This is the first in a three-part series on segmentation. We’ll dive deeper into topics relating to the use of behavioral data and micro-segmentation in subsequent articles, but we’ll begin with Attitudinal Segmentations You Can Also Target.
Below is an example of two very different types of segments that many of us have seen before. One is “knowable” and sparks a clear mental image of what this person is like in terms of their attitudes and general needs. The other is “targetable” and is described via common demographics and observable characteristics:
Time and time again, marketing and insights leaders are forced to make a tradeoff: should our consumer segmentation optimize for knowability (attitudes/needs) or actionability (targeting)? The former is often done using survey data to develop deeper emotional segment profiles and can be very helpful for brand, product, and messaging strategy. But the resulting segments can be hard to find. The latter is more antiquated and provides little insight as to what makes segment members “tick,” but it allows for more identifiable groups that align with how media “segments” are typically purchased.
Our Approach to Avoiding Segmentation Trade-offs
Infusing Activation
Thankfully these traditional tradeoffs are no longer necessary! With a commitment to upfront planning, we can ensure you’re set up for both. Our approach is to infuse the way you go-to-market into our process without sacrificing the strategic power of a Knowable (needs-based) segmentation. The benefit of planning with the end in mind is that even these more emotionally-driven segmentations can be targeted and are actionable, especially when it comes to media activation.
From the outset we bring in all parties who will have purview over using the segmentation. Perhaps unique to other processes we’ve seen, we specifically include those who buy media, whether internally or from an agency. In order to set you up for media activation, it’s important to infuse the way you buy media directly into our process. We can’t wait until after the segmentation is already developed to have these discussions or it will be too late.
The Path for Segmentation
From there, our segmentation follows a specific path:
- Gather: Generating survey data and collecting individual-level secondary data on media behaviors (to enable future targeting).
- Explore: Understanding the relationships between variables in your dataset and how they can indicate the outcomes we care most about (buying, using, advocating, etc.).
- Model: Generating dozens (or hundreds) of potential segmentation solutions while optimizing for specific performance metrics.
- Debate: Identifying 2 to 3 solutions we love, and then pressure testing them with our clients before collaboratively selecting the optimal one.
- Deliver: Creating actionable recommendations and, most importantly, a way to connect your segmentations to media buys. This is the end goal that we always keep in mind. To get there, our “Gather” and “Model” phases are particularly important.
How it works
For simplification, let’s talk about the ideal scenario: our client’s agency subscribes to and can access a rich, syndicated dataset like MRI, GWI, Crossix, or similar. These datasets contain detailed profiling information, including attitudinal statements that tend to have a lot of predictive power for recreating your segments:
- During our “Gather” stage we discuss available datasets and/or where media is typically purchased.
- We then “Explore” by utilizing a hypothesis-led approach to curate a list of the attributes that likely have signal for predicting our segments.
- We ensure that our final “Model” includes those attributes that we can activate against when creating a secondary typing tool (algorithm) and can be applied where media is purchased.
The Result: Knowable AND Targetable Segments
Recently, using the above process for a leading healthcare client, we were able to create survey-based “knowable” segments that we could identify with 75% accuracy using only variables acquired from syndicated sources. Moving forward, the client plans to target and buy media using this approach for upcoming consumer campaigns.
Since all situations are unique the process can be customized and has a track record of high predictive accuracy even for survey-based, emotionally-driven segmentations, using tools and attributes that will allow your media agency to activate the segmentation.