The Programmatic Path to Profit for Publishers

The Programmatic Path to Profit for Publishers

Title image

The Programmatic Path to Profit for Publishers

  • Add To Interests
  • PDF

  • A New Model Emerges

    Despite the growing pains, there is a new model emerging that enables publishers to better capture the programmatic opportunity and use both direct and programmatic strategies to boost sales and margins. Our workshops and interviews highlighted four key practices. While many publishers pursue some of them to some extent, only three companies in our study (12 percent) fully embrace them all. The ones that do outperform the market.

    Apply a cross-channel, data-driven strategy. As publishers look to rationalize traditional direct and programmatic sales, pricing structure and go-to-market strategies are key considerations. Forward-looking companies take a data-driven approach across both direct and programmatic channels. They do not get hung up on out-of-date distinctions; instead, they probe advertiser and user data to develop a clear view of the propositions that sell best in each channel, which then shapes both their sales and their pricing approaches.

    For these companies, programmatic is no longer just a way to sell remnant inventory—it has become a key tool for the same types of inventory that are sold directly, especially in more mature markets. “We [price] programmatic at least the same as direct,” said a senior vice president of a major US news publisher. “It’s the same inventory so it should have the same value, and this avoids any channel conflict. I don’t have remnant inventory anymore; instead, I just have inventory that I sell via different channels and models. We have seen both our revenue and CPMs grow, and our programmatic approach has been key to that.” A similar trend is at work, albeit more slowly, in less mature markets.

    The programmatic marketplace is set to accelerate as sales of programmatic-guaranteed inventory become more significant. To maximize revenue today and avoid cannibalization and price erosion in the future, publishers need to comprehensively think through pricing structure and go-to-market approaches across their direct and programmatic strategies, including their guaranteed and nonguaranteed models. One North American broadcaster aims to completely replace traditional direct sales with programmatic guaranteed within the next two years and is already working through the pricing and organization ramifications of the move.

    As channels converge, the right direct and programmatic strategies will differ by market and publisher. For example, publishers soliciting direct-response marketers require a robust programmatic capability that appeals to the aggressive approach of these advertisers. At the same time, premium websites seeking to attract high-end brand-marketing campaigns need to carefully consider how they price and expose inventory to maintain their premium status and prices. In some markets, large publishers can leverage their scale to lead the market in a direction that aligns with their strategy. In February 2015, eBay UK held a “programmatic-only week,” offering its entire inventory for sale exclusively on a programmatic basis. In addition to raising awareness of the advantages of programmatic trading, this initiative enabled the company to better understand the current role and requirements of programmatic buying, and it has since reorganized its sales teams to reflect what it learned.

    Segment and match inventory with the right buyers. Publishers can significantly enhance the value they deliver to buyers by matching advertisers with the audiences they want to reach. Smart publishers use both dialogue and data to build a clear understanding of which buyers prize which inventory and audiences, and they offer advertisers a number of different ways to target the right consumers. These include targeting based on the following:

    • The quality or position of an impression
    • An ad’s context (where on the site the ad appears or what the content of the web page relates to)
    • First-party data (the publisher’s data on audience attributes and behaviors)
    • Preferential access (often called “first look”)

    High-value targeting is typically offered to buyers through invitation-only auctions or through unreserved fixed-rate, or “preferred,” deals, in which buyer and publisher agree on a set price for a certain type of inventory. These impressions are generally sold at premium rates, reflecting their added value. This approach is rapidly growing in popularity. In Google’s ad exchange in Europe, the Middle East, and Africa, for example, unreserved fixed-rate programmatic transactions quadrupled last year and invitation-only auction transactions doubled.

