Behavioral Targeting

September 24, 2010 In Ad networks, Advertising, Behavioral Targeting, Display Advertising, Marketing, Online Advertising, Online Marketing, Research By Tony Winders

Study highlights efficacy of display advertising

comscore logoThis week comScore and ValueClick Media released a white paper I co-authored titled “When Money Moves to Digital, Where Should It Go? Identifying the right media-placement strategies for digital display.”  The study was based on over 100 campaigns run on the ValueClick Media network and used comScore’s Action Lift reporting methodology to evaluate the impact of various display placement and pricing strategies on site visitation and search behavior .

                       Read the press release.                        Download the paper.

While cross-media and search vs. display has been studied previously, to our knowledge this is the first quantitative analysis of the view-through effectiveness of various display placement strategies (Audience Targeting, Contextual Targeting, Retargeting) and pricing options (Premium/CPM, Efficiency/CPC). And while the study is by no means exhaustive, and may even raise more questions than it answers, it validates the efficacy of display advertising and provides direction on which placement strategies to deploy to drive advertiser performance, whether measured by brand or direct response metrics.

Here are a few of the key findings:

  • Retargeting works extremely well and should be considered for both direct response and branding initiatives
  • Some placements create new traffic while others find audiences that are already engaged with the subject
  • Marketers in different industries take advantage of the strategies differently
  • All multi-strategy marketers increased site visitation above the norm
  • Advertisers who used three or more placement strategies tended to have one metric in which they disproportionately beat the norm

While the study was exclusively on placements across the ValueClick Media network, it was not sponsored research, but rather done in partnership with comScore. Regardless, the campaigns were all sold and delivered on a single network, so a different network or analysis across an entire campaign may have shown different results.

That said, the data tells an interesting story about ValueClick Media. Most surprising to me was the lift, however small, driven by the RON baseline, something we attribute to the impact data and optimization is increasingly having on the ability to predict the performance of every impression — a huge validation of our technology. Contextual pricing probably skewed high due to inclusion of campaigns run on our more exclusive vertical networks. And the data suggests we just may be undervaluing our audience targeting and retargeting inventory.

Laurie Sullivan did a great job of recapping the results of the paper in her MediaPost article published today. I will be presenting the research with Anne Hunter on Tuesday at the IAB MIXX conference in NY. Our workshop is up against sponsored content from Google/AdMob, Yahoo! and AOL, so I’m eager to see what audience we attract. Whoever comes, I’m looking forward to seeing everyone during Advertising Week in NY!

0
July 24, 2010 In Ad networks, Advertising, Behavioral Targeting, Data, Los Angeles, Marketing Automation, Online Advertising, Online Marketing By Tony Winders

Fixing Advertising Los Angeles | July 2010

CaptureRarely does a panel discussion achieve an optimal mix of education, entertainment and controversy the way the Fixing Advertising session did on Monday night in Los Angeles. The education series, sponsored by Dapper, has now made its way to every major U.S. media market in an effort to not only make sense of the fragmented display advertising landscape, but to actually do something about it. Bravo!

Credit for the effective cadence and tone of the session goes to Pete Kim, General Manager of Yahoo! SmartAds, who clearly being qualified to serve on the panel himself had the audience on the edge of their seats just by knowing just when to dial up and down the intensity. He began by asking the audience what level of discussion they wanted to hear, which was promptly met with shouts of “deep dive” and “hard core!” And the panel did not disappoint. 

So what is being done to “fix” advertising? Kim began the conversation by asking the panel to articulate what’s broken — and everyone had an opinion about given the theme of the evening. 

According to Zack Coelius of Triggit, advertising is broken because it’s fragmented. It’s broken when it costs 20 to 30 percent of the media budget just to plan, buy and manage the process, especially when you don’t know where your ads are going and when you don’t have control over the buy.

Frank Adante of Rubicon echoed his sentiment from the publisher perspective, explaining how the sell side is fragmented too. Finding the money is difficult, he says, when an estimated two million advertisers are buying from one thousand sources and at least 500 sales teams worldwide. Not to mention the difficulties presented by multiple billing, collections, reporting systems. The solution, he suggests, is a central platform for selling and the  need for automation.

