Measuring your public relations results against that of your competitors not only improves your program, but also boosts the bottom line.

“The holy grail in PR measurement has always been how do we correlate outputs with business outcomes, and you can do that in a competitive framework,” says Gary Getto, executive vice president of Surveillance Data, a data and analytics company.

After conducting more than 100 studies that analyzed a total of 4 million articles, Surveillance Data proved that a correlation exists between sales and share of discussion-the quantity and quality of articles you have versus those of your competition.

“If you only look at your publicity and try to generate a correlation between outcomes, you won’t find one,” Getto says. “It’s only when you look at what you’re doing in a competitive framework that you see a correlation.”

“While it sounds very rudimentary, it’s a very simple and a powerful way to analyze the media,” says Bruce Jeffries-Fox, president of measurement consultancy Jeffries-Fox Associates.

Jeffries-Fox applauds Getto’s finding and hopes it might steer the industry away from touting clip counts as a valid measure of PR success.

“In the past, people used article counts-and only those that were favorable-and tried to match that with sales,” he says. “What [Getto has] done is very valuable. It helps move the industry to a more scientific measure if this holds up in more studies.”

Here are a few examples from Surveillance Data’s research:

Getto used Exogin software to analyze 106,000 articles about Excel Communications and its competitors. Exogin measures the tone of each article as well as the number of articles about each company.

Getto’s analysis revealed that Excel’s media coverage about its long-distance service correlated to its sales.

“When we only looked at Excel’s coverage, there was very little correlation, but that improved when we measured share of discussion,” Getto says.

The correlation between the number of articles about Excel and its sales was .06, but the percentage of articles about Excel over its competitors was .33.

“It was revealing to see that share of discussion impacted sales so dramatically,” says Travis Jacobsen, former executive director of Excel’s North American corporate communications. “It’s significant because it teaches PR professionals that they have to consider sales as a primary emphasis of why they do PR.”

Getto took the study one step further and instead of article counts, used the number of impressions to calculate share of discussion. The correlation between sales and impressions declined slightly to .30.

“We thought that as we improved our metrics, we’d improve our correlations, but if there isn’t enough substance, that may reduce the correlation,” he says. “Broadcast stories about Excel generated a lot of impressions, but had little impact because they were short mentions.”

When ad value equivalency was used to compute Excel’s share of discussion, the correlation rose to .38.

“We have since done numerous studies and we are confident in saying that ad value correlates to business outcomes better than article counts or impressions,” Getto says.

Analyzing share of discussion can pinpoint weaknesses in your PR program.

Surveillance Data analyzed 60,000 articles about over-the-counter vitamins from January 1999 through April 2002 for an international packaged-goods company.

Research showed that as share of discussion dropped, sales declined.

“The client believed that discussion drove sales, but they had no way of quantifying it and therefore weren’t managing the business to generate discussion,” Getto says.

Further analysis revealed that there were more and better articles about herbal remedies than vitamins.

“We could show the client that a lot of their competition was coming from herbals and single-entity vitamins, and they needed to generate talk about multivitamins,” Getto says.

Not only is share of discussion tied to company sales, it can predict sales patterns.

When Surveillance Data analyzed 96,000 articles about hormone replacement therapy over a four-year period, it discovered that new prescription sales from menopausal women declined as the level of discussion dropped.

The study also showed that share of discussion could predict an increase in new prescription sales.

“Peak periods of discussion were reflected 11 weeks in the future,” Getto says. “A lot of people have questioned whether discussion leads or trails sales, but this shows that discussion takes place beforehand and sales follow.”

Gary Getto, 610.834.0800, ggetto@survdata.com
Bruce Jeffries-Fox, 908.766.1414, bruce@jeffries-fox.com