    Best-in-class publishers go a step further and identify which segments work best for different advertisers. Some use special-event-based offers that help keep their media properties front of mind for buyers. Half the participants in our study regularly initiate private deals, while the rest do so only on an ad hoc basis, typically in response to an advertiser’s request. One European online publisher realizes premiums of about 40 percent for contextual advertisements, 65 percent for first-look packages, and up to 220 percent for first-party data. A UK news publisher that uses customized programmatic sales of inventory targeting its highest-value audience segments achieves CPMs up to six times those for direct sales.

    One brand campaign buyer told us, “I want to partner with publishers who have the right audience and quality of inventory—and the truth is, I work most with the publishers who are proactive in presenting these opportunities to me.”

    In a fast-evolving market, it is also important for publishers to stay close to buyers in order to anticipate their changing needs. For sales teams, this means keeping track of buyer preferences with respect to budgets, access, and relationships with trading desks. In the US, for example, trading desks at major buyers increasingly want “always on,” invitation-only deals with access to all of a publisher’s available inventory. Publishers that provide this access see higher revenue; those that don’t, struggle.

    Assemble the right technology. Publishers must navigate a complex technology landscape. Assembling the right layered collection of software—or “stack”—is critical to revenue generation, since it is the stack that helps publishers maximize revenue by providing essential inventory management and process controls, as well as the gateway to advertiser demand.

    Among the many technology decisions with an impact on revenue that publishers face, the following are some of the most important:

    • Choosing an ad-serving technology—the key tool for delivering both direct and programmatic sales, as it determines when to serve a direct campaign and when to sell an impression automatically
    • Determining how many, and which, programmatic demand sources and tools (such as SSPs) to use, taking into account, among other factors, access to demand and the functionality of the SSPs—including their “decisioning” capabilities, which help maximize revenue from one or more demand sources
    • Configuring the demand sources and tools (concurrently or sequentially), with the aim of maximizing CPMs and fill rates
    • Considering whether to implement a separate data management technology or to rely on functionality integrated into a programmatic platform

    Complicating matters further, publishers must make these decisions across multiple formats and platforms, including desktop, mobile, display, and video. While many ad servers, and the more sophisticated SSPs, work across formats and platforms, enabling publishers to remove some complexity and gain an integrated view across their inventory, this is not universally the case.

    Determining the right stack partners and configuration, and making the most of decision engines and algorithms, are critical. As the COO of a European Web portal told us, “As a result of reassessing priorities of different demand sources, the performance of campaigns running in programmatic is much higher than we’ve seen before. We’ve increased our fill rate to about 95 percent. We’ve seen an uplift in our eCPMs of more than 100 percent.” (See the sidebar “Technology Trade-Offs” for a discussion of the key considerations for a revenue-maximizing technology stack.)


    To avoid getting lost in complexity, publishers need to think strategically about their technology stacks—the layers of software that make up the ad-serving ecosystem. Many publishers elect to use a primary SSP from which they access and manage most, if not all, of their real-time bidding demand, since all the major SSPs provide a large pool of demand. While there is considerable debate about the amount of additional unique demand that can be tapped by accessing multiple SSPs, some publishers believe that it increases yield and that the additional revenue generated more than offsets the disadvantages of a more complex stack.

    In our study, we found clear benefits to using a unified stack—meaning the ad server and SSP are provided by the same company and are designed to work in concert. The more complex the stack, the more inefficient it almost inevitably is. Impressions can be lost at every stage of the ad-serving process, from ad server to SSP to DSP, and the associated revenue is lost with them. Our analysis found that less than 0.5 percent of “discrepancies”—ad impressions that are lost as they pass between different technology platforms—occur with unified technology stacks, while with other technology setups, 3 to 10 percent of impressions are lost between the ad server and the exchange. (See the exhibit below.)


    Even when impressions aren’t lost, data associated with them can be, reducing the value of the impression to buyers. In addition, a unified stack can prioritize demand based on live bids, which allows for real-time decisions about how to maximize revenue from each impression.

