Back to the Future: Direct Mail Data Is the Outcomes Era’s Missing Link

Paul Sobel, the CEO and founder of Dataline

Many conversations around digital advertising today focus on outcomes. Advertisers want their money to go towards media that inspires consumers to take actions, and they want tools that can accurately connect the media with those actions. Advertisers can then adjust their spending accordingly and optimize budgets across channels to drive the optimal level of ROI.

The funny thing is, this much ballyhooed “outcomes era” is hardly new. In fact, what digital marketers are striving for is already achievable – and has been for decades – in direct mail.

Eyes may be rolling into the back of heads right now, but bear with me. While direct mail is regarded as an old school format, the pillars of direct mail marketing are still relevant. The marketing data that powers direct mail campaigns is transactional data, which links it to a consumer outcome.

Bringing this data online and using it for digital audience modeling may be the missing piece advertisers need to succeed in an outcomes-oriented paradigm.

Modelling with incomplete data

There is no arguing that digital media dominates in terms of consumer time spent. Yet the models that marketers use to target digital ads are incomplete.

There are a few issues here. First and foremost, the data set is limited to online activity, and no human being lives their entire life online.

Then there’s the issue of inference. Online browsing activity shows what consumers looked at online, but doesn’t tell us what they were thinking or what they ultimately did. There is a lot of guesswork about consumers’ needs and interests based on their browsing habits alone.

Finally, there’s the issue of fraud. The sad truth is that some online activity used to create a model may not belong to a real human being at all. The advertiser’s budget then goes towards an audience that will never make a purchase or lead to an outcome. Beyond the wasted budget spent on impressions served to the bot, fraud skews performance metrics and leads to inaccurate views of which channels can actually drive outcomes.

Finding the missing piece

This is where bringing in direct mail data can fill in the gaps. Direct mail audiences are built on transactional data, meaning that the consumer’s interest profile is linked to something that they actually purchased. Building a model off of this data set creates a much higher likelihood that consumers will be interested in a product or offer.

Direct mail models also often pull in demographic and psychographic data on top of this transactional data to understand which customers are best qualified for a mailer. That’s a much more robust picture than simply knowing that a customer moved across multiple websites and consumed certain kinds of content. Even if a consumer ultimately makes a purchase digitally, it’s still relatively easy to determine if they received a mailer and/or used a promotional code or offer from a mailer.

Then there’s the fraud issue. Transactional data roots out any bots by definition, because bots cannot make purchases. Therefore, any digital audience models rooted in transactional data are going to eliminate the potential for the model to target bots or non-human traffic, painting a much more accurate picture of the addressable audience.

Moving past perceptions

Direct mail data can clearly fill in the gaps of digital targeting strategies, but there is one final piece that requires advertiser buy-in. Digital media’s popularity is rooted in the ability to get cheap scale. Even if response rates are lower, which they usually are, marketers are, in effect, throwing everything against the wall to see what will stick.

Meanwhile, direct mail is often seen as a channel with higher costs and fewer respondents. That’s what happens when you only target ads to real people.

Marketers need to get honest about their pursuit of scale. Are they using digital to drive outcomes, or to claim they hit the biggest audience possible? Even in this movement to outcomes, cheap scale is enticing.

Incorporating direct mail data into online audience models still provides scale, but it’s defined scale. Only those consumers who have made a purchase are included in the models. This is indeed smaller than the broad reach that comes with undefined scale.

As mentioned above, eliminating fraud also impacts scale. Digital targeting models built around transactional direct mail data will show fewer audience targets and a higher cost to reach each of those consumers. But the end result is that marketers will see better overall response rates. The outcomes and ROI will therefore be greater.