SEM Reach and Frequency

In traditional brand-centric media buying, reach and frequency are core metrics used for planning and evaluating campaigns. Reach is the percentage of a target audience “reached” by the campaign (some marketers use numeric reach, percentage is more common). Frequency is the number of times the average “reached” person sees your ad.

In traditional offline media, media buyers are forced to estimate reach and frequency based on formulas derived from syndicated and custom research. Each medium has a method to help planners get as close as possible to actual reach and frequency.

In Internet marketing, frequency and reach can often be quantitatively measured on banners, interstitials, and rich media. Highly accurate control and measurement is possible, particularly when a portion of a buy runs on one network or originates from one ad server.

Agencies that buy online media are wondering how reach and frequency might apply to paid search campaigns. Brand marketers want to leverage search engine marketing (SEM), too. Before determining if a reach metric is appropriate for brand marketers in SEM, let’s redefine “target market,” or denominator of the reach equation, for SEM.

In traditional media, the target market denominator is based on demo- or psychographics. For example, the media plan is to reach 60 percent of recent homebuyers age 30 to 45. In search marketing, a paid placement campaign is keyword-/key-phrase-centric. The audience you want to reach are those who search a particular set of terms at search portals in a defined universe.

Let’s say your key phrase is “cell phone.” If 100 people search for “cell phone” and 50 of them see your listing, your reach is 50 percent. For high reach, listings must be in the top at major engines. If you use auction engines, the top three positions generally maximize reach across the networks. However, MSN, one of Overture’s largest syndication partners, doesn’t always display Overture results. Lower-priced bids frequently don’t appear in MSN, making reach difficult to confirm.

Paid listing interrelationships of engines also make calculating reach difficult. Engine reach is reported based on monthly search behavior, resulting in reach numbers that add up to well over 100 percent. Searchers use more than one engine, engines power each other’s paid results. Overture’s full reach percentage is 80 percent. Google’s reach for AdWords is similar.

Media buyers must keep other factors in mind when it comes to paid placement to avoid crucial mistakes. For SEM, lots of keywords and key phrases can apply to a site and, therefore, may belong in a campaign. But the more keywords in your campaign, the more listings you need to have top positions to achieve the target reach as a percentage of total available inventory (for the full set of appropriate keywords). Each search portal or paid placement network has a set number of searches for any given keyword or key phrase. To get the listing in front of a high percentage of those searchers at the primary portals, listings must be visible for the appropriate percentage of time.

To achieve a specific reach percentage, misguided media planners narrow the number of keywords down from the true universe into an artificially small target audience. Keeping keywords in a specific position (or a range of heavily syndicated positions) gets extremely expensive. With $30,000 to spend to achieve a reach objective and a full set of keywords costing $65,000 to stay on top (high reach, highly syndicated positions), a media buyer might slash a large buy down to only a few keywords. It’s what the buyer can afford to remain in a premium (high reach) position for the month.

Do media buyers who don’t run appropriately broad campaigns delude themselves and fail to serve their clients? Probably. By fudging the reach number, some planners forget the true purpose of the reach metric. The principle is to put the message in front of as many target market members as possible. Artificially shrinking the target market definition to make numbers work is a big mistake. When presented with two definitions of target audience based on search listings, smart clients select “true reach” over “media reach.”

A simple example illustrates the folly of fudging reach numbers. The keyword “recipe” was searched 1,275,142 times on Overture in April. The top three positions have a CPC of approximately $0.30. A brand marketer with a recognizable name and a good recipe database might expect a 4 percent CTR on those impressions, resulting in a $15,000 spend (assuming 100 percent reach). Add to that search inventory for its brand name and category keyword “cheese” (52,195 searches at $0.25). With a 50 percent reach goal in Overture’s distribution network, the monthly spend is significant.

Unless the agency wrote the listings with a heavy branding message to discourage click-throughs (to achieve reach with fewer click-throughs, which can get a listing removed from both Google and Overture), most of the listings’ media impact would happen on the advertiser’s site. Once searchers reach the advertiser site they become immersed in the brand, perhaps printing recipes or registering for sweepstakes.

If the agency had used a true reach concept, it would have instead broadened the keyword list to include several hundred different kinds of recipes, several kinds of cheese, and lots of permutations of the brand name. For example, the following keywords are available (all can be coupled with “recipe” to create a key phrase): “chicken,” “Mexican,” “low-carb,” “crock pot,” “food,” “cake,” “vegetarian,” “low-fat,” “healthy,” “dessert,” “free,” “shrimp,” “diabetic,” “barbecue,” “easy,” “restaurant,” “punch,” “top secret”… you get the point. Not all are appropriate for this campaign, but many would deliver equal (or better) visitors to the site.

Instead of a position-based strategy relying on reach as a percentage of impressions, the agency could deliver perhaps double the visits to the client site by lowering the CPC on the primary terms and finding less expensive keyword listings equally or better targeted to the client’s site content. Same budget, more visitors. “Shrimp recipe” only costs $0.10 CPC at top position.

The agency might also use directory inclusion, such as LookSmart, or even submit the site to search engines through an XML paid inclusion program (if the CPC is appropriate). With a broad campaign, more individuals in the true target audience have the opportunity to interact with the brand. The budget would have been spent far more efficiently.

Sophisticated brand managers might know certain behaviors on their site that are well correlated with brand metrics improvement or increased sales. A BEI-style metric can be used to further optimize the campaign, taking it even further than optimizing on maximum true reach.

Is it more work to manage hundreds of keyword listings, directory listings, and XML listings? Yes. Do brand marketers deserve the very best SEM campaigns possible?

I know how the brand manager and marketing VP would answer.

Meet Kevin at the Jupiter ClickZ Advertising Forum in New York City on July 30 and 31.