What is Microtargeting?
Microtargeting is the theory that very niche audiences (micro-segments) with yet-unknown, and sometimes counter-intuitive preferences, have a high propensity for interest in a specific product or service.
Here’s a (made up) example: people who have just gone to the dentist are 10x more likely to buy a stereo in the next 3 days.
This fictional microtargeting insight would lead stereo advertisers to target people who have just gone to the dentist.
This all makes sense – but you must be asking: how do we know that people that just visited the dentist are more likely to purchase a stereo? This is where the technology comes in. Having seen explosive growth in competition in recent months (and number of competitors), online networks (ad networks, ad exchanges, DSPs, etc…) are trying desperately to differentiate themselves. One of the ways they are doing is this is by uncovering micro-segments (like the dentist-stereo example). They way they do this is by following people around the internet and collecting data. They tag users (with cookies) that have viewed specific content – in the example they would tag users that viewed dental health content – and then follow them around the internet observing their browsing activity for the next 30 days. Capturing millions (or billions) of activities and crunching the numbers with computers, networks can make statistical arguments that these micro-segments exist – and that they can leverage them to benefit advertisers.
Here’s why it’s a fad:
Micro-segments are rare. It’s true that they do occur sometimes (and these rare examples are always highlighted in glorious sales presentations) – but the notion that they exist for every product is inconceivable. Most large advertisers spend hundreds of thousands of dollars a year on consumer segmentation services that interview and survey consumers ad nauseam. These services may overlook an occasional insight, but the idea that they would be ignorant to any substantial segments is not plausible.
It is very interesting (or scary) that networks are able to collect such deep data about consumers – but just because the data is available – does not mean it’s useful. Or, perhaps they just haven’t found a useful way to use it, yet.
Thoughts?