Sizmek uses consumer, context, creative, campaign, and cost data to feed AI on optimization and prediction

Sizmek said today that uses 5 data sources to “predict which type of creative, what context, and what consumer will be most responsive.” Here the five Cs of data used by Sizmek:

  • Consumer data is what we know about an individual consumer, such as age, device ownership, past purchases, and behavioral data, like past website visits.
  • Contextual data is what Sizmek knows about a specific moment in time. What device is the consumer using right now? Is he playing a mobile game or reading the news? What time is it? What’s the weather like outside?
  • Creative data is what Sizmek knows about the specifics of the message, such as creative elements like headline, call-to-action, or product images. Putting creative data to use lets advertisers tailor that message to the appropriate moment.
  • Campaign data measures the ad exposures and results of a specific campaign, things like ad frequency, clicks, conversions, viewability, and demographic in-target rate.
  • Cost data represents the total price paid for each ad impression, including costs for data, serving, reporting, and win/loss auctions for media.


Sizmek says “what makes real-time AI special is that it “self-learns” by comparing its predictions to results—and then adjusts its predictions so the next prediction was better than the last.”