What are some positives about?

Exactly what audience segment attributes are being measured in online display advertising to generate "signal", "positives", and a high "index"?

• This question is about measuring signal and positives of an "audience segment" (i.e., group of users/browsers/cookies with similar audience attributes) that the marketer might want to bid on to serve the ad to in a real-time bidding (RTB) environment. The question is NOT about measuring the effectiveness of the creative of the display ad at driving click throughs. Audience Management Platforms (AMPs), like Media6Degrees, talk about 'signal' and 'positives' in modeling and creating relevant audience segments. What are the attributes of signal, positives, and high 'indexing' of such audience segments?

• Answer:

  These will be defined differently based on the end goal, of course. We built our DMP from the ground up to capture and score buying signals associated with tens of thousands of different products in a vertical. Examples of signals based on user behaviors and actions may include: product and/or category searches, reading multiple reviews on multiple sites of same/similar products, price-comparison shopping, clicking merchant links in editorial, asking social networks about specific products, users' previous ad interactions, etc. With enough data and the right algorithm, the machines learn which signals are most likely to precede actual purchases, which ones tend to yield false positives, as well as how to recognize the impact and take advantage of ad placement and bidding strategies in achieving success with high-scoring cookies. Many providers today focus primarily on retargeting users who've been to the marketer's site. This suggests that they consider this one datapoint to be the only "signal" that really matters.