Branding Metrics for Search Engine Marketing?

12 december 2002, 08:55

Brand marketers have an advantage in their ability to maximize the volume of clicks they receive from search engines. But it’s not all roses. Brands have it tough using metrics to measure effectiveness and to adjust search engine campaigns. Sure, search text listings have some branding impact. Yet it’s far more likely a brand advertiser wants to create a brand experience on its site, not on a search results page. That requires post-click metrics.

That’s a dilemma. A marketing professional with a branding objective wants to use search engine marketing (SEM) efficiently. In an e-commerce or business-to-business (B2B) SEM campaign, not all clicks are created equal. Metrics are used to differentiate between the best and less-valuable traffic. Most marketers are forced to “do a study” to compare media effectiveness or different creative executions. Many studies are cosponsored by industry organizations such as the Interactive Advertising Bureau, which coordinated research to examine online media as part of effective overall campaigns, or Dynamic Logic, which employs metrics that compare creative and media. None of these metrics lends itself to effectively measure the differential between keyword/position/engine combinations search engine marketers use.

We need easily measurable data that correlate with branding metrics to make necessary, ongoing adjustments in search engine campaigns. I propose a new metric, based on a behavior mapping strategy, called the Branding Effectiveness Index, or BEI (pronounced “buy”). BEI is a work in progress but has already proven useful in comparing campaign results.

Each marketer’s branding needs are unique. Each determines which measurable actions and metrics combine to create his own BEI. The result is a flexible metric that measures branding impact. It won’t measure ad effectiveness, rather the effectiveness of a given site on a visitor driven by an ad. A visitor interacts with the brand on the site and, ideally, positively changes the way she feels about that brand. If results aren’t positive, there may be serious problems. You probably shouldn’t spend on media until your site provides a good user experience.

A site does the branding in this case. So instead of using metrics designed to measure advertising brand effectiveness, we use the following proxies to build our own BEI:

– Purchase (see McDonald’s example below)

– Registration

– Newsletter subscription

– Information request

– Page view

– Content depth or involvement

– Webinar view

BEI is a flexible, relative metric. It’s used to compare campaign elements to achieve optimized results.

To generate your own BEI, set the most valuable measurable action to a value of 1. Other significant actions are assigned a percentage value of that variable, based on importance. For example, if a brochure request is 70 percent as important as newsletter registration, assign it a value of 0.7. All actions are added up and multiplied by their factor. These are totaled and divided by spending to determine BEI.

Any campaign or campaign segment (even a single keyword listing) can be measured by how well it did achieving BEI. Using the example above, 20 newsletter registrations and 10 newsletter requests for a $1,000 spend results in a BEI of ((20*1)+(10*.7))/1,000, or 0.027. If the number of newsletter registrations rises to 40 for a different campaign segment, BEI is 0.047. This value is useful in comparing campaign segments when using the same formula.

If a second campaign with a different set of measurable actions exists, results aren’t comparable. A simpler way to think of BEI is as a variable, like a weighted average cost per action (CPA) that incorporates a host of post-click actions. I’m working with statisticians on how BEI might be normalized to allow comparison across various sites. That requires a more complex formula to set “factors” for actions.

Let’s look at AdAge’s top brands to see what metrics they might use to build BEI.

AT&T: Though AT&T is a brand marketer, customers can buy online from it. The company might include a variety of other metrics in a BEI campaign index. If customer acquisition is the highest valued action (1), loyalty program signup might be 0.4 and interaction with a plans and services page might be 0.15.

Chevrolet: Chevrolet has a brochure request, a dealer locator, a “Contact Us” form, content, and a clubs listing. All these measurable actions could be combined to form an overall BEI.

Ford: Ford emphasizes newsworthy information, such as a special financing rate. Perhaps that’s its most important metric for building BEI; or maybe it’s a visit to “Build Your Ford.”

Toyota: Toyota is pushing the Camry. The company may use interaction with the brand’s pages as a large BEI factor. If an event such as a page view is likely to repeat, it can be assigned a BEI factor relative to another, less frequent action. Different pages can be assigned different factors as well. Sweepstakes entries might have a factor of 1, page views within the Camry section a factor of 0.03 per page view.

McDonald’s: McDonald’s has a restaurant locator and may assign that the factor of 1, purchase of merchandise at 0.9, and visits to the “Treasure Planet Happy Meal” site area a 0.2.

I’ll spare you more examples.

Whether you think BEI is a wonderful new way of combining existing metrics to capture a site’s branding impact or the nonsensical ravings of a guy who’s looked at too much data, the fact remains Web sites impact brands.

Marketers should plan and adjust campaigns based on brand awareness measurements. My branding clients are glad to have an empirical way to effectively manage their paid search campaigns.


Marco Derksen
Partner bij Upstream

Oprichter/partner Upstream, Marketingfacts, Arnhem Direct, SportNext, TravelNext, RvT VPRO, Bestuur Luxor Live, social business, onderwijs, fotografie en vader!


Marketingfacts. Elke dag vers. Mis niks!