“The Sell Sider” is a column written by the sell side of the digital media community.
Today's column is written by Rachel Parkin, senior vice president of strategy and sales at CafeMedia.
While ad tech has seen a constant stream of innovation over the last few years, one thing has largely stayed the same: private marketplaces (PMPs).
PMPs fill a critical role in the programmatic landscape, providing a direct conduit between buyers and publishers. PMPs have many benefits, from the security of working with known partners and fee transparency to enhanced performance and unique data overlays.
Yet what we're implementing today is pretty much unchanged from what we did when the PMP protocol was added to the OpenRTB 2.2 specs more than three years ago.
With many impending OpenRTB protocol updates and shifting auction dynamics poised to fundamentally shift how media transacts, it’s time to revisit the PMP standard.
PMPs come in many different flavors and forms. There are one-to-one deals, for example, one-to-many deals and many-to-one deals. Demand-side platforms (DSPs) have their own specific terminology too – preferred deal, private auction, you name it – and names tied to platforms mean different things to different people. In some sense, we don't have a standard. Today, buyers and sellers have to translate, so we’re all speaking the same language.
PMPs have clear strengths, but nothing about activating them is easy. They take time to manage on both the sell and buy sides. Buyers must weigh the cost versus benefits of whether the extra effort needed to transact on PMPs is worth it.
Even when PMPs surpass that hurdle, troubleshooting tools don’t provide enough granularity to solve the issue without some guesswork. Uncovering the “typical” reasons for lack of bids is just one baby step toward surfacing the data that’s actually useful.
Rosy estimates for future dominance of private deals as 80% of programmatic spend by 2019 seem generous (at best) based on how PMPs are being used today. Much of the so-called PMP spend runs through deals where there isn’t anything special over the open exchange. The plethora of always-on multi-advertiser deals publishers set up within every DSP are often randomly activated. Some buyers may be haphazardly selecting deals without any insight into priority or quality, and other buyers may be automatically opted into private deals without even knowing it.
Turning on PMPs without any information exchange between buyer and seller misses the real advantages that come from a collaboration around the advertiser’s goals and inventory comprising the deal.
Ideas For The Next Generation Of PMPs
PMPs still have a place in our market. Ads.txt may circumvent spoofing, and the first-price auction can mitigate fee transparency, but neither solve for everything. PMPs offer access, curation and dependability. With a few new twists in functionality and mindset, it's possible for PMPs to feel shiny and new again.
We must first create clarity around the PMP model by divorcing priority from pricing. It’s time to retire platform-specific names and concentrate on what each PMP lever does to accomplish an advertiser’s specific goals. Priority, auction type and price are all independent. Any PMP can be set as a fixed or floor price, regardless of the priority or price.
The priority of a PMP can either be set based on the price or strategic relationships. The PMP price should reflect the desired win rate or benchmark return on ad spend. Aligning on each of these elements should become a part of every PMP conversation.
We also need to focus on using PMPs for truly unique opportunities.
Private deals provide a way to transact units that aren't typically sold on the open market or even through programmatic pipes. New and creative units could range from the extra-large format units to highly custom mobile units or exclusive native placements.
Publishers can also curate PMPs based on first-party audience or data to provide value advertisers can’t get elsewhere. Relying on publishers for what they know directly over any third party maximizes reach because this information is imperfectly measurable, both in context and coverage. Shifting the PMP’s role toward fewer, higher-value opportunities will increase the ratio of the payoff to the effort.
Automating more information sharing between supply-side platforms and DSPs will simplify the burden of managing PMPs. Passing the tactic ID should just be the tip of the iceberg. Buyers could also share KPIs, such as viewability or conversion type, and their associated thresholds.
And publishers could enable advertisers to make better extrapolations if the OpenRTB protocol enabled self-reported deal settings – audience (woman 18-34), performance (80%-plus viewability) or waterfall priority level – to be sent in the bid request. The more we share, the more easily we can troubleshoot, too.
There needs to be a better solution for programmatic guaranteed deals. Most of the working models in place effectively disable the DSP from doing what it was designed to do: give buyers RTB control. Bidding forcibly on every impression, with no opportunity for universal frequency capping or audience overlays, is a bad mechanism for what should be a robust, rich platform. If we can enable the forecasting and technology to handle guarantees in a true RTB fashion, programmatic guaranteed deals will give advertisers consistent access to inventory sources, compared to the uncertainty of the open market.
We have to change our mindsets and recast PMPs as a partnership above all else. A PMP is much more than just the technology of a Deal ID. At its core, it’s a private partnership ascribed to how we approach deals, as opposed to how they are transacted. Collaboration can even extend beyond the Deal ID to the open exchange, although this still leans on third-party technologies to fill the data gaps.
The value of private deals isn’t to be automated – it’s to be effective. Nothing happens in a vacuum. When advertisers and publishers work together toward a goal, the whole really is greater than the sum of its parts.