The Washington Post this morning unveiled a new app enabling readers to fact-check speeches in “near real-time.” The new app, dubbed “Truth Teller, was created by the Post with funding from the John S. and James L. Knight Foundation.The concept was the brainchild of the paper’s politics editor, Steven Ginsberg, who had a light bulb moment after listening to a politician give a stump speech “filled with inaccurate statements.The Post described the initial version as a “prototype,” and said it was the first step toward creating a more robust truth-telling app that could be used during live speeches and discussions.The initial version focuses on the looming debate over tax reform and users can play videos from President Barack Obama, Speaker of the House John Boehner and other politicians and instantly see which statements are true, false or misleading. Future iterations will incorporate a broader array of political news topics. To create the prototype, the Post is utilizing several technologies combining video and audio extraction with a speech-to-text technology to search databases of facts and fact checks.
RTB-based mobile ad network AdTheorent this morning unveiled what it described as the “industry’s first real-time learning and predictive modeling platform.” Dubbed the “Real-Time Learning Machine,” the platform leverages machine-learning and predictive modeling to generate “on-the-fly” insights for mobile advertisers in real-time.During a pre-release trial over the 2012 holiday shopping season, AdTheorent said the platform delivered average engagement levels in the “200-300%” range above industry norms for some undisclosed retail brands.The platform, which was developed AdTheorent’s Chief Data Scientist Saed Sayad, analyzes 50,000 bid requests per second on a single server, filtering out bids with a low probability of click, conversion or awareness lift, the company said, adding that the system “learns from incoming bid requests and builds and modifies predictive models as it learns, applying such models in live campaigns to match each mobile advertisement with the optimum mobile impression.”The company said campaigns tested to date generated “uplifts” as high as 500% over previous norms.AdTheorent described the platform as the foundation of it’s “second generation mobile ad network.
Independent agency trading desk Accordant Media this morning unveiled a platform for buying mobile audiences via a real-time bidding. The platform, dubbed Accordant Mobile, enables advertisers and media buyers to trade mobile as part of an integrated RTB process, including targeting, tracking, reporting and machine-based optimization. Accordant said it had been developing the mobile platform over the past year, and made it active with a number of clients in August. The platform enables users to bid for real-time impressions on both mobile Web and in-app mobile inventory. Accordant VP-Platforms James Rooney said the platform utilizes audience-targeting data derived from “the device ID” and the IP source to create a “short-lived” user impression. One of the major obstacles in developing a robust programmatic marketplace for mobile ad inventory has been the inability to effectively target mobile users with cookie-based audience profile data. Rooney said Accordant will “build” on the rudimentary targeting features of its mobile platform, and that in-app inventory will enable advertisers to utilize a “permanent ID” associated with mobile users.
Google product listing ad (PLA) share, as a percent of total search clicks, rose 210% in 2012, according to Marin Software, but the real insight will come through a deeper understanding of intent from the clicks. The percent of PLA impressions vs. text ads rose 60%, leading advertisers to increase their share of search budgets directed toward the image ads by nearly 600% in Q4. Google transitioned Google Shopping from a free to paid model in October. The transition to a paid from a free service made advertisers pay more attention to clicks and what they pay, according to Marin SVP Matt Lawson. Consumers clicking on image ads reveal much more information about themselves than those clicking on text ads. Intent and interest in specific products or brands become clearer. For example, image ads support long-tail marketing strategies because they are specific to a product or services. Text-based ads are typically more generic and based on quality, price, and selection, Lawson said. "Both try to capture a slightly different intent," he said. Marketers could collect the data from PLA clicks to build specific audience segments for retargeting. It's not clear whether Google uses the PLA click data to get smarter about ads it serves cross-device: desktop, tablets and smartphones. Similarly, Rimm-Kafmann's Group's Q4 2012 report notes that PLAs generated 28% of Google non-brand clicks for the quarter. PLA cost per clicks came in at 26% lower than CPCs for competitive text ads. The research shows a larger-than-average gap between bids and CPCs paid for PLAs, suggesting light competition for the new format during the quarter. Marin drives $4 billion in annual ad spend through its tech platform. The company claims to manage about $1 in every $10 spent on Google AdWords.
adBrite, an online ad exchange founded in 2002, Monday informed advertisers that it would cease operations on Feb. 1. The San Francisco-based company attributed the decision to “market conditions” and “certain liabilities.”
