching the expert swodsman wield a sword over and over in expert fashion. Jones took out his gun and shot him.
Expertise just gets in the way when there is a cheaper, better or more efficient method. Online marketing is closer to direct response media than advertising and marketing.
- Posted by DE Shaw
August 20, 2008 10:44 AM
Jeff, thanks for the post. I do think, however, that what you are really implying is that only with multi-variate analyses can the patterns of the internet be exploited. There are just too many variables to account for.
I firmly believe that only the aggregate amount of viewing time is the true measurement of consumption. X eyesballs per minute.
Niraj
- Posted by Niraj Patel
August 21, 2008 11:24 PM
Jeff,
Interesting article and subsequent debate. Where I think people are losing you is that they're equating your "evolutionary" approach with a see what sticks approach. People... that's not it at all.
What amazes me about most marketers is that we can spend so much time on research, on picking the right terms, etc. and then put complete trust into what the networks and ad servers say they delivered and then focus our efforts on deconstructing the campaign to satisfy the human need for reassurance that the right decision was made. In Jeff's approach, and correct me if I'm wrong Jeff, you don't worry about that stuff as much and you let the multivariate testing sort things out and base success on what happened and not focus as much on why it happened.
If you get a little more skeptical about the world of online advertising and instead track the ads back through your entire site, rather than just landed at your front door or resulted in a sale, I think you’ll see what Jeff is saying is very true. There will be very clear winners for traffic drivers, interaction (engagement) drivers, and revenue drivers. Pop these groups into a Venn diagram and look at the intersection if that helps. This is the "survival of the fittest" – the group of creatives that likely provides the most overall bang for your buck (or highest level of cost/benefit optimization if you’re more inclined to think in those terms). Base your next round of MVT on them and move on without focusing on the individual campaign attributes. It’s just a stepping stone to the next iteration of an infinite process.
Keep 'em coming, Jeff.
Greg
- Posted by Greg Asman
August 29, 2008 8:31 AM
You're spot on about traditional marketing people. I had the surreal experience of going from an online direct marketing company that at the time was doing $150M/yr without a single traditional marketing person to an online SAAS company filled to gills w/ traditional marketing background.
The dynamics were amazingly destructive. Everyone (repeat everyone) had an opinion about how to do online and none of those opinions were based on experience. So some people would go for anything, others wouldn't go for anything they didn't understand. And since they didn't understand online, they took shots at every idea or initiative. It made for a highly ineffective, borderline schizophrenic situation.
- Posted by Michael G
November 12, 2008 2:30 PM