Some use the words “segment” and “audience” interchangeably; I have from time to time as well. It’s understandable as both words refer to grouping consumers by their similarities. But there are important distinctions.
What is a consumer segment?
Let’s think of a consumer segment as a grouping of consumers who have similarities that have marketing importance (for informing innovation, communications, etc.). With segmentation, typically consumers are assigned to one and only one segment.
What is a consumer audience?
Audiences are non-exclusive, collections of consumers who can be targeted with advertising and are usually collected by media content, first-party data, or by building lookalike audiences. The same consumer can be in many audiences. For example, a consumer can be in a targetable third-party AdTech audience of males 35-54. They can watch America’s Got Talent, The Blacklist, Judge Judy n CTV, etc., and be in all those audiences too.
Strategic thinking considers both segments and audiences
As researchers, one distinction to note is that segmentation research puts people into mutually exclusive and exhaustive buckets. Audiences allow for the same people to be classified in multiple ways.
Those in media, think of it this way: segments live inside of audiences at some incidence or density rate that varies but is usually well below 100%.
Good media strategy requires bringing together two components:
- Define segments such that they are mathematically designed to deliver higher than average response to your advertising
- Deliver advertising disproportionately more to those audiences that deliver a high concentration of that segment.
Here’s the upside to this blog post: Marketers can do a much better job of implementing this plan with your help!
When segmentation goes wrong
I was consulting with a media company that had been asked to optimize ad schedules across their properties for an auto marketer’s “conquest” segment (those who own another make of car who never owned their cars). The optimizers worked as they should: leveraging smart TV data, profiled their shows across all their networks, and reduced the target CPM! But the marketer saw almost no lift! Was it the optimizers? The creative? No! It was the initial segment definition.
Applying the theory of OBM2 and targeting Movable Middles, I could immediately see that the auto marketer was targeting a consumer segment that was unintentionally engineered to not respond to their brand’s advertising and this would produce unprofitable results. The marketer failed on challenge 1 (getting the segment right) and everyone looked bad as a result.
One way to nail audience selection
I have found that AdTech third-party audiences have very variable quality, that is, the percent of IDs purported to possess a profiling attribute who actually do possess it. Even gender can vary from random (truth rate equal to national average) to 2.5 times random. So, choosing the right AdTech data partner would make a difference here. But there is one other step you can take.
Working with a company called Truthset (disclosure: I am on their board of advisors), you can sort out those IDs within an audience who have a high probability of possessing the asserted attribute. By grouping IDs into two buckets those with high vs. low “truthscore”, we saw validated rates of over 70% vs. under 3% (holdout sample vs. zero party data).
A marketer can unpack an audience and choose just that set of IDs who are “in the good bucket”. The math of this process was a “wisdom of crowds” approach. Truthset works with up to 20 data aggregators and when there was agreement on the same ID, the score gets higher (approaching 1); disagreement makes the score lower (approaching 0), using a form of Bayesian estimation.
Unpacking media strategies
When you think of the difference between segments and audiences and realize that you target audiences, it puts digital addressable advertising and traditional media on the same footing. This is great for simplifying media planning.
For example, if you are looking to direct Spanish language advertising to those who primarily speak Spanish, you can target “cleaned up” digital audiences but don’t forget the high concentration rates of the desired segment within linear TV programs on, say, Univision.
Unpacking media strategies into segments and audiences really simplifies things. Deploy ad impressions to the audiences that have higher-than-average concentrations of segments designed to outperform; this should be the core of your audience strategy. Those who are not in the segment within those audiences are more likely to be lookalikes for high-responding IDs so you will win on short-term performance and long-term conversion of non-buyers into buyers. Succeed at this and you should consistently outperform competitors who are not following this simple and universal gameplan.
