November 5, 2010 by Andrew Q. Kraft
The ability of marketers to now find highly precise specific audience sets online, enabled by data companies such as BlueKai and eXelate, is piquing clients' interest. There's no question that data is suddenly the "it thing" and data is available more easily than ever before. Audience targeting is a key component to optimise campaigns, drive business results, and crucial in meeting advertisers' overall needs.
Data, however, can be a trap, as well. It is "easy" to sell to a marketer the audience they think they want (e.g. "auto intenders according to Bluekai and eXelate who also are moms according to DataLogix"). The problem is those audiences tend to be small. No data provider has managed to tag even a majority of these pools. Why? How many auto intenders do not go to Kelly Blue Book (or now Autotrader)? Most of them. How many moms do not go to moms sites? Most of them. The key question is... how can you find more of these users? Data Amplification is the answer.
While data cannot change consumer behavior in and of itself, it can reveal important attributes about the user. These attributes can indicate intent. Companies like Nielsen have been making a business of this for decades, positing (successfully, as proven by the fact that they get used by marketers over and over again) that people who live in a particular zip code have a tendency to buy certain products. This sort of analysis is possible today in the AMP® system and to customers of the AMP Network™. The key is creativity. If you want to help a marketer find an audience that you know is 5% of the marketplace, do you want to target the 0.5% of the marketplace (10% of the audience you seek) and leave the other 90% of their target to their competitors, or do you want to help the marketer reach the 10% of the marketplace most likely to include the 5% that are in market? The pendulum has swung too far in the direction of explicit targeting. While useful - and it should be the core of any campaign - Data Amplification, the ability to model out users to find others who are not yet tagged explicitly but are most likely to have the same attributes or purchase intents, will help you get that bigger buy from your clients.
The key is having enough data from enough disparate sources. The AMP Audience Cloud™ brings together data from over 35 data providers (and growing) to power your AMP platform or your AMP Network reach extension buys. The Audience Cloud allows you to gather data from your advertisers (aka retargeting), your own sites (using our Personifi classification system), and 3rd party data providers not only to target, but also to provide Data Amplification to your clients through lookalike modeling. Sometimes the data isn't necessarily "obvious". If you're targeting luxury auto intenders, an obvious Amplification includes PRIZM segments that index well to new car purchasers. A not-so-obvious one might be that most high-end vehicle auto intenders in the market during the summer may also happen to like gadgets and technology. One of our clients shared with us a client success story where they found that people most interested in a hotel chain also happened to be pet lovers (it was a pet-friendly hotel), and when they targeted pet lovers as an Amplification set, they actually had better response metrics than the original audience pool that had been sought.
Each user in the Audience Cloud has a signature - like DNA - based on the data gathered from all the data sources we import. Imagine looking at a client's retargeting pool and discovering that their users have some key attributes in common... and targeting those attributes as a lookalike pool in order to power a much larger buy than straight retargeting. That is Data Amplification in action.