When ads for Zappos started to appear to follow users around the Web, some people found the experience a bit eerie. A person would shop for a pair of shoes but not buy them, and then see ads about those shoes show up on other sites they visited. Watch a video on YouTube? See an ad for the shoes. Check on a TechCrunch blog? See a similar ad for the shoes.
Welcome to the world of real-time relevance, where online ads react to user needs by context and action, and where media buyers can bid for highly targeted display space in a split second.
While retargeting campaigns like the one from Zappos might raise the hackles of privacy advocates (the ads now include a conspicuous link to an explanation of how the ad works and an opt-out), they’re also wildly effective for a wide range of different online purchases from jeans to webcast hosting. For instance, comScore found retargeting provided a vastly greater lift than five other media placement strategies when it came to brand website visitation for trademark search queries.
None of this would be possible without the powerful back-end analytics that are transforming how agencies and marketers assess their media buys by giving them a more complete look at overall online user behavior. Consumers have become increasingly numb to banner ads, and marketers have had to look at new levels of metrics to measure effectiveness, engagement and ROI. Click-throughs, impressions and conversions, once the key ways to measure consumer interest, are a thing of the past for many marketers. Which is one reason why technology providers and agencies are moving beyond one-size-fits-all marketing into a version of narrowcasting that focuses messages and offers to highly specific segments with highly specific interests.
“We’re moving toward a more targeted behavior pattern,” says Kent Emeson, senior marketing manager at digital ad metrics provider Adometry. “You’re going to see ads much more targeted to individuals.”
This level of targeting is one of the reasons why U.S. online ad spending continues to experience double-digit growth. eMarketer forecasts a 10.5 percent rise in spending to $28.5 billion this year, climbing to $44.5 billion in 2015.
Because campaign metrics typically have been based on user actions, much online advertising has started to take on the look of direct response. Ads and campaigns are focused on actions, leads or sales. It is hardly a coincidence that performance-based pricing hasn’t just overtaken CPM models for Internet ads, but is crushing them. In the first half of 2011, according to figures from the Internet Advertising Bureau (IAB) and PwC, performance-based pricing accounted for 64 percent of Internet ad revenue, while CPM/impression-based pricing was 31 percent, its lowest ever.
This use of analytics is allowing agencies and marketers to optimize their ad spending. Managing traffic and tracking behavior at a company’s website is just one part of the equation. The greater power can come when companies are able to track audience behavior across the Web and use that data to make more informed ad buying decisions. This goes far beyond cookies and into rich data mining and analysis. By seeing this information in real time, campaigns can be quickly adjusted so that they ensure maximum return.
The implications of this real-time model are significant. One repercussion has been a greater emphasis on real-time bidding for online display inventory. Not only is this method being used to help publishers sell off excess ad inventory, but analysts believe there is increasing pressure for real-time bidding to be used in place of many person-to-person—or direct—sales. (See sidbar at left.)
Simply put, real-time marketing decisions based on analytics are driving greater automation of ad sales. This may not necessarily reduce spending, but it speeds up decision-making and ideally gives a much clearer and timely picture of ROI.
Going beyond the click
Typical online metrics have focused on the “final click”—the last action a user takes based on an ad. But ad technology providers want to get advertisers away from this analysis and base their buys on the path that the user takes to get to that final action. They also want to be able to gauge the longer-term implications of this final click in terms of loyalty and lifetime value.
Nowhere is this more pertinent than in search, where text-based ads may drive a user to the site, but the value of that ad can’t be judged simply on the first click nor the last click. The first click—going from search to the advertiser’s site—may ensure the value of the ad, but it doesn’t take into account whether the person clicking on the ad is a qualified prospect. The final click—an actual purchase or other necessary action—might identify the conversion, but it doesn’t necessarily gauge the process the consumer went through to get to that stage.
So when a more-involved sale gets promoted via search, click-throughs can’t be the only KPI. For its client Volvo, Media Contacts, an interactive arm of Havas Digital, had placed ads across Google, Yahoo and Bing, as well as secondary sites using search syndicator adMarketplace. It not only evaluated the results based on front-end behavior such as click-throughs and impressions, but incorporated back-end activities into the analytics. These included whether users opted to take advantage of Volvo’s online tools to locate a dealer, build their own Volvo, download a brochure, sign up for a test drive or request a quote, says Tanya Sabharwal, search engine marketing specialist at Media Contacts. The secondary site program allowed real-time adjustment of placements based on these metrics. “adMarketplace continued to show improved performance and decreased our cost per key performance indicators based on optimizations they’d done, providing us with richer data spelling out week over week performance,” adds Sabharwal.
