How to make programmatic less problematic
Automated marketing is like a gifted adolescent: The long-term outlook is rosy, while the short term is troublesome.
Marketing would be so much easier to quantify and strategize if it just stopped moving so fast. Were programmatic advertising a person, age-wise it would be in the fourth grade. Yet it's progressed so quickly in its short time on the planet that it's on a fast track for grad school at MIT.
According to a recent state of the industry report from programmatic platform AdRoll, the percentage of the more than 1,000 marketing and ad executives polled who are retargeting on mobile jumped from 54% to 82% in 2015. But more amazing than that is the fact that a third of those surveyed for AdRoll by Qualtrics said their companies had yet to field a mobile app. A quarter of them hadn't yet optimized their websites for mobile. It's no wonder marketers have mixed feelings about programmatic.
“‘My advertising doesn't work, the website sucks, and my salespeople don't follow up on my leads.' You hear this repeatedly across the B2B marketing landscape,” notes Peter Isaacson, CMO of Demandbase, an account-based marketing platform. “There is frustration, but there's also an air of resignation that, ‘Yes, I know some of my ad dollars are wasted, I know they're not reaching the audience I want, but how else am I supposed to reach a broad segment?'”
Erich Wasserman, cofounder and chief revenue officer of MediaMath, lays out the rocky path for B2C marketers.
“What programmatic has done is create a great deal of media liquidity in the marketplace,” he shares. “There are billions of opportunities on a monthly basis, and the sources of liquidity have multiplied — video and mobile and desktop, rich formats and social formats. The question is how to manage that enormous opportunity.”
Wasserman's answer is software, and MediaMath and scores of other providers of cross-channel solutions and data management platforms are zealously tossing them out like life preservers to the marketers wading deeper into programmatic's inviting sea of scalability. In 2013, just 7% of marketers spent more than half of their budgets on programmatic advertising. That percentage doubled in 2014 and, by last year, according to Qualtrics, one in five marketers were flowing the majority of their funds into programmatic.
Much of that increased investment is being spent on mobile, literally a moving target across both time and location. One solution finding favor is Go2mobi, which promises marketers a way to deliver programmatic ads depending on the time of day, the weather, and the recipient's proximity to home and work. Like many cross-channel solutions seeking to track people's daily migration from screen to screen, it uses probabilistic matching to follow desktop IDs to mobile. Also like most solutions fighting to gain traction in a whirlwind of change, it's a work in progress.
“There's still not a perfect closed-loop system for tying an actual cash register ring to another touchpoint, like when [customers] make a phone call. It depends on how sophisticated the marketer is,” explains Go2mobi president and cofounder Tom Desaulniers. “Some are really ahead of others and some are totally flat-footed. There's a huge opportunity for companies in our space.”
Solutions providers, therefore, are spreading the table with a banquet of options allowing marketers of all sizes and capabilities to tuck into programmatic campaigns that are more targeted and efficient.
Getting what you pay for
If we return to the metaphor of programmatic being an ocean of scalability, most marketers are like first-time boat owners who blunder out into the sea lanes in their new Bayliners without the benefit of a boating course.
“Here we have these people who are basically the pioneers of artificial intelligence and they don't understand the sophistication of programmatic advertising,” notes Jay Wilson, VP of omnichannel platforms and marketing technology at Healthgrades, an online service for finding healthcare providers. “They should take a look at it. It's so cool. When you think about the time it used to take to buy and sell media, these transactions consist of 22 steps and happen in 30 milliseconds.”
Programmatic ad platforms have come a long way to make the sailing smoother for marketers still finding their sea legs. NinthDecimal, for example, helps brands use their own CRM data to tap into more than 200 standard customer segments online or to create custom segments attuned to their go-to-market strategies. Programmatic agency Rocket Fuel is putting machine learning to work in a system it calls moment scoring that, like Go2mobi, looks to serve relevant ads to people by location, time of day, and purchase intent, as well as by customer segment. It calls on its own eight years worth of data to inform buys, and it claims to have made an advance beyond the marketing segment most common on the programmatic scene.
