Welcome to our second article in our three-part series on segmentation. Our first installment focused on Attitudinal Segmentations That You Can Also Target. Today we are diving into another ask we often hear from clients: enriching segmentations with behavioral data.
When You Want to Know More, More Accurately
If we gave you a list of 50 common things, could you tell us exactly how often you do each of them? How many hours per month do you spend on TikTok, and is it more than you spend watching Netflix? How much more? These are topics that take up precious survey space and produce unreliable and undifferentiated results – humans aren’t great at answering them. But these topics are also critical for knowing how your segments differ to help plan campaigns, collaborations, and partnerships. We can help you unlock these answers, using real online behaviors – also called observed behaviors – of the actual respondents answering the rest of the questions in your survey.
We can go about this in one of three ways. The right one for your brand depends on your specific needs/use case, data availability, and resources dedicated to the effort.
Three Potential Options for Leveraging Behavioral Data
1) Observed, Behavioral Data
This first category is the gold standard, for enriching segmentations. Specialty providers give marketers and researchers access to true, observed behaviors for the respondents taking your survey. In other words, the same person who fills out a 15-minute questionnaire also provides 6-12 months+ of their browsing, search, and app usage behaviors. The data is verified to be the respondents, and it allows researchers and marketers to see inside of so-called walled gardens (Instagram, Facebook , TikTok, YouTube, etc.) to better understand types of content consumed, frequency of app usage, specific titles watched or searched for, etc., all appended to your custom survey.
Pros: true/observed behavioral data, insights into apps/platforms that typically do not share info, 1:1 connection of behaviors and opinions
Cons: can be expensive, and feasibility is limited to large/general segments of the population (not niche audiences, due to the size of the panels)
Our Take: when available and budgets allow, it’s the best option for enriching your segmentations
2) Modeled Data From Surveys
Some of the most common sources of this are MRI-Simmons and GWI. While not technically behavioral data (since this info isn’t directly observed), it’s often available to marketers and researchers via subscriptions or one-time fees. Media is purchased through these fields, and the data are often appended to custom surveys.
Pros: rich attitudinal questions (in addition to product usage), a fair level of ubiquity in the industry, great fodder for creating/re-creating segmentations
Cons: data from a few thousand people are modeled to cover 100s of millions, all data is stated by the survey taker rather than directly observed
Our Take: Given the challenges of relying on the underlying self-reported behaviors, we like it more for activation (i.e., media buying) than we do insights and/or precision of how one segment differs from another on a key topic
3) Aggregated Behavioral Data
Suppliers like Epsilon are well-known players in this space, and use a blend of public records, transaction data, publisher data, and sometimes surveys, to create rich profiles on adult consumers. While some data is actual, observed data, modeling is used to fill in other gaps for unknown fields for vast swaths of the public (similar to the above Modeled Data From Surveys).
Pros: closer to true observational data, and includes observed purchases; also has a level of ubiquity in the industry
Cons: in our experience, the data can be very sparse (and lacks full coverage for some respondents when appending to surveys) and cumbersome to work with; the coverage relies on modeling
Our Take: similar to the second category, we like it more for activation (i.e., media buying) than we do insights and/or precision of how one segment differs from another on a key topic
This topic is a complex one, and typically the best approach is to fully understand the limitations of each type of dataset before committing. We’ve helped clients in multiple industries navigate this landscape in order to develop actionable and targetable segmentations.
A couple of quick acknowledgments: This information covers typical consumer work and isn’t as well suited to more specific audiences like doctors or patients with rare diseases. Also, the richest datasets are generally more readily available in the US than in other markets.