July 24th, 2007
Following up on my post about the poor state of online ad relevancy, I’ve been digging deeper with some of the companies that are working to improve that state, which lead to the following interview with Larry Allen, SVP, Business Development and Marketing at TACODA. Believe it or not, we completed this interview before the announcement that AOL has acquired TACODA. Relevancy is in the air (and so are ad networks).
Publishing 2.0: How would you assess the current state of relevancy across all online display advertising? Where/how are display ads being served with the greatest relevancy? Where/how are display ads being served with the poorest relevancy?
Larry Allen: In general, marketers are still using context to drive most of their online buying decisions for media. This creates some challenges as it negates relevancy for many visitors to a particular page. There is still a large gap between the volume of ad inventory on the Web, especially when you consider news and social media, and the amount of ad dollars being spent. The earliest adopters of online media have been direct response focused clients willing to pay only on a lead or transaction basis, which has promoted massive amounts of untargeted advertising to run freely across the Web.
P2: How does behavioral targeting for online advertising work?
Allen: TACODA observes anonymous online behaviors across a wide network of publisher partners in order to determine active and passive interests of visitors in aggregate. These interests are then packaged into targetable and predictable audience segments. These pre-qualified segments are then sold to advertisers to promote relevant products and services to those individuals. TACODA’s publisher partners contribute ad inventory that is evaluated, and then appropriate advertisements are delivered based on the audience segmentation of the visitors visiting those Web page.
P2: How has behavioral targeting improved relevancy for online display ads across websites using the technology? Are there elements of Tacoda’s system that make it uniquely able to improve relevancy?
Allen: TACODA has shown on an advertiser-by-advertiser basis that the consumers reached in aggregate for their campaigns are more likely to be interested in their products than a typical run-of-site or run-of-network advertising campaign. TACODA is uniquely positioned to provide insights about how various audience segments respond to creative messages and products in an effort to improve relevancy over time. One major challenge we as an industry face is the massive amount of ad inventory on the Web, it is a daunting challenge to dramatically improve overall advertising relevancy given the number of advertisements required to fill every available impression. TACODA is trying to improve this one ad impression at a time.
P2: How is TACODA “uniquely positioned” relative to other behavioral targeting companies?
Allen: Other BT companies do not provide publishers or advertisers with the advanced analytics or deep behavioral audience information. These analytics provide advertisers with a blueprint for the discrete audience segments that deliver the highest quality results.
P2: It seems that the surplus ad inventory problem reflects the current display ad economy, which is based on page views and impressions rather than on people. BT overcomes one major aspect of this problem by focusing on people and what their behaviors their intentions and interests. But doesn’t valuing display ads based on impressions still reinforce the whole ad inventory problem by forcing publishers to value their content based on how many impressions it generates? Do you agree, and how much is this an issue for TACODA, i.e. how much does TACODA work with advertisers on a CPM basis vs. cost-per-click or cost-per-action?
Allen: TACODA only works with advertisers on a CPM model. However, marketers are very aware of the back end metrics and how that CPM converts to ROI for their business. Large auto manufactures or CPG companies are looking at brand metrics like engagement and consideration. These metrics do not lend themselves to be purchased on a CPA or CPC basis. TACODA is about delivering a marketer the right audience to give their message the best chance at resonating and driving to a sale, whether online or offline.
P2: For large publishers like the New York Times, how would you compare the relevancy of ads served by the TACODA network to ads served by other ad networks and to ads sold through the publisher’s ad sales force?
Allen: TACODA is delivering ads to consumers based on the targeted segments prescribed based on intent and previous behaviors. This creates a significantly higher level of relevancy for the consumer over contextual and other optimization techniques. Other ad networks use creative and placement optimization to make ad delivery decisions, which tend to show ads to people based solely on their willingness to click on ads. Our research has shown that the majority of click responders come from a very small percentage of the overall online audience. Therefore, this method forces advertising to be less relevant for the masses.
When publishers are selling their own property, they are often selling the demographics of their audience and the context of their site to marketers that would be more relevant to their readers. The challenge is most web sites have more inventory than their sales team can sell directly, which creates the need to fill the inventory with a paid advertiser regardless of relevancy. This is especially true for news, social or sports sites, which tend to have less obvious commercial relevancy to advertisers. Auto and Financial sites tend to be in high demand, creating sold out scenarios and generally a higher level of relevancy.