But what of brand ads?
It’s not that brand advertisers are abandoning digital. In fact, the opposite is true. As consumers have integrated the Internet more into their lives, brand advertisers have looked to digital for messaging and building relationships. The Cambridge Group’s 2011 Advertising Outlook found that 54 percent of advertisers surveyed believe online ads are highly effective at “enhancing brand/product image,” compared to just 38 percent of respondents in 2008.
The challenge now is to gauge the impact of these digital brand ads in real time so that media can be bought in real-time bids. Response-based metrics such as impressions, click-throughs and engagements are being used more because they’re convenient than because they’re valid. “Anytime something new comes up, people feel naked without a metric,” says Dan Beltramo, CEO of brand analytics provider Vizu. “Click-through rates are misapplied to brand [ads]. They’re still being used just because they’re there.”
As a result, Vizu created “Brand Lift,” a real-time method for measuring the percentage increase in the primary marketing objective of a brand advertising campaign.
Balancing engagement and privacy
The push and pull between ad personalization and privacy remains a concern in this new era of real-time analysis. Some users may appreciate being reminded of what they were shopping for—not all online purchases are impulse buys—while others may feel they’re being stalked. Ideally, retargeting should mirror the offline buying decision without disclosing too much information.
One retailer noted that his company’s ads could serve up everything from the specific product shopped for to the size the person wears. But his company feels this is too much information to share and could scare off customers. Instead, retargeting ads focus on categories such as “men’s pants” to ensure relevance but not overwhelm the customer.
Other retail marketers are using online media to mirror offline behavior, building greater engagement through greater relevance. Staples has turned its weekly print circular into a virtual circular served up based on location. Radio Shack has developed a Facebook app based around its weekly specials. The challenge for some providers is to get rich media formats supported by real-time platforms.
“As [real-time bidding] inventory is more readily purchased by brand advertisers, there are limits to what you can do in some of this exchange-traded inventory,” says Rob Gatto, CEO of PointRoll, which developed the campaigns for Radio Shack and Staples. “We’re working with many of the top DSPs [demand side platforms] to enable rich media to run in those exchanges.”
Social and mobile ROI
The preponderance of traffic to social media, as well as the increasing tendency for consumers to access messaging via their mobile devices, may provide data about consumers’ emotional attachment to brands. Still, they have complicated the picture for analytics-based ad purchasing, since they don’t make it easy to track behavior.
The old formula would say that a “like” on Facebook is a solid metric. But that’s hardly the case today. “You have to be very careful in evaluating metrics around fan acquisition, like cost-per-fan, and when determining the value of a fan engagement,” says Justin Merickel, VP of marketing at Efficient Frontier. “All Facebook fans are not created equal.”
According to Merickel, social ROI should be based on how a company’s social presence drives its desired actions. But there are difficulties in measuring this, as Facebook doesn’t allow third parties to serve ads.
Similarly, analytics providers are being challenged in how to follow mobile behavior. Mobile messaging may come via an app, via an ad placed in a publisher’s app, through a game or via the mobile web. So mobile targeting needs to mirror online targeting. “If you’re playing Angry Birds, and you’re a 25-year-old guy and an ad pops up for Pampers, that’s the opposite of what we want,” says Matthew Koncz, managing director of product distribution of Fluent.
Video may be the fastest growing digital ad segment--eMarketer has projected that video ad spend will increase by 50 percent this year to $2.16 billion. Yet this is one area where real-time targeting has been slow to arrive, with limited automation of ad buys.
“A lot of media ad buys have been handled manually by the agencies,” says Mark Trefgarne, CEO of LiveRail, a provider of analytics and services for video ad placement. “They’ve been slow and cumbersome [and don’t provide] the targetability or measurability that online has the potential to give. Building out video online solutions has the potential to help those agencies do a better job for their clients and remove a lot of heavy lifting involved in large-scale video buys.”
But video analytics require engagement measurement that is different from typical online metrics. Beyond click-throughs, advertisers need to be able to assess whether the ad was watched in its entirety, whether the user paused it, or whether or not the volume was muted. Analysis of this information will be critical to the content and the placement of the ad—if it’s on a network that is generally muted (as might be the case in a business-to-business environment where it is watched in an office), then the creative needs to take this into account.
Real-time metrics are driving performance-based pricing, now used twice as frequently as CPM (Source: IAB)