Understanding The BlogHer Ad Tier System: Part One
This is the first in several posts explaining the statistical evidence I have found that suggests there is a tier system that — together with traffic — governs how much money publishers on the BlogHer Ad network make from running advertisements on their blogs. For a review of what I’ve discussed already on the topic, you can start with a post that questions how bringing a very high traffic site onto the network might impact other posts here. Then, you can read my kind of convoluted metaphor for the money-making scheme here, and my more general introduction to all the issues involved from the last post. Today, I’m going to show you how the tier system can be shown just by looking at how my own statistics have changed over the course of my time serving BlogHer ads for about a year and a half.
What I Have Learned About Tiers In Online Ad Networks Over The Course Of The Past Week
BlogHer is not the only online ad network that uses tiers. Tiers are used by many (if not all) online ad networks to package small sites with big ones in order to create a more “effective” CPM rate and, therefore, be more attractive to advertisers and big ad agencies. For example, a premium, high traffic site like ThePioneerWoman.com is where big brand names are going to want to advertise, but it would be tough to sell a package of ads to an advertiser with all premium sites like this. Why? Because ThePioneerWoman.com is only going to be running ads at very high CPM rates (like $12 CPM or more, say). So, to be attractive to a big brand, the advertiser will package ThePioneerWoman.com with a small site, so that they can sell the “pack” at a rate of like $8 CPM, and the advertiser gets some hits on ThePioneerWoman.com, and some on the smaller sites, and the network gives PW the CPM of $12, and the little guy the CPM of like $3, and pockets the difference.
That’s one kind of tier. A CPM tier.
Then there’s a tier for available ads, which is what we’ve been talking about with BlogHer. From looking at stats, it is my *theory* (again, just a theory, because without anyone admitting it, that’s all I can say is that this is a theory) that there are several different tier systems at work at BlogHer: one for CPM rates, one for available paid ads, one for who gets house ads, and one for who gets defaults. And possibly more. It is also my theory that the tier system is not tightly controlled by anyone, but can be altered for various reasons. My subsequent posts on this topic will explain how I arrived at these conclusions, beginning today with a look at my own stats, and the conclusions one might make from them. Bear in mind that my stats actually provide the weakest of the evidence for tiers: only after looking at the numbers for many other people did I become totally convinced of the tiers being real, like I said before, it’s just too much to do all at once, so I’m still going to just start with my own numbers.
My own traffic
There are a couple of things about website traffic that constantly change — one, your traffic changes throughout the year for various reasons — your blog grows, online traffic ebbs and flows, maybe your posting changes. And two, how many ads are sold changes throughout the year, because of the fact that ad budgets (as Blogher is so terribly fond of pointing out) are funded in a cyclical fashion. Because of this, it’s really tough to compare things in terms of dollars and cents from one quarter to the next. This is why I started looking at the percentages of unpaid ads I was running now versus before. Below is a graph that shows what percentage of unpaid ads I was running throughout my career as a BlogHer ad network member. Just so you understand, in this case, a high percentage is a bad thing, it is a high percentage of ads for which you are receiving no money — these are public service announcements, remnant ads (“leftover” ads that didn’t run from the previous month, this doesn’t really happen anymore but it did back in 2008), or completely blank ads, or house ads (technically you might receive some compensation for these ads but the rate is very, very low, so I’m counting them as “unpaid”). You want a low number here, because any time you serve an unpaid ad, you’ve got crap on your site that you’re not being compensated for, and you are still being held to the storied BlogHer editorial guidelines and all that jazz, and running BlogHer headlines, and all that jazz, even though you’re not being paid.
OK, so that’s kind of a volatile graph, but you can see that I was running a bunch of unpaid ads at the beginning, and then eventually I went down and stayed down pretty steadily for a while, and then went back up, and then kind of went back down, like halfway down and stayed there until they booted me off the network. These kinds of fluctuations are usually explained by the “cyclical nature of advertising sales.” In the interest of addressing those arguments, let’s consider that graph according to quarterly boundaries, noting that Quarter 4 (October, November, and December) is traditionally the most lucrative quarter for advertising because of the holiday season. If it’s a lucrative quarter for advertising, then my unpaid percentage should be going down, because there should be more ads on the network — provided that the fluctuations in this chart are only being caused by seasonal changes to the ad market, that is.
If a revenue decrease were as simple as seasonal changes, there wouldn’t be these kinds of inconsistencies. You wouldn’t see a sharp increase in unpaid ads during the holiday season like I did in Q4 of 2008, and then see them flatten out in Q4 of 2009. If anything, you would see them decrease during that time, and then increase afterwards, when things got bad, and the curve would look similar both years, only changing in tune with the economy.
So what accounts for the sharp differences in the graph? Let me see if I can illuminate things by telling you a few things about what happened during that timeline, both on this blog, and elsewhere in the BlogHer Ad Network, and you can draw your own conclusions.
In late April of 2009, I was chosen as BlogHer of the week. The next month, my percentage of unpaid ads changed drastically. To be exact, it went from almost 61% unpaid ads (that means that only 39% of the ads I was running were actually giving me revenue) to about 28.5% (meaning that about 71.5% of the ads I was running were paid). The next big spike in my percentages occurred in January of 2010, when my unpaid ads jumped up to about 53%. That was also the month that Ree Drummond joined the BlogHer network. In February of 2010, I wrote a post questioning whether our revenue had been going down because of this move. That month, my percentage went down again, to 36% unpaid ads, right smack in the middle of Q1 (the worst of the financial quarters).
When it’s one person, it might be dismissed as coincidence. My next posts will make it seem more difficult to do that.