Three Audience Targeting Tactics You Didn’t Know Were Possible

By Robert Konkos
Vice President, Strategic Data Partnership Development
Media Center of Excellence, IRI

Buying audiences is a fairly routine exercise these days. The campaign brief has been written, the insights and planning work have been completed, and you know who you need to target. Now you buy the demos and lifestyle segments to ensure you have a contextually relevant audience, and you should be good to go. Alternatively, maybe you even bought a modeled audience claiming to be based on actual purchase data. But, how much purchase data was actually used, how recently it was collected, and the quality of that data is usually not disclosed. (This is not ideal, but not unusual, and you may think there isn’t much you can do about it). 

The campaign begins, the response data flows in, and you begin to optimize the best you can based on response metrics, such as engagement, click-through, etc. You may wish that actual conversion data was available, but since 95 percent of CPG sales still happen in a brick and mortar store, you often don’t have the luxury of incorporating conversion data.

The above scenario is very typical for advertisers – but it’s not the best one available. There are three specific audience targeting tactics that can significantly improve your campaign performance.

  1. Use 100% Deterministic Data. Audience data built using 100% deterministic methods is available to build and use CPG audience data in your campaigns. By using a provider who has amassed a robust single-source data set across hundreds of millions of de-identified frequent shopper loyalty cards, you can build audiences using actual purchase data and do so at the scale needed to fulfill your frequency and reach objectives. Lapsed buyers are good examples.  If you want to target households who bought the advertised product during the previous 26 weeks but not the current 26 weeks, you can only do that with 100% deterministic data.
  2. Refresh Your Deterministic Audience Based on Purchase Cycle. If a product has an eight-week purchase cycle, you don’t want to waste impressions by delivering media to households that are known to have bought the product recently. Instead, start with an audience that you know buys the product or category but just hasn’t done so recently. Even better, you can add households to the audience that are outside of the purchase cycle and based on their purchase behavior.
  3. Use Recent Purchase Data to Optimize Campaign Execution. If you are working with a DSP and using 100% deterministic data, you can send the DSP all the households that have purchased the advertised products. From there, the DSP can optimize its models by matching the deterministic households to the households they know have been exposed to campaign media, and begin to tune their algorithms to favor the campaign tactics that tended to drive households to actually purchase.  

Our initial work has shown that levering in-flight optimization techniques can increase ROAS by up to 80 percent versus a non-optimized campaign. Considering that performance is everything, advertisers that embrace audience targeting in an ongoing and dynamic way as described here can gain an edge that differentiates their campaign execution and ultimately drives relatively outsized performance.