According to Jon Aizen of Dapper, people enjoy the web and get great value, but they don’t like display ads, which cover 10-20% of the visual real estate. After all, banners haven’t changed much since 1994. Advertisers actually have sometthing people want, but the advertising isn’t reflective of their offering, which Dapper hopes to change by matching visitors to relevant content.

Amy Lehman of United Online made a compelling case for how expensive it is to manage campaign reporting, metrics and attribution and how insane it is that we have not dealt with this as an industry already. Furthermore, she said, the industry is “beyond commodotized” and we make enable fragmentation which only makes it harder on ourselves. Ultimately, according to Lehman, most facets of the fragmented ecosystem (analytics, rich media, creative optimization, ad verification, etc.) belong resident in the ad server. 

As automation of these processes takes hold, will jobs actually be eliminated? Probably not, since machines can’t do creative or strategy, but more junior level roles centered around manually running reports and  managing pivot tables may evaborate, or at least their jobs will change, as the industry continues its rapid trend toward automation. According to Adante, automation is partly the cause of the fragmentation, referencing how a huge SEM/SEO services industry was built upon the backs of the major search engines.

Jon Aizen spoke about page saturation, consumer immunity (banner blindness). Unlike how a half page ad in print is half the cost of a full page, more ads on the page online are sold at the same rate, thus creating banner blindness. In Aizen’s view, sometimes it is more prudent to know when not to serve an ad. He also claims display units are too small and not intrusive enough. After all, when was the last time a banner ad made you laugh or cry?

The days of arbitrage models where middlemen add no value are over. If Terrence Kawaja’s now infamous GCA Savvian fragmentation slide is an indicator of some future consolidation, the Kim asks “by whom?” According to Coelius, “it’s going to be a going out of business process, not a buying process.” For those vendors who help to add insight and extract real value for advertisers, however, the outcome may certainly be acquisition by those larger media players and agencies who must continue acquiring such technology to compete long term.

Partly justifying the need for so much data and analytics is how much more multi-dimensional and dynamic display is compared to search. The mere fact that campaigns are distributed among thousands of sites in and of itself is complex. Then add in the critical creative component, which according to Michael Baker, a recent DataXu study found was the single most important factor in driving conversions, followed by consumer and context.

Being hosted at The Rubicon Project, Adante diligently represented the voice of publishers, whose role in all of this cannot be overlooked. According to some, a publisher backlash related to how networks use their data is looming, but there shouldn’t be any at all if publishers are simply paid for each impression based on what it is worth to the advertiser, which is what DSPs and sophisticated ad networks are set up to do.

Best quotes of the evening:

“We’re trying to kill online advertising and replace it with content.” – Jon Aizen

“Arbitrage just needs to die.” –  Zach Coelius

 “These little buy and sell side technologies are like plaque in the teeth of Google.” – Michael Baker

DSC_0842

From left to right: Amy Lehman, SVP Advertising, United Online; Zach Coelius, Founder & CEO, Triggit; Frank Addante, Founder & CEO, The Rubicon Project; Jon Aizen, COO & Co-Founder, Dapper; Peter Kim, General Manager of Yahoo! SmartAds; Michael Baker, CEO, DataXu, James Beriker, CEO, Dapper.

0
July 3, 2010 In Ad networks, Advertising, Behavioral Targeting, Brand Marketing, Online Advertising, Online Marketing, Research By Tony Winders

A way forward for brands online: Audience Measurement 5.0

arf logoI’m no researcher, but in recent years have acquired an appreciation for the role of research in audience targeting and measurement as well as its effectiveness in B2B marketing. Last week I had a rare opportunity to rub elbows with some of the digital media industry’s top research experts during the Advertising Research Foundation’s Audience Measurement 5.0 conference in New York. As part of the “Media Smackdown” track, I was there to co-present new research on “Media Placement Strategy and its Impact on Online Ad Effectiveness” with Anne Hunter, VP of advertising effectiveness at comScore.

