Completing the visitor targeting cycle

Copyright © 2009 James Taylor. Visit the original article at Completing the visitor targeting cycle.Syndicated from Smart Data Collective
Thomas Rose-Bolden from TaxBrain and Joshua Koran of ValueClick presented on visitor targeting. Joshua started discussing targeting. Online marketers try to reach the right visitors in the right context with the right message. A good result for […]

Copyright © 2009 James Taylor. Visit the original article at Completing the visitor targeting cycle.

Syndicated from Smart Data Collective

Thomas Rose-Bolden from TaxBrain and Joshua Koran of ValueClick presented on visitor targeting. Joshua started discussing targeting. Online marketers try to reach the right visitors in the right context with the right message. A good result for clicking on a banner ad is 1 out of 1,000 so the click through rates that represent success are very low. Behavioral targeting is a way to improve this rate. Contextual marketing used to be top of mind with ads being driven purely by the content on a page. Yet this takes no account of what else you might be looking at or how often you do the looking.

Behavioral targeting takes all your online behavior and then builds a usable behavioral profile that can drive more suitable ads. Simple approaches, using single events or simple patterns, work pretty well. More complex approaches can use a variety of data sources but must deal with differences between vendors. Must also deal with personal identifiable information which comes from various vendors but that ValueClick don’t want to keep.

The first attempts were based around clustering just like CRM but there is so much more data online that putting someone into a single cluster was too limiting.  Another approach is to allow a marketers to specify the rules for different ads but this lacks any kind of statistical validity unless a lot of testing is done and the dynamism of the internet makes it hard. ValueClick media uses a predictive approach – they build models to the metrics of the marketer and have these models adapt as data comes in. People can have multiple interest attributes too, allowing for broader targeting. The engine can also find unexpected and non-intuitive behavioral segments.

TaxBrain has used this approach. They wanted to spend less money on advertising while reaching consumers who wanted to do their taxes online. TurboTax and HR Block buy so much advertising that it is hard and expensive for TaxBrain to do so. They used audience maps to discover the ideal behavioral segments and then developed three campaigns:

• Geo-targeted ads for reach
• Retargeting to get last year’s customers and visitors who did not buy but visited last year
• Profiles to find in-market prospects

Some of this drove results this year, some developed leads for the future. They got much better click through rates from the targeted campaigns with up to 50% improvement. Behavioral targeting has been a big success for TaxBrain.

Link to original post