Discover some of the most crucial strategies to segmenting using a data management platform, including examples of how a DMP can enhance predictive abilities.
The emergence of data management platform (DMP) usage has crystallized the data-driven marketing revolution, making marketing decisions more about deduction than intuition. There is an investment increase in DMPs, which marketers favor for their ability to “identify high value users and run lookalike ad campaigns to attract more of these.”
With DMP adoption at an all-time high, many marketers still find themselves wondering how to make the best use of the technology in order to target customer segments with high specificity and favorable conversion rates. Here are some of the most effective strategies they are discovering:
The purpose of data-based segmentation is to go beyond mere observational capabilities. A spreadsheet can tell you about how many people fall in a certain age and gender group in a target region, but only data-based systems like DMPs allow you to go beyond such superficial qualities of consumers and examine their behaviors.
Segments based on data-like cross-site attributional and behavioral profiling, which can approach the granular level such as how many mouse clicks a visitor uses, are combined with qualitative attitudinal data to create segments not just based on who a consumer appears to be but why their actions make them different. For example, two single mothers may have very different attitudes when it comes to which types of foods they choose to serve their children.
These 360° perspectives are an “instruction manual for how customers should be treated in the future, considering all different customer touch points along with their analytical profile.”
Event-Driven and Dynamic Campaigns
Another element of segmentation is how campaigns respond when they recognize that a prospect is in a particular segment.
For example, frequency capping can limit the number of times a prospect is served the same retargeted display ad. Targeting exclusions can be applied, for example, if data indicates that the prospect is not a truck owner, negating the need to encourage them to buy a bed-installed toolbox.
Dynamic creative can recognize that this customer was examining articles about off-roading and they own a Nissan Xterra, allowing the marketing automation program attached to the DMP to adjust suggested and cross-sold products based on how far along the owner has come with customizing the vehicle. For example, a recent purchase of more rugged roof racks could indicate that the owner has a high chance of buying a new roof cargo carrier.
The ultimate potential for DMP-based marketing strategies is not to make educated guesses but to make more accurate guesses over time based on customer behaviors and the gap between expected and measured campaign performance. DMPs enable optimization of frequency, audience, inventory, and overlap by identifying cross-DSP audiences without restricting activation through a single DSP.
Examining past segmenting data such as credit card purchases and education level can allow data management platforms to observe trends like “individuals who completed only college are more likely to rent storage space after marriage.” Campaigns can respond in kind by predicting that 16% of such prospects will rent storage space. As actual campaign performance numbers roll in, the predicted hypothesis can be refined and also compounded on, such as saying “12 percent of this segment will rent storage space, but those that do rent storage space are 80 percent more likely to purchase moving services.”
As campaigns continue, more data gets captured, and hypotheses are tested, data management platforms get smarter at things like prediction, cross-selling and targeting while marketers get smarter about essentially predicting the future. The true potential of data-driven marketing is not whether these predictions are right as much as how much we can learn from each small failure.