Publishing companies have started uniting into publisher-controlled ad networks, which covers both web and mobile users. In this cXense whitepaper we present this concept, one active ad network and its initial business results.
A cXense WhitepaperPublisher-Controlled Ad Networks
For more than a decade, publishers have been searching for a solution to the overarching strategicproblem of decreasing revenue stemming from reduced print circulations. The much sought-after totalonline ad revenues have been steadily increasing, but new players have arrived from unexpectedcorners with innovative products to dominate and control these new revenue streams. Recently, newdigital business technology has appeared that enables media companies to counter these trends, andcreate their own complete digital advertising environments. To optimize these digital businesses,publishing companies have started uniting into publisher-controlled ad networks, which covers bothweb and mobile users. In this whitepaper we both present this concept, and one such active adnetwork with its initial business results. INTRODUCTION In the US, only five companies now control 65% of all online ad revenues. None of them are from the traditional publishing industry. They have originated in technology, but are now by all means competing in the media value chain. For publishers, managing As the online content offerings have developed, the online ad network relationships market has become crowded and complex with different effectively is becoming a technology-driven media middlemen. At present, these middlemen competitive differentiator. sit in the value chain between the publishers and advertisers, and New technologies also are creating opportunities for reap un-proportionally large shares of the overall online revenue, publishers to form – not thanks to solutions that help them understand the end users and just join – ad networks. exploit the corresponding data to arbitrage on ad prices through better targeting. So far, publishers have been at a disadvantage Media Management Centre, Kellogg School of Business and at a loss at how to regain control over their online advertising business. Now, new pure-play technology companies, with no intentions of intercepting the media value chain, are appearing as a reaction to this situation. They offer publishers complete solutions across the whole online ad value chain, and thus help media companies eliminate the costly middlemen. With this technology, several publishing companies have taken the challenge to regain control over their digital business. They now offer and deliver a highly attractive, self- made proposal to advertisers: large and sufficient inventory for online ads (which means sufficient traffic volumes), effective understanding of content consumption and user behavior, and an effective matching function where the best ads are displayed to the right users to maximize relevance for the end user, as well as ROI and revenues for the advertiser and publisher, respectively. These networked publishers use advanced and integrated tools that enable them to monitor, analyze, model, and predict patterns of content and users behavior, through a dynamic, real- time closed loop that connects the information about their content, their users, and their ads. Through this improved visibility into and understanding of the content and users intent, A cXense Whitepaper Publisher Controlled Ad Networks 2
publishers are able to utilize their traffic inventory optimally for ad placement. This is a criticalsuccess factor for any ad network operation.However, publishers also need to have a critical mass of inventory and traffic, so that they canattract a critical mass of advertisers. To achieve sufficient traffic, these companies have unitedinto publisher-controlled ad networks.CASE STUDY: PUBLISHER-CONTROLLED AD NETWORKScXense works with a number of such ad networks worldwide. The ad network we describe inthis white paper is formed by a group of publishers consisting of all the largest newspapers inone European country. The network represents more than 60 sites including national, regional,and local newspapers, as well as other online offerings like classified, search, and verticals. Thetotal online traffic for the group is more than 1 billion page impressions per month.The different partners in this ad network have previously used several third party network-based platforms for their search- and performance-based ads. One of the main drivers for achange to their own network was their desire to gain andkeep with the publishers the knowledge about the readers Network Propertiesand the readers‘ navigation patterns. The combined user 60 sitesknowledge in this group is broad and deep, since most people Both newspapers and classifieds 1 billion page views per monthin the country leave implicit signs of their preferences on the 75% of online users in the countryclassifieds site; they also daily use one or more of the online Committed vertical sitesnewspapers in the group, implicitly marking their interestthrough the chosen newspaper articles and vertical sections. In addition, the new platformallows a collection of user data based on their devices, location, social media participation, etc.A central aspect of this network is its size. A high percentage of publisher traffic (more than75% of all of the online users in this country) is available from the start through initial trafficcommitment from participating publishers; over time, there is a considerable revenue upsidethrough publishers’ plans to provide additional traffic to the ad network, and to expand the adnetwork to new types of content sites. Already, the technology this publisher group usescollects data from a wide variety of external public sites in the country. This enables them tobuild trustworthy long- and short-term user profiles, since every new source of contentincreases the understanding of user’s interest and intent.To build an attractive ad-serving base and high match rates, the publishers also benefit fromhaving a good base of relevant vertical content pages to enable even higher quality userprofiles – in addition to sufficient traffic. Hence, to achieve this, the publisher network has alsocommitted their vertically focused sites to this new solution. Already, their analytics showsignificantly improved match rates and relevance leveraging the user profiles and semanticcontent profiles, resulting in increased ad click-thru rates (CTR). For example, a leading traveloperator in this country reports a four-times increase of conversion and ROI on its ads with thenew ad network, as compared to the conventional third-party ad network.A cXense WhitepaperPublisher Controlled Ad Networks 3
