POV

This June the ANA released a study revealing that the advertising industry wastes $20 billion each year in programmatic ad dollars. The waste happens when advertisers buy low-quality websites, click-bait sites, and, frankly, a myriad of other murky, transparency-free zone shenanigans along the digital media supply chain. One statistic that really popped – $13 billion of that is on made-for-advertising websites alone, translating into 21% of all impressions and 15% of ad spend. With that in mind, please walk with me behind the curtain for a moment, to reveal ways brands fall for inefficiencies and can mitigate them.

The way to think about media optimization is both bottom up, and top down. From the bottom up, there are tactics media teams can deploy to eliminate fraud, focus on quality bidding and data sources. These tactics enable the ad platforms’ own machine learning to optimize against quality measures, rather than vanity metrics. From the top down, there are tactics media teams can employ to enhance attribution measurement, ensuring media investments are impacting business outcomes.

RELATED: Getting Smart About Ad Waste

As the ANA indicated, programmatic media – the automated purchase of digital media from ad buying platforms – can be rife with opaque, daisy-chain inventory management. Below are the categories of cautionary tales from the bottom up and top down.

1. Data for Audience Targeting

Media teams have the option of leveraging first, second- or third-party data to find desired audiences. For example, a pet food brand may want to add a data layer of pet owners. Where to find that pet owner is answered with the evaluation of first, second or third party data. Think of first-party data as browser data – browsers (and the humans behind them) that come to your brand’s website and are captured; second-party data as that which is owned, collected and shared by others; and third-party data as available for purchase.

These data sets can be uploaded and/or purchased from within the programmatic platforms. But here’s the thing – there are literally hundreds of third-party data providers from which to choose. How is a buyer supposed to know which of the dozens of data partners touting an audience of pet owners to select? The buying team should have a process and ranking system for data providers, especially paying attention to those which are deterministic and those which are probabilistic. A database of pet owners from purchase behavior is pretty deterministic; a database of pet owners from readers of “cute pet tricks.com” (totally made-up website, but you get my point) is probabilistic.

For consumer facing-brands, there is yet another category of data, the “0” data set – permission-based name, phone, address, email. This is the holy grail of data, and where many consumer-facing marketers are moving as best practice, some eliminating third-party altogether.

2. Regulatory Data Concerns

Following the European Union, many states in the US are getting very serious about digital privacy and codifying legislation of consumer data. This is a big concern for media teams, and something that should be taken extremely seriously by brands. Getting sued for data privacy regulations is costly.

As a recent example, a healthcare company’s media team leveraged first-party data (remember this is browser data) to retarget their website visitors in Facebook. This retargeting strategy is standard practice for B2C brands. The healthcare brand didn’t disclose any personal identifiable information, so all is good, right? Before I answer, remember, browsers represent humans. And the human behind one “browser” successfully sued the healthcare brand, arguing the healthcare brand did not have permission to share the browser’s data with Facebook, even if the sharing was through tags and not PII.

It makes sense, in a way. A person researching bariatric surgery may not want to see a bariatric surgery ad on their Facebook page. Data use can be very effective and efficient, in aggregate. But, also, data use can be very liable and costly, individually. So, take care.

3. Bots, Kidnappers and Other Digital Swashbuckling

From the $20 billion price tag, and continued increasing investment in programmatic media, you would be right to surmise the digital marketplace is extremely lucrative. Whenever there is that kind of money available, inevitably bad actors follow. There are industry-standard anti-fraud layers.

Programmatic media teams should curate desired site lists/publishers and continually scrub made-for-click sites, as well as non-brand-safe sites. Believe it or not, this is a bit of a manual process. There exist several industry evangelists who’ve developed sophisticated detection systems – above that which you get from typical IAB standard solutions. Using these additional tools is not only the right thing to do, but provides demonstrable improved performance.

4. Appropriate Campaign Measurement

Not to mix a culinary metaphor, but cookies have been historically the bread and butter of digital media efficiency. They’ve been essential in measuring campaign effectiveness, because they track consumers from exposure to merit event. In fact, most multi-touch attribution solutions have been built on audience identifiers, like cookies.

With cookie deprecation, and with increasing privacy and data regulations, persistent audience identifiers as the root methodology for attribution is increasingly challenged. If you’ve gotten wonky results from an MTA solution, this is probably why. At a recent ANA meeting, the Media Ratings Council hinted that “everything old is new again,” saying they are seeing emerging attribution solutions based on large-set probabilistic models, rather than deterministic, persistent audience identifiers.

Regarding attribution and ROI measurement, there needs to be a mindset shift, with a renewed focus on using machine learning and econometric models to measure channel impact, rather than individual journeys. Because individuals don’t like journey stalking.

5. Educate Procurement

Thinking about curated publisher partners, thinking about second party data partnerships (with potential clean room activations to safely share data); thinking about developing engaging experiences to gather 0-party data; thinking about modern attribution solutions – all this is to say that not all CPMs are created equally. Procurement needs in on the action, and should get educated around where hidden fees or cheap data or click-bait sites are used to compete on price.

At the end of the day, quality data governance drives improvement across several key areas of organizations. It increases sales with existing customers; it enhances marketing effectiveness; it increases the customer pool; it reduces customer acquisition costs; and it improves administrative costs.