Contextual Advertising – The Cure for Database Poisoning

Whoever has the misfortune of the email address bob@bob.com definitely hates me.  If he exists, I’d like to apologize to him for the hundreds of sites that I’ve avoided providing my real details to by usurping his good name.  Of course, I also have my own email address set up just to register on sites where I need to check the email (for passwords, etc.) too.  In this way, I am just like any other consumer on the Internet.  When presented with a registration form with no distinct payoff or reason for filling it out – I am likely to make up or fake as many details as I can get away with in order to avoid being contacted or make the registration process faster for myself.  An article from Pamela Parker in ClickZ yesterday discussed the concept of "database poisoning" (where users enter fake registration details) – and used the site www.bugmenot.com as an example of a consumer driven eAdvocacy campaign to influence publishers to get rid of these annoying registration forms.  The site offers user-contributed passwords to bypass registration forms for more than 84,000 sites.

Perhaps the most telling aspect of the Bugmenot site/movement is in their FAQ section where they note that the top five most requested sites by their users are all online versions of newspapers.  So for newpapers, and the thousands of other content-driven sites that rely on advertising (and segmenting their audience to sell to advertisers) to survive – how can a poisoned database be avoided?  And unless they ask for registration, how will they have any demographic information on their users to share with advertisers?  The answer is: they won’t.  The biggest myth in online advertising is that you need to collect demographic data in order to succeed.

Search engines have realized huge success tauting their ability to connect with users irrespective of demographics and based on their online activities.  Google doesn’t need to know who I am to present relevant advertising.  If a grandma from Kansas or a new mom from New York are both searching for a new stroller – an advertiser selling strollers must reach both.  On search engines, you can do that.  So, how would this model possibly work on sites that need to serve banners, interstitials, or any other ad unit based on browsing versus search behaviour?  iVillage is one of the few sites leading the way in this regard through their ability to target users based on individual articles or piece of content on their site.  So, if there is an article about strollers, on iVillage you can segment your ad to appear on that page.  Imagine the performance of an ad placed here versus the run of site, geotargeting, or even demographic targeting we do now on most publisher sites.  If we could present ads in context, where they had the most relevance, imagine the performance these ads might realize. 

This is a model that other online publishers need to seriously consider implementing.  As their ongoing attempts to get readers to register details causes more and more consumer backlash in the form of faking details on registration forms and signing petitions on sites like www.bugmenot.com – the details held in any of their databases will be meaningless in a matter of months.  Even the good/accurate data will be impossible to filter from the "bob.com" registrations.  Ultimately, unless publishers improve their methods, advertisers will continue to shift their money to more consistently effective online spending such as search marketing.

WE RECENTLY REMOVED COMMENTING - LEARN WHY HERE >