Notwithstanding how excited I was to participate in the conference and introduce the research ValueClick Media and comScore will release in the coming weeks, I took the opportunity to revisit my own understanding of the state of online measurement and targeting. The main question on my mind: how far away are we from seeing brands invest more heavily in digital, which today accounts for a disproportionate +/- six percent of total media spending? Through my conversations, I reaffirmed my understanding of what’s possible today, and I got fairly consistent feedback on where things need to go in order to move us toward the measurement and targeting capabilities necessary to give brand advertisers the confidence to spend more online.

Randy Cohen, founder and CEO of Advertiser Perceptions, reinforced what comScore suggested in its 2008 “Whither the Click” research, when he reminded me that “we’re chasing the wrong metric in performance.” The real money, he says, is in upper funnel measurement and our ability to prove that online advertising is capable of moving consumers into the consideration set for a particular brand. More succinctly put, he suggests “return on ad spend is a metric, but what matters most is return on brand.” Although he predicts innovative vendors will partner to deliver breakthroughs in audience targeting and measurement brands can embrace, like most people I spoke with, he doubts a single metric for brand engagement is realistic.

A related theme shared throughout the week and captured best in my conversation with Jack Myers, is the need to move ROI measurement beyond online sales impact and brand lift studies and into better accountability for its impact on offline sales. While offline sales impact studies have shown proven lift from online campaigns, they are expensive and fall short of brand marketers’ true desire: to target ads on the basis of offline shopping behavior.

The ability to target based on offline sales data was first explored by Yahoo Consumer Direct (with Nielsen/Homescan) in 2003 and improvements can surely be expected through a more recent announcement of a joint venture between IRI and Nielsen. It is unclear yet how this will impact comScore’s relationship with IRI. comScore has partnerships with other offline data providers, including dunnhumbyUSA, Polk and others, which until recently existed primarily for the purpose of measuring offline sales impacts as a result of online advertising. But the more interesting development for comScore is its introduction of Audience Advantage, which combines offline data with comScore panel data and a look-alike modeling methodology to allow networks and portals to “pre-score” media for its propensity to identify consumers who have exhibited similar behavior to those who purchased particular products offline.

While Consumer Direct and Audience Advantage, with their look-alike models have been around for awhile, third party data sources for online behavioral targeting have emerged as a key component of the display advertising landscape over the past two years. To the extent these providers can deliver specific audiences who have exhibited a recent behavior with any scale, it would stand to reason that this would be the more effective targeting method – perhaps worthy of a future comparison by an innovating brand. Whatever approach is best, there are actionable brand targeting and measurement solutions emerging, which when proven, promoted, refined and repeated will be a boon for display advertising online.

My most entertaining conversation of the week, and perhaps the most insightful, was with Andy Fisher, EVP Global Data and Analytics Director for Starcom, who suggested that brands really do want to spend more online, if nothing else because they know their future job security depends on digital. If only we could demonstrate for them a clear reason to do so. Brands, he said, want to know who they are reaching. Not just what audience segment performed well for a particular metric, but the specific individuals reached by a campaign. To support this ideal, measurement and ratings vendors should strive to report on granular audience delivery metrics, much in the way they do for media measurement today. One shortcoming, it seems is the lack of a common taxonomy for measuring behavioral audience segments. Even if there were, the user profile data it would rely upon lives in the proprietary databases of advertisers and media companies, each with their slightly different description of the same users and valid reasons for not wanting to share the data freely.

I was unable to find anyone to argue in favor of context over audience, however, as our study with comScore will demonstrate, what strategy to deploy is highly dependent upon price, reach and each marketer’s specific objectives – and deploying multiple strategies may be most effective. To that end, one of the more interesting point made by Mr. Fisher was how in television, media equates to audience because all of the creative is delivered against the same media at the same time, making it easier to measure and scale. Whereas the distributed nature of online carries with it many more complexities because of the requirement to reach and measure audiences across several media placement options. I’m sure I am oversimplifying this thought and hope he will elaborate.

Another concept that arose from a guided luncheon discussion led by Ms. Hunter and reiterated by Joanne Burns, EVP of marketing, research and new media at Twentieth Television, was the need for research and marketing to work more integrally together. When speaking with brands, in fact, Robert McLoughlin, director strategic insights at AOL said his approach is to first ask “what is your research goal” and only then dig deeper into the typical line of questioning related to marketing objectives.

