So many impressions go unsold that the average cost-per-thousand (CPM) rate for online media — given our $8 billion in ad revenue — should be around $0.39. That’s the figure you get by dividing the amount of money spent on ads by the number of impressions available (the macroeconomists’ method for determining the rough point of equilibrium in a free market).
What a renaissance we would see if sites warmed up to that idea. It’s pretty easy for brand managers today to throw their budgets away on print and TV when CPM rates seem comparable to those of online. If online media undercut the pricing by one or two orders of magnitude, those brand managers would have to write a lot of strange memos to justify continuing putting those dollars into traditional media.
Sites would multiply revenue, winning a share of budget from other media. The industry would see greater economies of scale.
So, Why Not?
It comes down to one simple obstacle: ad serving. The cost of serving 1,000 ads ranges from about $0.30 to $1.00. This saps the brunt of the revenue from sites and creates a different scale of pricing that disallows online media to significantly undercut traditional media.
A site that could make a 25 percent margin on a $0.40 CPM (winning $0.10) finds itself with a margin of minus 50 percent when it factors in ad serving. To date, the solution has been to raise the CPM price to account for serving. But when the site doubles its prices to account for serving, its value proposition begins to look less attractive. Heck, you can do print for that cost.
When most homeowners switched from coal heat to oil heat in the 19th century, they did so partly for convenience but mainly for cost. Had companies added a 100 percent surcharge back in the 1890s, we’d still be burning coal. It took a big price advantage to persuade people to use newfangled technology. Only after a certain number of people joined the trend, did economies of scale bring the cost of delivering oil much, much lower. Initially, delivery costs were very high. Now, they account for only a small portion of the cost.
In the media world, we’re still burning coal. We have very poor economies of scale. Ad serving keeps us there.
DoubleClick: A Case Study
DoubleClick gets about $21 million each month in ad sales, according to Securities and Exchange Commission filings. It serves about 55 billion impressions in that time. That works out satisfyingly close to our industry average estimate (it appears to receive an average of $0.387 cents for 1,000 impressions after subtracting serving costs and the amount due the site).
The problem is, the company doesn’t go around charging people $0.39 CPMs. Instead, it charges much, much higher rates and sells a much smaller portion of its inventory. Rather than dazzle advertisers with deals to bring in more revenue, it seeks to acquire the same amount of revenue with fewer impressions. This minimizes the overhead cost of banner serving and other transaction costs, but it also minimizes the value proposition of online advertising to brand managers and ad buyers.
I hear salespeople in the industry complaining there’s too much inventory out there. That tells me they’re shortchanging the value proposition of the medium to their clients.
If these people ran a theater, they’d admit everyone for $5 but charge a mandatory $45 for a playbill. Currently, due to the overhead costs, they’d rather sell tickets to 10 people at $50 each than sell 100 tickets at $5 each. What they fail to realize is there’s a huge potential audience out there. They might become regular customers if the price were to come down.
Will pricing reformation start at DoubleClick or any other network with an interest in a technology solution such as ad serving? Don’t hold your breath. The bottom line for DoubleClick is that it makes roughly a 70 percent gross margin on the serving part of the business and only about a 30 percent gross margin on the media sales side. In other words, ad serving may be a scourge for the rest of the industry, but it’s DoubleClick’s bread and butter. Its solution to the issue harks back to the Standard Oil horizontal monopoly of the 19th century, when it made money on related services, such as delivery. Owning other oil-related services helped keep out competition and kept prices nice and high.
For individual sites, though, DoubleClick could have made a bold move and radically changed the pricing model, that is, increase the scale of the deals for the same revenue. Except it couldn’t. Banner serving costs prevent it.
Here’s what those sites should do. They should use ad serving (and charge for it) only for direct response buys, to get the necessary tracking. For branding buys, they should move to an audience-based measurement, such as ratings. That will allow them to sell huge numbers of impressions for no incremental delivery cost. When they start offering a $0.40 CPM, those marketing budgets will find them.