To ensure efficiency across those nanoseconds of a programmatic decision, however, one must understand with specificity how several touchpoints add up to a sale.
“It can be easy, say, when the user clicks through on a display ad to a website and makes a purchase,” Wasserman says. “In the more complex case of online sales, though, we're still able to make connections. A customer sees a mobile ad and makes a purchase. We're able to ID that user's device, and, lo and behold, that user is a member of the retailer's loyalty program.”
Rocket Fuel's path to programmatic efficiency is a pricing system that flies in the face of usual programmatic bids that place limits on buys, such as a $1.75 CPM max for women 18–34 years old or $6 for retargeting. In its moment scoring system — depending on the client and the product and, say, time of day — a bid of $1.25 might be the better play for a hair care brand to reach someone like Sally at 9 p.m. But it could be worth a $7 CPM to an electronics marketer to retarget Jim, who is predicted to be on the verge of buying new headphones. Rocket Fuel professes to be able to call on more than 11 million customer attributes in making buys, at the same time weeding out bad actors by noting any lack of purchase behavior.
The one-to-one moment scoring example Rocket Fuel doles out to customer “Annie” might give the shivers to the average privacy-conscious consumer (even though personal identity is not considered). On a mobile ad for one consumer product, Annie gets positive scores for being on a mobile phone and having clicked on an ad at 8:22 p.m. But she gets negative ratings for having already seen three ads and having read a fashion blog. Her positive moment score of +0.24 merits a CPM bid of $5.56.
“Moment scoring is a way for us to articulate what we're able to do instead of trying to explain the artificial intelligence applied. What we're saying is that you can do one-to-one marketing to some degree in programmatic rather than retargeting someone into submission until they buy,” states Rhonda Shantz, Rocket Fuel's VP of marketing.
Facebook flips the proposition
Persistent identification — tagging that golden record of a single individual across the cookied and non-cookied digital universe — is one of the hallmarks of customer engagement at this juncture in the digital age. It's something that cross-channel marketing technology companies claim to specialize in despite the lack of connectedness across channels. There is one company, however, that has come a long way in a short time in relevant ad delivery.
“Facebook has been so successful with Custom Audiences, and it was faced with a huge challenge,” recalls Signal founder and chief revenue officer Marc Kiven. “Every Web page in its universe was the same. The only thing that differentiated them was the people visiting. So Facebook flipped the proposition. It didn't offer advertising; it offered a curated audience. That's what we do: help our customers create curated audiences.”
Kiven explains that what his company does is enable “addressable media,” something different from display ads and even retargeting in that it's about creating an ongoing dialog across customer journeys. Essential to the process is constructing what Signal calls a recognition graph. This entails moving all known offline behavioral data online, dragging it through CRM systems, and endeavoring to recognize all customers across all channels.
“In our world, it's a deterministic system. We convert PII into an anonymous identifier and spread it out over the graph,” Kiven says.
This may be a boon for B2C marketers, but working without the names, emails, and phone numbers of prospective customers is something alien in the B2B world, where personal relationships and nurturing have always been the key to big-ticket sales and long-term contracts. A recent study from Wakefield Research for Demandbase discovered a B2B marketing community adrift in an anonymous digital world. Seventy percent of B2B marketers surveyed said advertising regularly failed to meet their expectations and 96% agreed that their efforts inevitably reached significant amounts of people who never would buy anything from them. Most frustrating is that B2B dealings are migrating more and more to the mobile channel.
“The bulk of B2B advertising is still cookie-based, so the mobile user is hard to identify,” divulges Isaacson of Demandbase, which is able to end-run cookies by identifying prospects through their companies' IP addresses, whether they're on mobile devices or desktops. To address the person-to-person issue, Demandbase recently acquired WhoToo, a data-as-a-service provider with millions of behavioral, functional, and personal files that let Demandbase identify individual decision-makers in its enterprise database.
Data providers are digging deeper to bring users of programmatic closer to their most important customers. Earlier this year MediaMath introduced a new business unit called Helix that presents retail clients with the opportunity to take advantage of super-
charged behavioral data by hooking into a huge second-party data co-op. Helix was built on shopper data from a third of the top 100 retailers that had been collected by MediaMath's Adroit Shopper unit.