    Publishers that use multiple programmatic demand sources recognize that there is substantial overlap of demand among the major SSPs, but they believe they can leverage buyer behavior and the way they configure platforms to generate higher prices and increased revenue. One technique is to compare bids for a given impression, either live or based on past performance, in order to sell the impression via the highest-bidding demand source; however, this requires the stack to have the relevant functionality and “decisioning” capability. Another method is to send an impression to different SSPs sequentially (or to the same SSP multiple times), starting with a high floor price and dropping the price as the prospective sale moves through subsequent SSPs. Publishers use this tactic because programmatic auctions are so-called second-price auctions, meaning that while the highest bidder wins, it actually pays either the second-highest bid price or the floor price, whichever is higher. Publishers can attempt to capture incremental revenue by minimizing the spread between advertiser bid price and auction close price.

    There is no single answer to what constitutes the right technology stack and configuration: each publisher has its own needs, strategies, capabilities, and supply-and-demand dynamics. In our study, around a third of publishers used a single programmatic demand source and half used two or three sources. We found many examples of publishers that were successful at increasing revenue with either strategy. However, those that used multiple SSPs successfully were careful to assess the impact of new sources and configurations. They also carefully consider potential downsides, such as fragmenting demand and the technology and process inefficiencies associated with using multiple SSPs.

    As the line between direct and programmatic approaches continues to blur, technology strategy will play a crucial role in determining the best channel for different kinds of inventory. We expect the use of solutions such as enhanced dynamic allocation (EDA), which is offered by Google’s DoubleClick, to proliferate. EDA optimizes deals across guaranteed and nonguaranteed impressions, allowing publishers to maximize revenue while ensuring that guaranteed impressions are delivered. In our study, publishers using EDA achieved an increase in programmatic revenue of as much as 24 percent and an average increase of 12 percent.

    Build strong go-to-market and analytic capabilities. Capabilities matter—and forward-thinking publishers are already developing their teams, especially in such critical programmatic functions as proposition development and pricing, sales, and analytic yield management. Half the publishers in our study have already strengthened their programmatic teams—or plan to do so in the next year—by hiring programmatic sales specialists and data scientists. Publishers are integrating standalone programmatic teams with traditional sales teams and increasingly expect all team members to understand and make programmatic deals (although only 4 of the 25 publishers in our study have fully integrated their traditional and programmatic sales teams to date, and even those with integrated teams have one or two programmatic specialists who support more complex sales). Publishers that leverage their traditional sales relationships typically generate more programmatic sales from core clients while avoiding problems with channel conflict. Over time, high-performing teams will rely more and more on a consultative approach, supported by data analysis, with deep knowledge of different buyers’ needs, including optimal targeting and an understanding of when different sales channels and models are most appropriate.

    At all points in the go-to-market process, publishers need to leverage data and technology that can help optimize ad exchange setup, analyze performance, provide insights into which segments are valuable to which advertisers, maximize yield, and ultimately increase revenues as well as improve efficiency. Best-in-class publishers develop sophisticated analytical capabilities in order to deliver continuous optimization and measurable revenue improvements using methods such as multivariate and A/B testing. One big US publisher hires finance-sector analysts for their quantitative skills in yield optimization, and the company runs between five and ten A/B tests at any one time.

    While digital markets today vary in maturity, there is no doubt that programmatic deals are central to the future of digital advertising globally as well as across formats and platforms—display, video, desktop, and mobile—and that programmatic and direct channels will become less clearly delineated over time. Publishers need to ensure that they are set up for programmatic success today and tomorrow by reassessing their strategy, sales proposition, pricing, technology, capabilities, and organization—across both traditional direct and programmatic channels. They will benefit immediately from more productive and more efficient programmatic operations and in the future from a stronger market position.

    Publishers already embracing our four recommendations are outperforming by increasing market share (both programmatic and direct), overall CPMs, and total revenue. More important, they are positioning themselves for explosive growth in the programmatic market over the next few years, when billions of dollars of revenue and market share will come up for grabs.