In 2012, we saw the incredible power of data to drive elections, and to help enable the next generation of hyper-efficient marketing. This year promises to fully demonstrate how big data, small data or just the right amount of data is driving online advertising success for brands and marketers. Right now is a great time for a gut check to reassess your strategy and audit how you’re managing the process. To help you along, here is an overview of how to look at your data management platform (DMP) and what you could (and should) be doing to make the most of your investment. We need to consider what the top agencies and brands are doing and follow their example. The leading players are leveraging data management platforms to address three specific purposes: 1) To enable marketers to integrate audience, contextual and geographic data across a company’s multiple touchpoints; 2) To enable a company to segment their audience and take immediate action on these segments; 3) To enable a company to discover new insights about their audience. Those who want to lead and not follow in the new techno-marketing era must ensure they are able to deliver on these abilities and they must ensure that their DMP is helping them get there. Integration To achieve the first purpose of integration requires a robust set of import capabilities from real-time pixels (also known as Tag Management) as well as batch files. Once the data is on-boarded, the system allows a marketer to transform the information, by bucketing transaction values from a shopping cart into a set of marketer-defined buckets. Transforming the incoming data enables downstream processes of segmentation and targeting to be faster and easier by standardizing the raw event data into more usable information. Segmentation and Action To achieve the second purpose, a DMP requires an easy-to-use interface for manually defining specific audience segments, such as combining third-party demographic information with first-party registration, activity and customer value information. These segments should be expandable through lookalike modeling to find similar audiences that are likely to behave in a similar manner to the “seed” audience segment. Many DMPs stop here. However, it is important to have real-time feedback to two additional criteria before generating the audience. First is the audience size. Marketers do not have the resources to craft specific campaigns to only a handful of consumers. Accordingly, a DMP should provide the feedback to the total addressable market that matches the segmentation rules both in terms of unique user counts and in terms of frequency of exposures at given price points. Even if a segment of users is large enough to warrant a campaign, a DMP should provide insight into the cost of reaching this audience with an increasing frequency of exposure. Stand-alone DMPs that do not have tight integrations into action-taking systems, such as Web site content management systems or ad-serving systems, are often unable to provide this feedback. Another important aspect of defining the audience segment, especially for digital channels, is the latency from creating the segment to being able to target this segment. Since the anonymous user identifiers used to target segments frequently change, to increase the latency between defining a segment and synching this information with an action-taking system would lead to a higher overstatement by the original forecast of available unique users. The lack of accurately understanding the number of unique users, the cost of reaching them in the desired context and the frequency are what we call premature “audience activation.” New Insights The third purpose of a DMP is to uncover new insights about a marketer’s audience segments. The DMP should provide easy access to understand what other attributes index most and least highly with the selected audience group. The platform should also recommend the best Web sites and content to reach that audience, as well as information about the best geographies to target. Furthermore, it should ensure that a marketer understands the cost implications of overlaying geographic, third-party attributes and contextual data with forecasting the cost of various advertisement placements. The online advertising ecosystem has never been more competitive. With consumers’ activity and attention increasingly fragmented across a variety of channels, such as video, mobile and social, it is of the utmost importance that marketers today have the technologies in place to ensure they fully understand this activity as they plan their segmentation and targeting strategies. Armed with a robust DMP, marketers are better able to locate, target and understand their desired audiences, ultimately increasing the effectiveness of their campaigns and improving ROI. By understanding and utilizing the capabilities provided by a complete DMP, marketers can ensure they are using the best planning and segmentation tools to achieve their goals.
If you keep your eyes just on the ad spend projections, the sheer velocity of usage statistics, and the panel after panel that we ourselves and many others put on stage at the countless industry events, mobile looks like the place to be alright. The supply side of every emerging platform is always about cheerleading, of course. But even on the buy side, agencies are themselves looking for the next big thing to sell to their clients. Keeping the hype, well, hyped serves a lot of interests. But it is always worthwhile to get outside that bubble of self-congratulatory fervor and listen to the media buyers and planners when they are not in a mobile-centric mood. At Thursday’s OMMA RTB event, the lead panel of agency honchos didn’t even engage the mobile question until an attendee asked. “Ewww,” responded moderator Joanna O’Connell, Forrester’s senior analyst, who admitted that she is trying to resolve this year to stop beating up on mobile’s weaknesses. Still, she said, “Mobile is a mess.” Concurring with Joanna’s “Urggggg” ( I am not doing the sound justice), however, Andy Chapman, leader, digital investment, Mindshare NA, said, “I would groan a little longer.” Agreeing with conventional wisdom about the huge opportunity in mobile, all three panelists voiced persistent concerns with where the industry is now. “We are all fumbling in the dark figuring out how to make this work,” added Chapman. He agrees that clients and agencies need to do a better job of putting “mobile up front and into the overall strategy.” But just as important is the mobile experience, which he argues is just porting the Web to a handset. Michael Lampert, VP, group director, media at Digitas, says that his teams are trying to get beyond tracking devices and instead figure out changing consumption patterns to see where they need to be. If the metrics aren’t there, that is not a reason not to do it, he says. They have already discovered that the value of “serving just ten display images is not as valuable as dispaying five display images and five mobile images.” “Users are there, advertisers are not…yet,” says Jeremy Hlavacek, VP strategy and business operations, Varick Media Management. Creative units and data are the major hurdles. They all seem to agree that the buy side needs to help the medium become as “smart and sophisticated” as the desktop -- even though the mobile platforms are far behind the Web in metrics and creativity. As much as they all agree that diving into mobile is inevitable and advisable, I think it is that first gut response from Joanna and Andy that tells us a lot about where mobile is for many on the client and agency side, especially on the tiers above those who have been assigned to specialize with mobile. It feels like a mess to many of them. Specifically, it sounds like they need more cases that go beyond illustrating the predictably strong results from isolated campaigns. They need more cross-platform studies that show how mobile is amplifying and improving the investments they are already making elsewhere. And perhaps they need more cases that focus more on integration within the plan -- how, when, where and to what effect mobile was a part of the initial concept of a broader cross-media effort. You can see the full video of Joanna’s panel here.