Yes, we know that digital is more than a messaging medium. Yes, we know that expressing digital's impact in terms of reach and frequency sells the medium short. But if you're planning to make a case for digital display to advertisers that have unified delivery metrics across all their other display channels -- like television, radio and print -- you're going to need R/F/GRP forecasting.

The problem is, the R/F/GRP model doesn't translate very well into digital. Our definition of "reach" is problematic, in particular. Here's the gist of the problem: When you place an ad in Sports Illustrated, you can pretty much assume that most readers of the publication will see your ad. When you place an ad on MSN, that's not necessarily true. To figure out how many people actually have an opportunity to see the ad, a media planner has to ask a lot of questions: How many impressions were bought? How are they targeted? In which sections of the site will they be running? And there are many more.

Existing solutions for R/F/GRP forecasting are rudimentary at best. They do a reasonable forecasting job if buys are placed on large sites in a run-of-site capacity. Add in targeting, frequency capping, editorial sponsorships or any one of a list of potential buy characteristics and the numbers begin to fall apart, particularly if the demographic against which the buy is being evaluated ends up being something other than a simple age and gender combination.

There are a few reasons why numbers fall apart with the existing solutions. As we've just seen, one of them is the incongruity of the definitions of "reach" and "opportunity to see" between online and offline media. Another is how we're determining who is visiting various ad-supported websites, where they're going within these sites and how often. With sample-based R/F/GRP forecasters, it's something of an issue that enough people falling into a specific demographic might not visit a certain section of a certain website enough to produce a projectable sample.

Mediasmith's David Smith probably cares more about reach and frequency than any two agency people out there. It's been something of a pet issue of his for many years, and it's doubtful that any one person has done more than Smith to translate reach and frequency into the language of digital advertising.

Smith believes the right R/F/GRP solution might come from Quantcast next year, and the company's census-based approach is preferable to sample-based solutions from other providers. Accounting for behaviorally targeted campaigns remains something of an enigma, but "[for] other targets, with Quantcast doing a census rather than a sample, audience definitions will be built in."

Advertisers need an accurate R/F/GRP tool in order to be able to engage in any sort of reliable scenario planning. They already get it in offline media, and they know that they can call their ad agency and get same- or next-day turnaround on a variety of potential scenarios for delivering media weight at a number of different levels. Digital display will miss out on a lot of opportunities until tools are at least at parity with offline.