It’s been a little over a week since we launched the RTB 500 and I apologize for not weighing in on it again sooner, but this little ad festival over in Cannes, France, more or less got in the way. I thought about writing about some of what people were saying over there about programmatic marketplaces, technology, data and RTB (mostly, it wasn’t good), but I figured I should come back to the RTB 500, because it’s the first chance I’ve had since the launch. So here are a couple of things. First, the reaction to it has been overwhelmingly positive. Actually, there have been no real negatives yet. And the closest thing to anything negative has been some confusion about what it is, what it really represents and means. And I take that as a good thing, because part of our mission is to help figure that out. So let’s pick that up again here and I welcome anyone to step forward, weigh in, and provide more context.
Among the comments I’ve received were some professors of business and economics at a major university who want to discuss exactly that with us, and I will report on whatever comes from that shortly. In the meantime, I’d like to discuss one initial reaction we had to the index, and some insights that grew out of the discussion. It came from Mike Nolet, CTO and Co-Founder of AppNexus, who commented on my previous post that the RTB 500 looked “interesting,” but he didn’t feel it looked “very accurate,” because it didn’t match AppNexus’ quarterly trending data indicating that demand and pricing for RTB inventory builds during the course of a quarter, and rallies near the end of the last month due the way advertisers manage their quarterly budgets.
“I've always seen a gradual increase in price as the quarter progresses with a strong peak in the last few days of [the quarter] when advertisers pump unspent budgets,” he commented on RTBlog, adding, “The first few days of a quarter are always low as advertisers are late setting up the next quarters campaign.”
Interestingly, I heard almost the exact same pattern when I previewed the RTB 500 to Art Muldoon and his team at Accordant Media. Muldoon said Accordant is analyzing that pattern and may issue an analysis of its own as part of its quarterly reports on the RTB marketplace, but he said it basically shows a similar flighting patterns to what Nolet describes, and his conclusion, like Nolet’s is that advertisers are simply dumping their unused quarterly budgets in the RTB marketplace to make their quarterly goals.
If that’s true, it is an interesting discovery about RTB marketplace behavior that has not really been discussed openly before, so I’ll take that as one of the first genuine insights to be uncovered by the RTB 500, even if only indirectly. Because if you ask me, it’s a behavior based on an artifact -- burning up unused quarterly budgets -- and not one a rationale read of the supply and demand of inventory in the RTB marketplace. And it is the exact reason we wanted to start publishing the RTB 500 index -- to give people a yardstick to compare their own behaviors to.
if what Nolet and Muldoon say is true, it’s a behavior based on a vestige that I’m not sure any individual marketer -- and maybe not even any individual agency may have been able to see. You’d have to have a diversified portfolio of brands actively bidding in the RTB marketplace -- the way say AppNexus or Accordant does -- to see that kind of behavior. I’m curious what the folks with an even broader view, like a Rubicon Project, or maybe OpenX or Google have seen about the peaks and valleys of quarterly RTB budgeting.
If it matches what Nolet and Muldoon are saying, my own belief is it says more about how immature and unsophisticated the RTB marketplace currently is -- that predictable swings can be made based simply on brands burning off extra budgets at the end of a quarterly cycle. Don’t get me wrong, it is a genuine market behavior, because obviously, people (and their machines) are behaving that way. It’s just that it probably doesn’t have anything to do with the actual supply-and-demand of biddable impressions available in the marketplace, like the way annual trends showing the marketplace building in the fourth-quarter and crashing on January 1. That’s a pattern of supply-and-demand that pre-dates the RTB marketplace and reflects the tightness of marketers competing for share of consumer attention during the peak holiday marketing season.
Anyway, I thought I’d come back to that one to see if anyone else has a POV on what it means, whether it makes sense, or is as I suggest, just an artifact vs. genuine supply-and-demand. I’ll be back soon with some other observations about the marketplace -- hopefully with ones made by you.
-- Joe
When we founded our most recent company, we deliberately chose accounting periods that didn't line up with yearly quarters for exactly the reason here - there's a disadvantage in competing with the pack for resources. I can't believe that any serious manager doesn't know the same, so why have they fallen into this trap?
The inefficient spending trends that programmatic buyers espouse has long been perplexing to me. We witnessed this very early at Lijit Networks in 2011 and it was like clock work every single month for the next 2.5 years I was there. I could tell you within a +/- 2 days of when the lift was going to begin occurring in the bidding. I'm actually shocked that this is considered "new." I created specific strategies for our publisher yield team to take advantage of this exact phenomena. A smart publisher will still take advantage of this asymmetry of information and throttle their direct deal campaigns up in the beginning of the month and slowly throttle back while ramping up their programmatic(mainly network rtb buyers) in the latter part of the month. This takes advantage of the monthly trend and helps assure greater overall revenue yield. Andrew Casale wrote about this last year (http://bit.ly/1kqbFRv) so I'm again confused on why this is a challenging question for Mediapost? The bigger question is why aren't smart marketers changing their ways? FYI, Nolet left Appnexus last year.