By Behnam Rezaei, CTO and Co-Founder, NetSeer
There is no doubt that digital advertising has been a boon for many companies and industries. Allowing for greater measurement, targeting and access to consumers than ever before, it’s hard to uncover any drawbacks in its practice. There is room for improvement, however, especially in three areas, which I see as digital advertising’s most critical pain points. These include the use of inaccurate attribution models to justify fake ROIs, consumer data leakage and poorly targeted ad messaging.
Bad Attribution Equals Great ROI
Criticism of digital advertising’s attribution models is nothing new. For years, people have argued that they don’t give us an accurate picture of what actually prompts a specific customer conversion. This is true, as multiple touch points are usually involved in a purchasing decision, and it’s difficult to determine definitively which one led to the conversion.
What people don’t talk about as much is why, even with all the flack they receive, companies continue to employ these faulty attribution models. I believe this is because these broken models make ROI numbers look good for CMOs and other marketing types, who are under the gun to demonstrate they’ve made a profitable investment on their ad spends. The head of marketing for a major bank, for example, can use last-touch attribution to show her CFO that for a $50 ad on one specific channel, a new customer signed up who is worth, say, $80 in interest and fees in the first year he is with the bank.
An attribution model that accounted for the actual journey the customer took to that bank, however, would be another story. It’s likely that multiple impressions on multiple channels led the customer to the conversion, not just the last channel. Examining the ad spend on all these channels would likely demonstrate that the customer cost a lot more than $50 in advertising. That’s not going to make the CFO happy, so the head of marketing will continue to use an attribution model that doesn’t pinpoint the true value of a specific channel to a conversion. Also, using attribution in this manner heavily incentivizes retargeting, painting an even more muddled picture of a conversion’s value.
Who Owns The Data?
Another pain point has to do with the data leakage stemming from consumers’ interactions with advertising publishers. Say a woman reads an article about pregnancy on the website of a prominent magazine. By reading that article, she offers a clue about herself: She might be pregnant, and looking for services and commodities related to pregnancy. As things currently stand, any company with a tag on that page, such as an advertiser, an ad-server outfit or social-media outlet, can learn that this woman consumed information about pregnancy.
What’s more, each of these companies has data partners with which they share that information. Those partners then share that data with their own partners, in a daisy-chain effect that dissipates the information throughout the digital advertising ecosystem. All of these entities can take this data and monetize it in some way. They are making money off the potentially pregnant woman without her even being aware of it. The magazine, because it must rely on third-party services such as Facebook and advertisers to make money itself, is also giving away data that others will monetize. Data leakage, then, is more than a consumer privacy issue. It’s also a financial problem.
In the digital advertising ecosystem, who owns, and therefore can monetize, consumer data? Is it the publisher that posted content drawing a reader’s interest to its website? Is it the ad technology company that served an ad on the page the consumer read? Is it the advertising agency that placed an ad on that page? Right now, nobody knows, and it’s led to a free-for-all of data sharing that is bad for the consumer and bad for the content publisher. We must resolve this.
Strong Targeting, Weak Messaging
A final pain point relates to targeting in advertisements. Digital advertising gives us tremendous insight into consumer preferences, yet few brands leverage this to create ads with highly targeted messaging. Instead, they generate a few ads for a wide, undifferentiated swath of consumers. A mobile carrier, for example, put out a few nationwide ads enticing a consumer to purchase its latest smartphone. The content of those ads contains nothing targeted to specific segments of that carrier’s potential customer base—even though doing so could almost guarantee those customers would convert to the product.
A better tactic is to change an ad’s content or messaging dynamically, based on the target consumer. We see this mostly in ecommerce, where ads draw from data mined from previous customers interactions with a brand to present him or her with the products or services the person is most likely to purchase. Even without retargeting, however, brands have access to data about their potential customers they can use to create better-targeted ads. They should be adopting single-tone messaging for selected audiences instead of disseminating a few static ads to consumers who may or may not be right for a specific product or service.
Companies that have the information they need to target an ad, along with the creative resources to generate that ad, are particularly adept at this approach. In a sense, like Apple, they have both the hardware (targeting technology and resources) and software (the creative). The advertising-technology provider Criteo, for example, works with Internet retailers to serve personalized online display advertisements to consumers that have previously visited the retailer’s websites. Unlike some of its competitors, which only offer targeting technology, Criteo also creates and inserts highly personalized ads for its customers on various channels. More companies should take their lead.
Digital advertising offers unprecedented opportunities for learning about and engaging with consumers, yet the pain points I outlined here make it difficult for companies to take full advantage of these benefits. Everyone in this practice should work together to resolve these issues, as doing so will help pave the way for the next big breakthroughs in the field.