P2: The mix content with high commercial value and content with low commercial value on a newspaper site, for example, is the legacy of the previous “bundled” media age, when some sections subsidized others. In the digital age, where each unit of content stands — and has to be monetized — on its own, is there now an incentive for publishers to only publish content with high commercial value? Or can BT essentially provide the same subsidy by allowing publishers to more effectively monetize content with low commercial value?
Allen: We believe that BT can provide publishers with the ability to monetize the undervalued pages that consumers find very important. There will continue to be fragmentation in the media business, and publishers will need a variety of ways to monetize their pages in order to offer content that satisfies the demands of their consumers.
P2: How would you compare the relevancy of behaviorally targeted display ads to that of keyword-driven contextual text ads? Is there a difference in relevancy between ads served contextually on search result pages and ads served contextually on content sites (vs. behaviorally targeted ads)?
Allen: There is definitely a difference. Contextually targeted ads require the content on the page to provide some immediate clues into the interests of the user. If one is reading about a plane crash you certainly don’t want to be showing text ads for travel. One of the clear benefits of behavioral targeting is that it uses historical references to produce more relevant advertising, and it can also be predictive based on look-a-like modeling. BT creates the best opportunity to show relevant ads to consumers regardless of the context of the page, especially in those environments that are less commercially relevant, like weather or news.
P2: Can you explain “look-alike modeling”?
Allen: We are able to build a blue print of those audiences that are working for an advertiser by evaluating their behavioral traits and then create segments of new consumers that exhibit similar traits online.
P2: It seems that the strength of behavioral targeting is for contexts that have little or no commercial relevance. From that perspective, can BT be viewed as a complement to contextual targeting, e.g. if someone has viewed travel content and then goes to view auto content, it’s better to show a contextually targeted auto ad on the auto content than a travel ad on the auto content, i.e. better to save the behaviorally targeted travel ad for when the user is viewing content with low commercial value? Or does behavioral targeting play a role in targeting ads in contexts with high commercial value?
Allen: We certainly do not see BT replacing large branded sponsorships within contextual environments. BT offers marketers a way to scale reach and frequency of those targeted audiences outside of context in a very efficient way. We do work with a number of highly relevant context sites in our network like auto that find value in our service. Every site, even the most relevant, have some level of undersold inventory due to a variety of reasons.
P2: Where do you see the greatest opportunity for improving online display ad relevancy?
Allen: As more publishers and marketers adopt networked behavioral targeting the level of relevant ads online will naturally increase. There is great opportunity to improve ad relevancy in general but especially in contextual environments that do not directly correlate to an advertising category; news, weather and sports come to mind immediately.
P2: Why do you think some publishers have been slow to adopt BT?
Allen: The largest reason is scale – most publishers have limited information on their audiences available for BT. We have seen very fast adoption of our networked behavioral model across the largest media companies as we provide the best solution for their needs.
P2: What are the greatest barriers to improving display ad relevancy, including data, technology, and human resources (e.g. for creating customized ad creative)? How do you think these barriers can be overcome?
Allen: TACODA is focused on creating a platform that scales and improves over time to provide the most relevant online display ads to consumers. The greatest barrier is the ever growing amount of inventory and the marketers’ adoption of behavioral and other targeting methods.
P2: Why do you think some marketers have been slow to adopt BT?
Allen: It is simply about education and comfort. Most media buyers have been trained that the internet is a direct response medium only. As an industry, we need to be educating the market about how it can be used for brand building. You are starting to see more major marketers adopt behavioral to drive awareness and consideration, from Pepsi to P&G.
P2: What will be the next big advancements in ad customization?
Allen: Predictive audience segments creating increased opportunity for ad relevance, Dynamic video creative insertion, dynamic creative messaging.
P2: How do you think the trade-off between privacy and ad relevancy/customization can best be addressed?
Allen: TACODA believes that the consumer must be in control and have a choice when it comes to their privacy. As long as publishers and ad networks are clear and concise with their notifications on privacy, consumers will be less likely to have concern. Consumers want more relevant advertising and are willing to give up some level or privacy to ensure that. As long as there is a clear method for opting out of the program, consumers generally do not have a problem.