"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Casey Saran, senior vice president of publisher products at Yieldmo.
Since its inception in 2009, real-time bidding (RTB) has been a total game changer, accounting for approximately 50% of digital display ads in 2014, according to eMarketer.
But the display landscape has changed since 2009, and in order to continue this growth, RTB standards must also continue to evolve.
The most important need involves delivering high-impact, purpose-built creative in new environments with the scale and real-time decisioning capabilities offered through RTB.
The Evolution Of The Supply-Side Platform (SSP)
We all know that when RTB enabled real-time decisioning on ad spend, much waste was eliminated from media plans. It’s a no-brainer why the buy side was quick to optimize budget toward a channel that could enable audience buys instead of content, which isn’t always a great proxy for demographics. Yet the reason most marketplaces don’t develop overnight is because they require adoption from both buyers and sellers.
The sell side, at this time, was drowning in inefficiency trying to manage the many ad networks that they would daisy chain in their ad servers. Rubicon, Admeld, Pubmatic and other companies began offering software as a service (SaaS) to manage and optimize a seller’s unsold inventory. These early “yield optimizers” didn’t always bring publishers new demand. Their original role was to run a price-priority auction based on which ad network was predicted to return the highest value and build the best possible daisy chain in real time.
The timing of this adoption on the sell side coincided perfectly with the emergence of bidders. And this was the technological primordial soup that enabled RTB: Bidders could capture large chunks of media spend because the RubiPubMelds of the world had already bundled up the inventory for them. Furthermore, these yield optimizers were eager to accept demand-side platforms (DSPs) into their ecosystem because it justified them being part of the revenue flow and delivering additional demand to publishers, rather than existing merely for operational efficiency.
Another component of this is Google, as always. When the yield optimizers were integrating with DSPs and becoming SSPs, Google already had a massive display business with Google Display Network, also known as GDN, the display ad component of AdSense. When AdExchange launched and Google acquired Invite Media, which became Doubleclick Bid Manager (DBM), it went live with only one supply source, GDN. So within less than a year there was effectively limitless inventory for buyers to access due to the existing consolidation of inventory in RubiPubMeld and GDN.
This trough of inventory meant DSPs could grow up quickly, and they did. They quickly became specialized to include retargeters, performance marketers, self-serve buying platforms and advanced data management, among others. As real-time bidding became a more critical part of the sell side’s monetization strategy, the tools and controls to support sell side evolved as well.
For example, publishers could set floor prices by advertiser or user frequency, view detailed histograms of bidding activity or set up private marketplaces through deal ID to better control who could access specific inventory and at what transparency. The skill set required to manage programmatic selling led to the creation of more technical and senior roles at publishers.
Despite this renaissance of controls and reporting, the sell side had become largely commoditized. While one SSP might provide a cleaner reporting interface than the next or take less time to set up, ultimately the decision of which one to use would come down to yield. Bake-offs, the running of split user tests with two SSPs to see which performs better, were a common way of choosing a partner, leading some SSPs to “juice” margins to try to remain competitive.
Bidders, on the other hand, don’t necessarily care about which SSP they are buying on either. They care about having access to as much inventory as possible, which makes it hard for an SSP to have any differentiated demand.
To prevent being further commoditized, SSPs need to find ways of actually increasing the value of inventory that they make available to buyers – just being the pipes isn’t enough anymore. I believe that the key is in the evolution of the creative.
In the pursuit of a successful campaign, there are often two forces that play against each other: custom executions and automation.
Custom executions offer superior creative and generally perform better, but are operationally inefficient. Automation offers tremendous scale but doesn’t necessarily deliver the message in a way that is as impactful. Native support in OpenRTB 2.3 is an attempt to solve for this.
When I say “native,” I am specifically referring to each asset that makes up the creative – including title, description and image – being delivered individually to the SSP, instead of being rendered by a third-party tag. This is what the IAB native spec allows for so when I say “native,” that is what I’m talking about. Whether or not the ad is designed to look like content on the page is irrelevant to this discussion.
By this definition, native can support a lot more than just making ads styled to look like they belong alongside publisher content. There is even the potential to deliver rich media experiences where the SSP builds the ad for the right user in the right environment. This will allow brand dollars to flow more easily into the RTB environment.
The next evolution of the SSP is rooted in the emergence of support for native ads via standard RTB protocol, enabling automation of custom executions. SSPs now have the ability to optimize the execution of the creative and apply their own expertise on ad formats to add brand new value to transactions. The good news is that we are already seeing many sellers move in this direction. In addition to superior performance, purpose-built ads generally deliver a better user experience as well.
DSPs will be the ones that need to catch up this time. Yieldmo (I work here), ShareThrough, Triple Lift and other companies are already bundling billions of monthly impressions for programmatic buyers. So just as when RTB first emerged, the inventory for this next wave is already here and DSPs need to participate.
Catching Up To Facebook
Facebook already does all of this. It is the buy side and the sell side, and it is attracting larger and larger ad spend because what it is doing works.
Facebook already allows buyers to provide their creative assets. It also has a massive amount of owned and operated inventory to experiment with the right way to target and build the creative to hit advertiser KPIs.
The native RTB spec really just introduces a standardized way to do many of the things that Facebook has been doing, only in an environment that preserves control for both the buyer and the seller.
We already know native ads work. We also know RTB works. Now it’s time to bring them together at scale.