The New Inventory Problem, Part 1

Much has been written about the inventory management problems Web publishers face. Inventory is hard, if not impossible, to predict. Network latency and underdelivery constantly plague the ad operations process. And managing potential conflicts between multiple campaigns competing for the same section or same audience is a problem just beginning to be addressed by vendor solutions, such as Solbright’s AdSuite.

As with many things in life, just when you thought things were pretty tough in the inventory management world, they got even tougher. A serious issue has emerged in the industry that will have significant effect on how publishers manage inventory: frequency capping.

Over the past several months, online media buyers have finally started to put publishers on notice. Campaign frequency capping will move from being used sparingly to being used universally. Anyone who’s heard Jason Heller of Mass Transit Interactive speak on a panel over the past four years knows this isn’t a new issue (he’s long been shouting about this from the rooftops). But now, the problem is moving front and center.

Frequency capping is when agencies use cookies and their ad servers to ensure a user or browser only sees an ad a certain, predefined number of times. In practice, this means though a campaign may call for 2 million impressions on a certain site section, or certain sections on unrelated sites, publishers will ensure the campaign is never shown more than four times to any one browser. They “cap” the campaign at a frequency of four and stop serving ads to individual browsers once that cap is reached. In theory, this ensures the campaign will then have a reach of 500,000, which may be one of the advertiser’s objectives.

Sounds good in principle.

It’s in implementation that the collision of principle and practice can create lots of trouble for publishers. First, one of the not-so-secret secrets of Web publishing is that for most sites, a relatively small percentage of their total unique users (maybe 20 percent) generate the majority of their page views (maybe 80 percent).

To meet basic ad delivery requirements of 500,000 impressions, many publishers will rely on serving the same ad an enormous number of times to the same people. If a publisher’s total unique monthly audience reach in a section is 500,000 and it delivers the campaign normally, the 20 percent of visitors (100,000) who generate the majority of traffic each receive an average of 16 ad impressions. The remaining 400,000 receive the ad only once, on average. All too often, this is reality.

Why does it happen? For one, because of the publishers’ inability to forecast and manage their inventory. For another, most publishers, particularly those with the most valuable vertical audiences, don’t have enough inventory to deliver differently.

Buyers tolerate it because they have to and because online advertising generally works so well clients don’t notice this enormous inefficiency in their buys. Unfortunately (or fortunately, depending on your viewpoint), this is changing. More sophisticated clients are now either directly tracking campaign deliveries on a per-user basis or having their clients do it for them. They’re starting to ask questions. They’re starting to understand. Agencies that get it are now trying to optimize.

What does this mean for publishers? A lot.

It means they need better forecasting. Publishers will now not only have to be able to survive delivering frequency-capped campaigns but also need to be able to predict how much audience they have in certain sections, and what their anticipated frequency will be, relatively accurately.

It means publishers will need better audience segmentation and targeting tools. Once they’ve identified the anticipated reach and frequency of their audiences within certain sections, they’ll need the ability to deliver as promised, and to deliver smoothly. If they don’t, they’ll serve impressions to advertisers that will be turned back, or not counted, due to frequency capping on their end.

What’s the answer? Publishers must either develop home-grown solutions or use commercial tools, such as AdSuite to create reliable inventory forecasts and assure campaign delivery. They can also use audience management to attack the problem.

For example, publishers can use behavioral data to maximize the yield of capped campaigns by dispersing campaigns throughout underutilized inventory. They can also use behavioral and demographic data for targeting performance campaigns to match results of typical search programs.

Many think of behavioral targeting as a solution for high-value ad targeting alone. It’s more than that. Publishers will find understanding audience behaviors will help plan reach and frequency to facilitate frequency capping. This can become a technology application that will offer as much value, or more, over time as targeting will.

Increasingly, advertisers will understand the complexities of inventory and audience. They won’t sit still for multiple deliveries of the same impression to a narrow slice of audience. As frequency capping becomes a requirement rather than an option, publishers who don’t or can’t guarantee it will begin to lose revenue opportunities.

Meet Dave at ClickZ’s upcoming AdForum. Join us for Search Engine Strategies in August.

Vote for your favorite product or campaign from July 20 through close of business August 2.

Related reading