Although I’m sure gaps remain in my understanding of what’s possible and where things are headed, I came away from Audience Measurement 5.0 much more confidence that the metrics CPG and other brand-oriented advertisers demand is on the horizon, if not here already. Regardless, I can assure you there is an army of intelligent marketers, publishers, analysts, scientists and researchers working to ensure online advertising isn’t relegated to being exclusively a direct response medium.

0
May 5, 2010 In Ad networks, Behavioral Targeting, Online Advertising, Politics, Privacy, Uncategorized By Tony Winders

Privacy Legislation Update: Part 3

boucherThe wait is over, but not the debate. Today, U.S. Representatives Rick Boucher (VA-09), Chairman of the Subcommittee on Communications, Technology, and the Internet, and Cliff Stearns, Ranking Member of the Subcommittee, released a discussion draft of legislation to assure the privacy of information about individuals both on the Internet and offline.

Congress may never pass a law, but for now at least, it appears our collective industry efforts to educate House Engergy and Commerce Committee members has paid off. I was especially relieved to read this paragraph about a “carve-out” for ad networks, which speaks directly to the huge educational efforts put forth by the IAB, NAI and several competitors and colleagues alike over the past year:

The bill creates an exception to the opt-in consent requirement for third-party information sharing by applying opt-out consent to the sharing of an individual’s information with a third-party ad network if there is a clear, easy-to-find link to a webpage for the ad network that allows a person to edit his or her profile, and if he chooses, to opt out of having a profile, provided that the ad network does not share the individual’s information with anyone else.

Consumer privacy groups are not satisfied with the fact that the legislation, as currently drafted, would rely heavily on consumers privacy policies. Yet, it would require companies to allow consumers a higher degree of control over the collection and use of their data. This requirement would be burdensome to advertsing companies, but in the long run may yield higher performance for advertisers and a more relevant experience for consumers. 

So it appears the online advertising industry dodged a bullet for now, but there is much more education and debate to unfold in the months ahead.

Click here to read an executive summary of the draft.

Click here for the full text of the bill.

0
January 27, 2010 In Behavioral Targeting, Politics, Privacy By Tony Winders

Privacy Legislation Update: Part 2

S1CA46WV97CAI3LKQOCA20ZYTPCAOBT33FCAX162QFCA3KDPBQCAX2F0G6CAKL2C4LCACKPME0CAABJO4BCAP88BIPCAESC850CACB4QJLCAG9OEN3CAMA43G6CAPE004JCAPWN253CAH6Y0GUCAC5X6XAThere have been a few developments since my last post related to privacy legislation governing behavioral targeting, so here are a few noteworthy updates. While I thought a draft bill would have been introduced to the House Committee on Energy and Commerce before the end of 2009, I certainly did not hear anyone in the online advertising industry complaining about it.

There have been, however, plenty of complaints registered in recent months over a plan currently being negotiated in the Senate Banking Committee that would create an independent Consumer Financial Protection Agency and reauthorize the Federal Trade Commission as part of pending finance reform legislation.

The plan attracted new attention from the industry last week when the IAB and other trade groups sent a letter to the Committee warning that the FTC could become too powerful under such a plan. Pace Lattin’s view was more entertaining than most news coverage of the topic and my favorite post since the ADBUMB founder launched his new blog, Industry Pace. The current political climate is one likely to see financial reform bill pass, however both The Huffington Post  and Fox Business report that bipartisan support for financial reform may hinge on dropping the CFPA issue.

Meanwhile, efforts in the House by Reps. Boucher and Stearns to introduce consumer privacy legislation that could mandate an opt-in consent for data collection remain in hearings before the Subcommittee on Commerce, Trade and Consumer Protection, most recently Exploring the Offline and Online Collection and Use of Consumer Information on Thursday, November  19, 2009.

Kudos to the entire industry’s aggressive self-regulatory efforts in staving off any harmful legislation so far, and especially the leadership of Randall Rothenberg ,whose recent op-ed piece in The Hill did a great job of defending our collective position on the topic.

0