Consumer Personas in Programmatic

Thanks to ambitions across the industry and the increased sophistication of data management practices, many brands are becoming far more datacentric. Of course, you already know this because you’ve seen all the headlines. So, what is changing?

Brands are realizing that their ideal target is not simply someone who fits into some broad category such as “Auto-Intender”, or “age 60-65”. They are cognizant of the fact that 3rd party data is often static, outdated, and inaccurate. Brands need more efficacious and transparent data, which they can control. Their ideal target consumer is (actually) a person with varying interests and beliefs. In various cases this realization has obviated the need for outdated solutions. Then, what is a more dated solution?

Consumer Personas. An ideal target consumer for a fast food restaurant does not choose to simply consume content online that is related to fast food. Their ideal target consumes this content in addition to all the other things they are interested in such as the Super Bowl, Marvel films, and spontaneous road trips to Colorado. Additionally, their ideal target doesn’t purchase fast food for the same reasons or at the same times. Some do it because they have a family, some do it because it’s cheap, and so on.

This is what we call a Consumer Persona. So, how are these created programmatically? Some say it can be done by evaluating the customer journey. Unfortunately, the specifics of this approach are often hidden behind Walled Gardens. Further, proposed solutions are often reliant on Personally Identifiable Information (PII), which may not be sustainable if regulatory agencies continue their trend of increased legislation.

At Semasio we create Consumer Personas with complete transparency in three primary ways:

1) By querying the Semasio platform. Semasio appends the key terms and phrases that individual users have consumed to their own user profile with strict data protective processes in place. Further, we have a view of >90% internet population. Thus, we understand, at scale, which users have an affinity for just about any topic. This could be 4 door electric vehicles, retirement investment accounts, the Super Bowl, veganism, The Avengers movie, and…. baby diapers? The generation of targets can be infinitely granular. When combined with market research this approach makes for a strong solution.

Note: We generate these segments in tandem with a client’s goals, which can be pushed overnight to a client’s DSP of choice. This is relevant to points 1, 2, and 3.

2) By general use of anonymized 1st party data. If a brand has customer data stored Semasio can discover which of their users have certain affinities. Semasio might discover that, for example, there are 4 or 5 differentiated groupings of users who share certain affinities for other topics such as hobbies, foods, vehicles, etc. We can then take those differentiated groups and model them outward to find more users like them. That is, we stack those users on top of each other to find what differentiates them from the rest of the population. Then, we find all of the users across the web who share that differentiation and choose a cutoff point. If a client is interested in gathering data from a specific page, Semasio can provision this via the placement of a pixel.

Note: We call our models semantic twins, not lookalike models, due to their clear differentiation. Thanks to data quality, Semasio needs only 350+ unique users to model from. The competitors we know of require far larger positive samples due to their use of high-quantity but low-quality data points. Our approach empowers granularity without sacrificing scale or quality.

3) Third, and certainly not last: By use of specific and anonymized 1st party data. Let’s imagine a marketer already understands which users fit into a certain Consumer Persona via structured data, such as market research or surveys. Or, perhaps the marketer has 3500 cookie IDs of individuals who signed up for a test drive and went on to purchase. Semasio would model this positive sample of users outward to find more users like them (as described in detail in point 2).

In a later article I will discuss a perfect relationship: granular audiences and creative assets.

For now, if you are interested in garnering complete control and transparency without sacrificing performance or compliance, please reach out! “Semasians” are always happy to make new friends across the industry.