lunes, 31 de enero de 2011

PPC is not for Miss Cleo

The development of forward thinking proprietary technology that can predict major Pay-Per-Click (PPC) outcomes such as conversions, revenue, and ROI is a current industry trend. Predictive optimization is one such advancement; a bid management tool used primarily at the keyword level. It establishes optimal bids for advertisers by leveraging historical and overall performance data.  If properly implemented, predictive optimization could allow those in paid search to focus more on driving revenue rather than costly bidding and monotonous keyword practices.

Predictive optimization involves analyzing complex outcome and predictor variables to come up with a best fit line. The line illustrates, with the least amount of possible error, the relationship between your predictor and outcome variables. In addition to the regression line, the process also produces massive migraines affecting laymen and search professionals alike. Feeling one coming on? Go pop an Advil. Back? Great. OK, simply speaking, predictive optimization helps you determine your ultimate ROI, cost-per-click and click-through rate.

For predictive optimization to truly be effective, it must adhere to a cycle of continuous improvement – ongoing testing, analyzing and evaluating. This is where many agencies cease and desist with their predictive program. That’s a problem because the power of predictive optimization is in the data, which changes rapidly in the short-term and the long-term. If an agency does not disclose the statistical significance of their models, how can folks be expected to trust their predictions? Accepting the results of a linear regression without at least knowing the significance of the model is like accepting a loan without knowing the interest rate. The results could be devastating.

Predictive optimization can be a useful, even revolutionary proprietary tool – if utilized correctly. It is the responsibility of performance marketers to thoroughly investigate such predictions before we implement them into PPC campaigns.

No hay comentarios:

Publicar un comentario