Longtime MediaMath user Rachel Silva, assistant VP of marketing at Pep Boys, was a beta-tester of Helix who says it's paid “huge dividends” in her programmatic investments.
“Over the past few years, people were doing a lot of behavioral analysis in buying programmatic programs, but it was all done in kind of a vacuum. It was like Legos that had to be stacked one on top of the other. But this changes the game,” she adds. “We use programmatic mostly for retargeting, and retargeting pools can be huge. But with this second-party data we can identify the people who are most likely to convert and reach for them first. It allows us to layer our spend.”
...in with the old
In the midst of such hard-charging martech advancements, is it possible that one of the most promising data-driven marketing solutions for e-commerce companies could be a programmatic version of good, old-fashioned direct mail?
“Who'd have ever thought we'd be applying big data to something as old as a catalog? It's weird and cool at the same time,” stresses Shawna Kaplan Hausman, VP of
e-commerce and digital marketing at Giggle, a retailer of high-end baby gear that uses services from programmatic direct mail provider PebblePost. “When I'm at conferences, people come up to me and want to know about PebblePost.”
Hausman and Giggle were beta-testers of PebblePost, a company that in the space of less than a year has built a fast-growing, programmatic direct-mail business. The company's founder and CEO, Lewis Gersh, who secured a trademark on that term, is a heralded martech venture capitalist who has funded companies such as Indiegogo, Madison Logic, and Tapad. But he quit his Metamorphic Ventures to devote all his time to what he envisions as a multibillion-dollar business of the future.
“Programmatic was sweeping through display, but it was getting overdone and efficacy was falling. PebblePost was born out of that,” Gersh shares. “Everybody looked at anything but direct mail as a solution. But direct mail is second only to TV in marketing expenditures.”
PebblePost launched last June on a promise of 8% response rates and 15% conversion rates. But when beta-testers like Giggle posted numbers closer to 20% and 40% in those categories, respectively, Gersh knew he was on to something. PebblePost is run on an ad server that is activated by customer-determined hierarchies such as product segments.
For instance, should a customer go to a retailer's website, look at some shirts and blazers, and then leave without purchasing, a postcard with an offer on an apparel purchase could be sent out to them within 24 hours of their visit. Of course, the retailer must have their physical address on file in its database.
“I came in [to Giggle] very skeptical of direct mail. It's such a huge expense,” says Hausman, who previously worked at Williams-Sonoma, which has a large catalog business. “Now I feel like I've become a proponent of direct mail. You've got to be so good at search today, for instance, or you end up on the second page. It's gotten to the point that direct mail can be disruptive.”
Giggle currently sends PebblePost postcards to every customer it recognizes on its site, but Hausman's looking forward to a day not too far away when PebblePost's Address Cloud comes on-stream. The company is filling the cloud with customer data from its users — now numbering more than 100 brands — that will be able to tap into it and send appeals to site visitors for whom they don't own any PII.
“We'll have ownership of the address and won't give it to the client. It's like when someone does a search and clicks on a link for an item. The retailer doesn't know who it is, though Google may know,” PebblePost CMO David Cooperstein explains.
Even in the fast-paced, algorithm-driven world of digital marketing, the old can become new again. Meanwhile, the new breeds more complications. Consider: Fewer than half of marketers in Germany buy ads programmatically because nine out of ten of them fear losing total control of their customer data. Consumers insist on being left alone on the Internet, using ad blockers and beseeching governments to protect their privacy. Federal Communications Chairman Tom Wheeler is pushing new rules for Internet Service Providers that would require them to have users opt in to allow marketing to use their data.Ultimately, a middle ground will have to be reached because automated marketing is — harking back to a phrase from the Great Recession — getting too big to fail. Programmatic buying accounted for more than half of display-related advertising on the Internet, generating $10 billion in revenue in 2014, according to the most recent report from the Interactive Advertising Bureau. Considering that surging mobile and video formats made up just 20% of that score, the process is clearly poised to rise up like a tidal wave. As is normally the case with adolescents, programmatic will face its fair share of problems, and most of them will likely be dealt with.