For some time, I’ve been calling analysts who cover the online ad space and complaining. But I haven’t seen any evidence that my complaints have had any impact, so I decided to take that complaint public.
Analysts attribute revenue to the online ad industry in the wrong way.
Never mind that I think most of the numbers are wrong, specifically that they’re low, at least the way they’re counted today. My big complaint is the way analysts attribute dollars to the advertising space.
Let’s start with the basics. A country’s advertising grows at approximately the same rate as its gross domestic product. Which means that overall, advertising stays linked to the same general percentage of the total economy. Advertising growth in the U.S. is usually 5 percent to 8 percent per year. Over the course of that year, advertising is attributed to various media types by the analysts of those media, and reports are written so investors and businesses can understand what’s going on in the ad market.
When analysts count ad spend by media, they do something very strange — even unprecedented — when they count ad spend for online media. Let’s take a look at what I mean. Here is an abstracted chart showing how the analyst population generally represents online media and traditional media growth:
|Ad Spend Growth, 2001-2010|
If we look at the way ad dollars are attributed for traditional media, and specifically at television, the attribution mechanisms are pretty clear. When broadcast television started being distributed across cable, analysts didn’t change the dollar attribution by distribution mechanism. In other words, broadcast TV dollars are always attributed to broadcast TV regardless of how the content was technically delivered to the viewer. Cable TV dollars are only attributed to cable-only TV programmers.
Typical and Ideal Ad Spends
So why do we attribute dollars associated with traditional media content to online when they are simply distributing the content over the Internet? We should be holding to the same traditions of attributing the dollars to the media that originated the content. At the very least, counting the dollars this way would more accurately reflect what’s going on with the newspaper companies. Why are we counting dollars spent on NYTimes.com separately from those spent on the print edition of “The New York Times?” It’s misleading — and artificially impacts the value of the publication as a whole (print and electronic).
And if we make the decision — as wrong as I think it is — to continue to attribute ad dollars to the distribution channel rather than to the pool that the various media companies have traditionally been counted in, then we should at least be coherent about how those dollars distribute out. I want to understand where the growth of online media is coming from: what dollars are coming from content that would normally be attributed to TV, print, radio, out of home (OOH), and so on.
Below, how the world looks today. Keep in mind, the online display channel is really a mix of pure-play online media companies and traditional media companies being counted in the online bucket. (Note: I’m being intentionally abstract with the charts here. I’m illustrating a point, not making financial predictions. And what we call online advertising today is made up of online display and paid search. Here, I’ve broken them out.)
|Typical Ad Spend Categories|
As you can see, online display is typically represented as a bit smaller than search advertising in this attribution mechanism. If we were to take all the dollars within online display that (by my reckoning) should be attributed to their traditional media counterparts, that red bar would be significantly smaller. But at least we would be truthful.
If we were to properly attribute growth in online advertising going forward, what would that look like? We’d see all the traditional media growth attributed to the online display column. But if it were broken out, we could tell where the spend was originating. And it would more appropriately attribute value to the traditional media companies that funded the growth. Below is how I’d like to see it attributed, with growth broken out by attribution:
|Ideal Ad Spend Categories|
The dotted lines represent future growth. And as I said before, I think the analysts’ numbers are low, but I’m not putting a timeframe on my chart because I’m making a point about attribution, not amount of money.
One reason I think the numbers for online display in particular are low is the way analysts handle online video. It’s completely wonky. Should the attribution of online video include television content watched on a PC? What about live television content delivered over the Internet to a Media Center PC?
A better way to handle all of this would be to attribute the dollars to the company being paid the ad dollars. So if the ad spend is going to Yahoo or MSN (both pure-play online distribution) the dollars would appear in the “online display” bar. But if the dollars are going to NBC or ABC, the dollars should go in the “TV” bar.
This has more implications than just how dollars are attributed and the resulting impact on stock prices (not insignificant, especially if you’re a newspaper company.) Another huge implication is where growth in online advertising will come from. Over time, the chart I’ve created above will be far more accurate than not. As traditional media is distributed across the Internet, the online display column will grow (hopefully, with proper attribution).
This is important because the online display part of our industry is built on models that are much closer to traditional media: the sales process, the counting process, the way we count, the way people get paid. From an ad technology perspective, that’s important. Even as ad exchanges and other auction-based models for display advertising become more prominent, the requirements of the advertisers who buy traditional media will drive the financial models. In other words, I don’t see display, especially video advertising, moving to a predominantly pay-for-performance model any time soon. This will have an impact on how we build the systems that manage this industry going forward.