High CTR leads, in turn, to higher CPC pricing for ads. In the very beginning, the eCPM from thenetwork is expected to be somewhat below that of the eCPM for the corresponding ads servedby the dominant ad players, due to the fact that these players have been in the market for along period, and have had a correspondingly longer period to develop the commercial factorsin their favor (large traffic, many advertisers, and high click prices). However, in a relativelyshort amount of time, the publisher-controlled network catches up on these factors. As theirinventory grows, their user profiles improve, and their network attracts more advertisers, theprice auction system will mature to same or higher levels of eCPM. Hence, the net advertiservalue over time exceeds revenue offered by traditional online ad models. In addition, thepublishers will retain full control of their online ad business.The publishers also have access to an advanced semantic content and ad analysis tool, allowingthem to survey and optimize both the content and the ads relative to user behavior, and todeduce important user behavior patterns. This data is used in a dynamic closed loop system sothat ad placement can be planned in real-time; thus they optimize their relevance and revenueon the whole time spectrum: from the long-term to the extremely short-term.TECHNOLOGY AND OPERATIONAL ASPECTSThe solution these publishers have chosen is a cloud-based solution, user friendly and flexible,and which can be in operation within a few days. It is highly scalable, since it requiresresponses within milliseconds for millions and millions of users, and is also highly secure andreliable. It uses highly advanced linguistic, semantic, and statistical analysis tools that recognizethe overarching meaning and sentiment of the content articles, as well as gathers informationabout the readers and their interests and behavior across the publisher sites. The system isextended continually with world-leading academic research results from these areas.Technically, user interest is, for example, clustered in a set of categories and associated withthe individual users’ long-term profiles. Categories include health, car and motor, travel,celebrities, telecom, food and restaurants, sport as training, gadgets, etc. Categories arefurther divided into sub-categories, and associated with meta-values; these meta-values aretypically matched with ads from advertisers. The content is also automatically profiled andcategorized. Furthermore, a wide range of contextual information is taken into account,including location, device, time-of-day, etc.The system is easy to integrate in the publishers business, both technically and functionally. Interms of operations, the publishers no longer need to handle a multitude of ad networks andexchanges. The small and focused team operating the ad network handles all kind of ads forthem, from agencies, direct customers, to direct downloads for other search ads providers.ROI ASPECTSThe projected business model for this publisher ad network shows an expected ad revenuegrowth for the participating publishers of 5 to 8 times over the next three years, and theA cXense WhitepaperPublisher Controlled Ad Networks 4
network expects to gain more than 20% market share in the performance-based ad sector inthis country. This growth is to a large extent based on greater inventory, higher ad relevanceleading to higher ad pricing, and on a better revenue share model than is typically offered bythe incumbent players.The revenue share that publishers gain from external networks is typically non-disclosedinformation, but in most cases in the traditional model the publishers’ share represents lessthan 50% of the gross invoiced advertising value. In the new model presented in this whitepaper, the publishers retain 70-80% of the total revenue. In addition, they retain the strategiccontrol position towards the advertisers, and the data and knowledge about all the users onlinebehavior can be leveraged by the publishers, so they stay in control of their future.Within the network, publishers are free to experiment with their own ad-serving strategies toincrease revenues. We have observed significant business innovation related to ads and userpatterns. Once they realize the potential of user and content data, publishers show greatinventiveness in how to optimally use the powerful dynamicanalytics and serving tools in the new solution. The trends we ROI advantagehave observed are extremely encouraging for the digitalfuture of these businesses; they show that with the right 5-8x revenue growth over 3 years 70-80% revenue shareinsight in the digital user, the publishers are eminently able to 20% of market share in thetake advantage of their industry expertise and bring the ad performance ad sectorbusiness back to their own territory. Based on their historical New ad serving strategies media savvy, they are the ones who can best interpret theemerging content consumption patterns. The generic ad players have been out-competingthem in the digital space; but now, the publishers can see and explore new user patterns andad models themselves and learn, in real-time, about their effectiveness and monetary results.They publishers can build their proprietary ad-related IP at impressive rates. As one example of such digital stratagem, one of the networked publishers has shown thatthey can better exploit the unattractive display ad inventory on the high-traffic generic pages.Because of the generic profile of these pages, it is typically hard to achieve high CPM rates onthese pages. However, with the right user and content data, the system can now place highlyrelevant ads for each specific user even on these sites. These precise targeting methods haveenabled the publisher to increase the CPM price for targeted ads 2-5 times as compared tocurrent CPM price for a run on a generic network ad. By using only 10% of display ad inventoryon a specific ad placement, this publisher has increased the ad revenue on these generic adplaces with up to 50% of gross revenue. That is a significant growth for a very limited risk.Over the past two years, we have worked with eight other publisher-controlled ad networksaround the world. They all show similar projections and early performance improvements. Byusing closed-loop targeting technology they have all achieved increased ad revenue, throughhigher volumes, increased CTR, higher prices, and better inventory utilization. The targeted adserver tools, with associated content and user analysis, have helped them all improve efficiencyin exploitation of their inventory.A cXense WhitepaperPublisher Controlled Ad Networks 5