The Dawn of Real-Time Guaranteed: An Interview With OpenX’s Dmitri Kazanski

A Sweet Proposition for Buyer and Seller

Possibly the biggest reason header bidding is such a game-changer is because it gives demand sources insight into more—if not all—of a publisher’s inventory. This enables buyers to better evaluate inventory and bid more acutely in real-time buying environments. (If you feel a need to revisit the horrors of the waterfall, read here.)

This advancement has paved the way for real-time guaranteed (RTG), or what you might call the “hybrid option”: guaranteed display buys that are transacted in an RTB environment. It’s the best of both worlds: advertisers can target premium publisher inventory in real time based on data and publishers—and advertisers, honestly—know exactly how much spend is going through, which is something they can predict against.

At the last several AdMonsters events, we pondered what it would take to make RTG work at scale, but it turns out that OpenX already has an RTG product in beta that leverages buyer-seller data syncing. While we were in the Los Angeles area for the October Tech Forum, we stopped by the OpenX office to hear more about this cutting-edge offering from Senior Director of Programmatic Direct and Ad Server (as well as veteran AdMonster) Dmitri Kazanski.

WITH THE SUPPORT OF OPENX
OpenX exists to help publishers grow their businesses by monetizing great content.

GAVIN DUNAWAY: Tell us about the development of RTG. 

DMITRI KAZANSKI: Right now, publisher yield curve is similar to a barbell: on one side there is direct guaranteed sponsorships, native and integrated placements with very high CPMs. On the other is programmatic, which typically yields lower CPMs.

Private marketplaces were created to fill the void between the two, but have fallen short of expectations. Publishers commonly see very low fill rates and buyers aren’t getting the scale they need.

While buyers may promise to spend $400,000 a month, they may only end up spending $500. If the targeting parameters become too granular after overlaying geo, frequency cap and day-parts, the number of eligible impressions shrinks dramatically.

Also, if the buyer and seller use different third-party data, viewability and geo-IP vendors, discrepancies in PMPs can be dramatic. The buyer’s definition of 70% viewable in-market for luxury cars from the LA DMA could be very different from the publisher’s definition for the same targeting criteria. So, if the publisher exposes 5,000,000 qualified impressions based on its tools, the buyer sees only a fraction of the impressions as qualified.

RTG effectively resolves those problems. It integrates an audience sync that matches the buyer’s audience to the publisher’s audience to find exactly how many impressions match the selected targeting parameters as per the buyer’s definition. The publisher gets a guarantee of revenue because the fill rate is really high and buyers no longer need to cherry-pick impressions because the audience is pre-selected and booked ahead. Both the buy side and the sell side want to know what the volume and spend is going to be, and RTG delivers that level of predictability.

ROB BEELER: As a publisher, if I can see my inventory through the buyer’s eyes I can make a better deal and hit my fill rates. We talk about PMPs being frustrating for the publisher, but buyers have their own expectations.

DK: Indeed. Having spent 15 years on the ad network and publisher side, last year I managed programmatic buying against custom audiences. PMPs were hugely frustrating. You wouldn’t know how much you were going to get. If the fill rate is low, making a PMP deal with a publisher is a waste of time.

With RTG, instead of attempting to mirror the desired buyer’s targeting definition on the publisher side or using content as a proxy for audience, we sync the exact buyer’s audience in form of cookie IDs and mobile Device IDs. Thus, there is no chance for audience definition mistranslation and low fill rates.

RB: Header bidding makes all publisher inventory available for demand-side evaluation, so you’re dependent on the market to tell you what you’re worth at the time of purchase. I’m old-school—I want to know what I’m worth next month. How can header bidding and real time guaranteed add up?

DK: Header bidding exposes more inventory to programmatic buying opportunities. Thus, the technology will actually accelerate the move toward RTG deals.

GD: So, what are you guaranteeing against: number of impressions or amount of spend?

DK: Both; since RTG deals are priced at a fixed CPM, the impression guarantee translates into guaranteed revenue.

GD: Can buyers switch up the rate they’re willing to pay if they’re not meeting the numbers they want? 

DK: This is something buyers do in private auction and open auction environments. You increase your max CPM bids, and hopefully win more often. Even with the bid increase, you might or might not be able to spend the budget that you want against your target audience.

Preferred deals are transacted at agreed upon rates, so the buyer does not have a way to increase volume by increasing the bid. Finally, with RTG, you don’t worry about not meeting your impression numbers. They are guaranteed.

GD: This sometimes gets to be a real point of confusion: how do you distinguish real-time guaranteed and automated guaranteed? 

DK: In automated guaranteed, the campaign gets automatically trafficked into your ad server using special tools, as opposed to buying it through a DSP. In the time I spent as a programmatic buyer, I frankly didn’t have time for automated guaranteed, because you need to use other tools, carve out separate budgets.

Furthermore, automated guaranteed buyers have no control over how the campaign is targeted and must trust that the publisher will deliver the automated guaranteed campaign to the right audiences. The trafficking is automated, so there is less overhead, but the buyer doesn’t have much insight or ability to optimize. With RTG, the buyers know that they are getting the audience that they’ve defined.

GD: Can you do hybrid deals?

DK: One of the flavors of RTG is a flexible guarantee, where the seller agrees to send a certain amount of inventory to the buyer and the buyer agrees to buy a specific percentage. This gives the buyer the flexibility to optimize within that buy.

GD: Instead of using the buyer’s or a third-party audience, could audience targeting be supplied by the publisher? 

DK: OpenX supports RTG use cases with an audience sync or without. If a buyer wants to buy against their own or third-party audience, OpenX will disclose exactly how many impressions can be bought against this audience. If you trust the publisher and believe in the value of their audience, then you don’t overlay your own audience and you base the buy on the publisher data. Many publishers have unique audience data not available to advertisers and third-party data aggregators, so this is a very attractive option.

GD: It sounds like more accurate TV buying.

DK: Right. You know ahead of time how much you’re going to get, as opposed to hoping you get enough, and then figuring out it’s not.

GD: Do you think people are going to immediately go running to this product? What do you think is going to drive adoption? 

DK: Based on the conversations that we have had, there is a lot of interest. Both publishers and buyers are signing up as we speak. Publishers want to plan ahead on revenue and carve out buys at a high rates and buyers are looking to buy specific audiences with predictability at a predetermined rate—RTG promises to deliver on both of those needs. That said, the technology has the power to replace both PMPs and automated guaranteed, it’s the next evolution of programmatic buying.