Digital advertisers should be able to agree that there are about 345 million people in the US, making up about 133 million households. But, historically, media buying has been all about scale, even if it meant outrageously false population counts on audience segments.
More recently, advertisers have started to accept that there is a correlation between campaign performance and accuracy of audience targeting. There are a few factors that converged to get us here, including a turn toward legitimate content instead of MFA and the use of first-party data and IDs for targeting.
Reaching real people effectively is now the task at hand. Achieving that requires assessing data for (1) accuracy and compliance with privacy regulations and (2) richness with regard to any overlap between an advertiser’s target personas.
Ensuring accuracy along the data daisy chain
Data should be consistent or at least predictable. While the US population overall is stable, what people do and where their data appears varies. Aggregating and normalizing data from disparate sources will expose advertisers to a lot of volatility. Advertisers need to understand those patterns so they can spot the difference between fraudulent and valid changes in audience activity.
The number of female lawyers isn’t going to change wildly week over week, but the population of Nantucket might, depending on the season. Similarly, the number of people who use a streaming app might change enormously if there is a major game, political debate or awards show. Tracking and analyzing the changes in volatility and favoring consistent data, unless there is a clear explanation for volatility, will help advertisers stick to accurate numbers.
Paying attention to the factors affecting data collection across the internet is vital. If a browser company changes their privacy settings or a publisher partner puts more content behind a paywall, data is going to fluctuate.
Data origin also matters. Data gets passed around the digital advertising market. This can make it hard for advertisers to pinpoint where the data came from, if it was collected in a compliant way and if it’s representative of real people. To get those insights, advertisers need to establish specific media buying requirements. There should also be a process in place to check any new data sets against a highly scaled set of IDs with info on most of the population.
Another way to avoid fraud is to work with direct data partners like publishers and retail companies. Paying extra for data is often worth it – trusted data collectors know more about their audiences, which helps with customization and segmentation.
Ultimately, a good data provider should be able to show – in detail – the path of every single bit of data. Any data that is opaque or lacks appended information is suspect and not worth the scale.
While first- or second-party data is ideal, it’s not fully scalable. Turning to third-party data providers is fine, as long as advertisers take an active role in data oversight.
Filling in identity blanks
Having 10 million hashed email addresses with no other insights attached isn’t as valuable as having five million profiles that include recent browsing behavior, past purchases and demographic info.
Filling in the blanks around a specific ID is a science. While the process can give advertisers a richer picture, it can also create overly specific targeting parameters, which could lead to small audiences. This is due to overlapping data sources, which are limited to the universe from which they came. Data partners don’t see the whole population, just those who visited their property, so it will be inherently limited.
Because of these limitations, segments get small fast. Not everyone in the Indianapolis region searched for airline tickets in the past 30 days, so that data is already very limited. Add on a requirement that the segment is women 18-24, and the result could be only a few hundred people. Not very impactful. The right approach balances accuracy with scale so advertisers can reach people effectively.
Adhering to data best practices requires advertisers to focus on specific clues so they aren’t fooled by fraud and unnaturally high scale.
Knowing these patterns, and learning how to judge and compare data sources, are key to long-term success.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow TrueData and AdExchanger on LinkedIn.
