Commentary
Let's Debate: RTB's True Cost Of Goods Sold
- by Mike Leo , November 14, 2012
I’m a big fan of debates.
In fact, they were my favorite part of the 2012 election season. After watching both presidential candidates dance around the issues for months on the campaign trail, seeing them discuss policy in a one-on-one forum was a welcome change. Over the course of three 90-minute debates, both Obama and Romney confronted more tough questions than they had in the past year -- and, in some cases, actually answered them. The debates pushed the candidates out of their comfort zones, cutting down on the say-nothing vagueness that, unfortunately, today’s campaigns are built on.
That said, I think real-time bidding could use some debate, too. There's plenty of talk about the RTB ecosystem but, for the most part, discussion about the toughest, most critical issues is avoided. Having an honest debate can help RTB grow, dentifying its problems and creating momentum to solve them.
Instead of calling for debate and walking away, I want to get the discussion started. In the space I have left, I’d like to offer up a debate prompt and some context to get the ball rolling. Then it’s up to you figure out what side you come down on and make the case for your view.
advertisement
advertisement
Resolved: RTB will overcome its true-cost-of-sale (TCOS) problem
TCOS Defined: The true cost of sale (TCOS) is the spread between what an advertiser paid and the revenue the publisher nets after all costs are considered.
Context: Here are the two main problems publishers have with RTB:
1. It bundles their good inventory together with others’ crap inventory, making it difficult to sell their ad space for its true value.
2. The different actors in the ecosystem get so much value that little is left for publishers by the time the check comes in.
For this debate, let’s assume the “programmatic buyer’” pays the true value for the publisher’s inventory. While I’m skeptical that the situation regularly occurs, the assumption will allow us to concentrate on the issue: RTB’s true cost of sale.
To understand RTB’s TCOS, it’s helpful to first examine direct selling’s TCOS, which is made up of the costs a company must pay a sales force to sell ads, an ad operations team to traffic the ads, and a finance team to bill. Add an ad server into the mix and the direct seller’s TCOS comes out to about 10%-12%. The costs, of course, are visible on the publisher’s P&L and can be painful numbers to look at.
The RTB ecosystem, on the other hand, removes the costs from view. The process is simple: a publisher plugs its inventory into the system and waits for the check. But, the fact that the costs don’t show up on a P&L does not mean they don’t exist.
The TCOS for RTB can be found in all cuts taken in the middle. A typical RTB scenario may have a trading desk, DSP, exchange and SSP all taking 15% cuts. After all those cuts, a publisher whose inventory sold for $4 may only end up with 70 cents. The TCOS for that transaction: over $3.
So, can RTB solve its TCOS problem? Will vendors ever be willing to take a smaller cut of the pie? Let's debate it. Without getting to the heart of this issue, and fixing it, RTB won’t ever amount to much more than a slightly better version of an ad network.
Interesting arguements. Real and scary for us publishers. Re: TCOS comes out to about 10%-12% -- Seems low actually...was it '10%-12%' from publisher/companies 50 million+ a year? What was criteria? Does this include sales and ad product related vendor fees/salesforce etc? I bet it's more like 40-50% for companies making under 20mm a year. Love to hear your thoughts.
When telling people in Germany RTB is even worse then the horrible discount selling on anonymous networks I was told not to lean against the trend. Now the TCOS approach shows exactly this while I find other issues even more disappointing: Not only do you mix your inventory with crap - the nivellation of inventory value adds to the fact that user profiles are generated by networks and RTBs - with YOUR users! and exploited across the networks. While your inventory is discounted you help adding value to crap sites. Of course if you would want to avoid that you'd have to get rid of all Google services on your site - and there all clear thinking surrenders to the cost evasion rationale (e.g. Google Analytics is for free). My claim: If you are a big player in your area of business pulling out of networks entirely would have immediate effect on prices paid for the kind of profile your site visitors establish. If someone want to buy into that he'd have to come to you again.
Really interesting article Mike. Hidden TCOS will always be a part of the business from agency down. Just how "hidden" this cost is in the RTB ecosystem depends on both a) the transparency of the platform that a publisher does business with (both buy and sell side) and b) the publishers’ willingness or desire to lift up the hood on what would otherwise be an automated auction.
To your point, many many publishers are happy to sit back and "wait for a check;" these less engaged publishers are also the ones who are most likely benefitting from the bundling you describe, landing on post-campaign reports alongside more premium pubs and reaping the rewards of audience targeting, remarketing, etc. That’s all fine, keep in mind that many of these large scale, less branded pubs paved the way for RTB by contributing scale and reach in the early RTB days.
More recent entrants into RTB, pubs who were waiting for pricing controls and programmatic direct/preferred architecture to evolve, are by nature now more willing to engage directly with advertisers to strike deals, pricing agreements, and CPM mins through RTB. As these deals flow and are directly tracked through various vendors to the end publisher, fee structures become less “hidden” and more transparent and can be addressed.
I agree with Joshua that 10-12% TCOS on direct seems low, but would also add that the comparison is not apples-to-apples; RTB is most often viewed as a supplement to direct not as a replacement (at least in the realistic near term)... the TCOS for running 3rd party network monetization, which RTB is more likely to replace and thus be closely comparable to, is probably much greater. So putting that 10-12% figure up against a figurative $4 → $0.70 CPM reduction is a bit extreme and feeds bad perception.
To get to my answer to the question posed, I don’t think a smaller cut for vendors is the solution here; as publishers become more empowered through tools and services to engage directly with end advertisers through the RTB channel they will have more control over the offered rate, can “unbundle”, can offer value add through differentiated inventory, and will end up benefiting from RTB efficiencies.
Mike Leo, the first problem you described is one of perception. The problem itself is non-existent:
[quote]It bundles their good inventory together with others’ crap inventory, making it difficult to sell their ad space for its true value.[/quote}
As a publisher, how do you rate value of your own websites? Let us rate sites based on actual data!
The #1 strength of RTB is to buy only the relevant impressions for the user currently present on the publishers site. For an airline advertiser, a weather website could have a much higher value, definitely when storm is ahead. For an automotive advertiser, the value of the site is only based on its reach and on the wishes of the advertiser to be exposed on that reach.
So I disagree defining this as a problem. Personally, I think it comes down to emotions, publishers are afraid because they cannot ask extremely high cpm prices. They have to trust the bidding proces.
This actually means they (publishers) have to start thinking how to make their placements as attractive as possible for their visitors. And connect to a network with a high demand (DSP).
BTW very annoying empty lines are deleted. No text mark up whatsoever. Nobody is actually going to read these 1-sentence posts..
Mike,
thanks for a good debate kick off. From my perspective, the costs look very different. I work for a publisher and our COS of display is definitely higher than 10 %.
On the other hand, the good-case scenario is: the client pays 10 % to a DSP, we pay 15 % to an SSP and it's done. And it realy works this way in many cases on our market. Of course, if there is an agency, data provider etc., the cost is higher, but I think generally RTB really saves money.