Why You Should Ditch Your 3rd Party Audience Segments
It has become common practice to add a 3rd party audience segment to an online advertising campaign in order to target more relevant viewers with your ad and thus raise the performance metrics of the campaign. The idea being that you pay an additional premium on your media cost to target a subset of high value viewers and in turn you generate a greater return on spend because the targeted viewers have a higher propensity to engage and convert.
The logic behind this campaign strategy seems sound and can even be convincing enough to persuade a marketing director to increase campaign budget to accommodate the costs of targeted audience segments; However, if you talk with media buyers they will tell you that adding a 3rd party audience segment very rarely raises your acquisition rate, in fact it can often result in a lower Return on Ad Spend.
It is not until you take a closer look at how the users in an audience segment are selected that you realize paying a premium for the audience is nothing but an inefficiency in the campaign. Most segments that are compiled by SSP’s and DMP’s rely heavily on descriptive models to break down the profile of a user based on available data points stored in a user’s web history. They are limited to using whatever variables come across in the ad request, which as it turns out, are also available to target against in most bidders for no extra charge. Once the segment provider has described each viewer they then classify individuals into groups based on different static characteristics and leave them in these groups for 30-90 days.
Two big issues here, 1. Static classifications based on contextual variables have very little to do with purchase propensity and 2. Individuals who have a high purchase propensity today will most likely not have the same purchase propensity in 30-90 days. Take for example your traveling to Las Vegas, you search “New York, New York Hotel” as in the hotel in Vegas, you were also recently on your favorite travel blog reading about how to get through airport security faster. Based on these two contextual variables you could very easily be looped into a “New York City Travelers” segment, however you have no intention of traveling to New York City. Advertisers who target the “New York City Travelers” segment will be paying a premium to put irrelevant ads in front of you because they think you are more likely to convert. Additionally, you will stay in this segment eating up advertiser’s premium programmatic dollars and receiving irrelevant ads for 30-90 days.
While most segment providers add little to no value there are a select few who generate audience segments geared towards conversions, however they are far and few between due to the high cost of creating and measuring conversion-based audiences. To build high quality conversion audiences a data provider must make a large investment in human capital, hiring multiple data scientists, as well as pay high dollar amounts to purchase 3rd party transactional data that can be used to measure and validate their segments.
Rather than making this investment many segment curators just revert to clustering groups of viewers together based on static data elements captured in their ad call with no regard as to whether these viewers are converting or not. This is the reason why adding one of these segments to a digital campaign does not raise conversions rates, the segment is not built to drive conversions and thus does not add any value beyond being a high-priced targeting layer. The proposed alternative would be to instead, deeply understand your first party data and leverage known variables about your customers to orchestrate a digital advertising program that is more cost efficient, but also highly focused on activating audiences you know will have propensity to purchase.