ADOTAS – And you thought demand-side platforms could be contained by the display space — on Tuesday, TubeMogul relaunched its PlayTime video platform as a video DSP, offering brands and advertisers access to inventory sources stretching most of the Internet and the ability to bid on 60,000 impressions per second.
David Burch, TubeMogul communications director and head of research, gave us some insight into the decision to transform PlayTime into a DSP as well as the sanctity of video inventory.
ADOTAS: So why go the demand-side platform route? What about the performance of DSPs in display inspired you?
BURCH: Today, brand marketers face a complicated market. Buying from different sources (publishers, networks, etc.), makes it difficult to optimize across ad types (pre-roll versus mobile, for instance) and vendors. It also precludes easily measuring campaign-wide, deduplicated audience numbers. Scale is also an issue. DSPs brought greater efficiency and optimization to display.
But that is where the analogy ends: video is different and requires going beyond than clicks and impressions, so we built this to be the best at delivering brand messages across audiences, ad types, placements and more.
What makes you think demand-side platforms viable in the online video space?
Inventory is obviously growing (i.e., YouTube making its pre-roll inventory biddable). It’s really a win-win for everybody. Publishers love it because they are getting built-in auctions to sell remnant inventory and often technical help on serving for their own ads. Brand marketers and trading desks love it because they get scale, de-duplicated audience metrics, optimization across inventory sources and the ability buy audiences everywhere.
What were the most dramatic changes you had to make to PlayTime to transform it into a DSP?
Before, we offered an ad network-plus, with ad serving, proprietary analytics and our data from and relationships with publishers. So technically, this was a tremendous undertaking to make it cross-platform, cross-format and plug-in third parties (you can see our partners here). In all, we are taking in over 60,000 auctions for ad space per second, making decisions in the milliseconds before a video player or page loads (no easy task). To offer it self-serve, we made a big investment in user interface and marketers can test the results — sophisticated campaigns are easy to set up in minutes.
Beyond that, arguably the most important developments are in regard to what brand advertisers care about: getting the most impact per dollar. We shift to the audiences and placements where brand messages are being watched the longest and impact is greatest across a raft of variables, with great results after extensive internal testing on real brand campaigns. Marketers can use our analytics (developed for over four years) to measure the results — both by site and by ad.
What’s the biggest advantage of being a DSP in online video?
Much of what I outlined I think are the benefits. Self-serve obviously is great for trading desks, but isn’t appropriate for many media buyers, who want to leverage external expertise and have others do the heavy lifting, which is why we introduced a managed service that comes with a lot more.
Do you think the development of the video space is mirroring display? If so, what comes next?
By nature it can’t exactly mirror it, since brand advertisers with TV ads or longer viral ads to sell things like toothpaste measure success differently than display. We built our platform from the ground up with that in mind.
YuMe cofounder Jayant Kadambi recently suggested that video is a distinct piece of inventory, fundamentally different than display — are you of the same mindset?
Yes. Premium pre-roll is obviously in high demand for a reason, and that is a big part of our platform and inventory. However, I think that in-banner video (i.e., videos viewers can choose to click to play in banner units) is going to grow. Google predicts that 50% of their display inventory will have video by 2015, and our joint research with Dynamic Logic (first two graphs) reveals higher purchase intent and awareness than simple flash/rich media display ads. But success is measured much differently.
What are your thoughts on display ad networks partnering or acquiring video ad networks? Wise move for either party?
It’s interesting. Not to punt, but I think the market will answer that one.