How To Measure Cyberspace

Part 2: Six Sets Of Tools
To Answer Your Questions

By Katharine Delahaye Paine

What kind of rulers do you need in cyberspace?

The precise tool or technique you use to measure your image or results in cyberspace depends on the objective you select. My ideal toolbox has half a dozen measurement devices.

1. A tool to find out what the cyber media is writing about you and what your constituencies are seeing about you (and your competition).
   This type of tool comes under the same category as "clipping services" and there are a number of them out there. Companies like CyberAlert can look at specific publications as well as the whole universe of Internet publications. (See the article elsewhere in this issue: Web Clipping Services Face Off: Cyberalert vs. Northern Light)
   
   One problem most of these services have is archival stories. While it is a plus to an organization to have archival stories on the web, since this increases their chance of being seen, the drawback is that it will skew your numbers if you are trying to develop accurate metrics for your cyber presence. If you don't want them included you need to contact the service to ensure that they exclude old data.
   
   Another problem that at least one firm, CyberAlert, has solved is the problem of "dead URLs." As you know, URLs become obsolete rather quickly. What CyberAlert does is "capture" the text from the URL as they find it and store it, so that if that URL goes dead, you will still have the text from the original hit.
   
   
2. A tool to find out the size of the impact you are having.
   This is where companies like Media Metrix and Nielsen//Netratings, Standard Rate and Data Service (SRDS) and Business Publishers Association come in. These firms, as well as the Advertising Research Foundation and the Internet Advertising Bureau, are all working on developing standards to determine a standard way of determining how many potential eyeballs you are reaching. Part of the problem is that while the larger web sites are represented in these ratings, many smaller sites that may have a huge impact on your audience may not be represented.
   
   I've always maintained that Hits stands for "How Idiots Track Success" and as a metric they are virtually worthless. The emerging standard is the monthly "unique audience"—the number of actual people that visit a site during a month (if 1 person visits 10 times during a month, that's 1 unique visitor). It is somewhat similar to print circulations, although perhaps a weekly unique audience figure would be even more similar.
   
   As with print circulations, the figure doesn't represent the actual number of people who saw a story; it represents the number of people who had the opportunity to see it. Another caveat is that Nielsen//Netratings figures are based on panels of some 30,000 Internet surfers who have software attached to their computers that track where they go. Similar to Nielsen TV ratings, it is a random sample of actual users.
   
   If Nielsen//Netratings doesn't list a particular website that is critical to you, BPA International, ABC Interactive or SRDS Interactive might list a unique audience figure; the next best thing is to take a self-reported unique audience figure from the web site itself.
   
   If all else fails and you’re stuck with hits, visitors or page impressions, you can still extrapolate unique audience figures from monthly page impressions or number of visitors using a multiplier. The multiplier is derived from looking at the relationship between page impressions, number of visitors and unique visitors in Nielsen//Netratings' database over a four-month period. After looking at several hundred publications, the ratio of unique audience to page impressions was .11, and the ratio of unique audience to visitors was .46 So when monthly unique audience figures are absolutely not available, monthly page impressions can be multiplied by .11, or monthly visitors can be multiplied by .46 to extrapolate monthly unique visitors.
   
   If the publication you are trying to track gives you weekly page impressions/requests, those have less similarity to print circulations than unique audience. Weekly page impressions are a good relative measure of the traffic/importance/audience size of various websites, but are useless as a comparison to print or broadcast media.
   
   The holy grail is actual unique page requests for a particular story. While that number might be much lower than print circulations, (since for print pubs we never know if someone reads a particular article), it would be a far more accurate determination of how many people have actually seen a story. In lieu of that, unique audience is the most comparable figure, because it's people, not double counted, who may see your story.
   
   Regardless of which figures you use, it is always dangerous to compare Internet audience figures with print audience figures because of the different methods of data retrieval.
   
   
3. A tool to find out what your constituencies are saying about you.
   To find out what people are saying about you, you need to monitor their chats and discussions in Usenet groups and other chat areas. Services like CyberAlert, E-Watch, Burrell's Cybertalk and Autonomy all gather and process the postings that your constituencies are making on Usenet discussion groups and other Internet chat areas. In general, the problem we face is too much data rather than too little.
   
   E-Watch was the first to deliver discussion group data, but they quickly overwhelmed their customers with data. CyberAlert and Autonomy have made great strides in filtering data, use highly sophisticated algorithms and promise delivery of only those areas or those topics that you want. It's a good idea to work with your vendor of choice to specify which sites you want to search, as opposed to searching the entire universe.
   
   Once you've gathered the postings, your next challenge is to analyze them to determine what they are actually saying. Studies have shown that people typically go online to chat about news that they've heard one to two weeks after an announcement, so if you're measuring a specific event, you need to monitor chat rooms for at least a month after the event. You should analyze the discussion content for tone, positioning and messages, just as you would your print or broadcast media coverage. One detail that technology has yet to provide us with is how many people are viewing each discussion group posting. For now, you can only monitor the actual number of postings, not their total impact.
   
   
4. A tool to determine what your constituencies think about you.
   
   The Internet has made it possible to get instant feedback from your constituencies, and there are now dozens of companies that offer web-based surveying tools. Typically these tools take two forms:
   1. Web-based programs that enable you or your web master to either put up a web site, or add a page to your web site to gather data from visitors. Examples of such products are Survey Said, Survey Connect, Survey Tracker, Inquisite, Survey Solutions, and Web Forms.
   2. E-mail services such as Custominsight.com offer automated questionnaire development email distribution. An excellent overview of these products was compiled by Ephraim Schecter and Henry Schaffer from the University of North Carolina and can be found at: http://ias.ga.unc.edu/~eis/websvyware.html. Additionally, David Solomon at Office of Medical Education Research and Development and the Department of Medicine, College of Human Medicine, Michigan State University has written a paper that outlines the challenges and opportunities of web-based surveys that can be found at http://ericae.net/pare/getvn.asp?v=7&n=19.
      
      
5. A tool to determine what action, if any, your constituencies are taking.
   
   Behavioral change is both the loftiest goal and the hardest to measure. However, new services from companies such as Loyalty Builders are making that task ever easier. Loyalty Builders takes customer transaction data and maps it over time to determine what impacts customer purchases and loyalty. When Loyalty Builders data is tracked against PR and Web activity, you can determine what impact, if any, the Web activity had on customer purchase. Other integrated marketing measurement firms such as Millward Brown and MMI are also beginning to look at web marketing results in conjunction with other marketing activities.
   
   If you don't have access to either the budget or data required by such firms, you could always do it yourself, which is what Miller Shandwick did for Kodak. Unsatisfied with conventional web tracking systems, Miller Shandwick manually analyzed referring URLs for Kodak's Picture Playground site to determine which were coming from Web publications. Some 208 hours later, they determined that during the promotion, 5% of the total traffic coming to the site was specifically from Miller Shandwick's online PR efforts.
   
   
6. A tool to determine whether it's all worth it.
   
   At some point, you need to put all of the above measures to the ultimate test: Does it help the business or ultimate goals of your organization? This involves factoring in the costs for the results. You need to track criteria such as "Marketing or PR cost per new customer acquired," or "Cost per minute spent with prospect," to help you determine what is or is not working. You can also compare the costs of on-line PR with other Internet traffic generators, such as banner ads and sponsored placements, to determine a relative value for your on-line PR efforts.

Go ahead to Part 3 of this article:
The Seven Basic Steps for Success, No Matter What Measurement Device You Use.

Go back to Part 1 of this article:
Understand The Challenges and Your Objectives.