Nov/Dec 2005

How do I interpret the data in my organization's Web server report?

Answer by Dr. David Crooke, Founder and CTO, Convio

In addition to gathering constituent data and tracking results from online activities such as fundraising or advocacy campaigns, every nonprofit organization should regularly analyze its Web server log report. By doing so, an organization has insight into several important indicators of its Web site's overall effectiveness, including:

• Types of information on the Web site that are most valuable to visitors;
• Whether the site is gaining or losing visitors;
• How much time visitors are spending on the site; and
• What areas of the site need improvement.

In short, while they do not substitute application-specific data such as constituent subscriptions and donation rates, Web server logs do capture meaningful data. Following are some fundamental Web server statistics that can be helpful for optimizing any organization's Web site.

It's about decisions, not numbers

Organizations often concern themselves with questions like, "How many 'hits' did we get today?" However, it can be all too easy to get caught up in numbers like these. The number of site hits each day is not useful for its own sake. To get real value from Web log data, organizations always should ask the question "What will we change based on this data?" For example, if the number of site visitors from month to month appears to be waning, it might be an indication that the information on the site is stale and needs to be updated or that the site navigation needs to be improved because visitors are having trouble finding the information they are seeking.

It's all relative

At its very basic level, the data in a Web server log tracks number of "hits," or page requests, on a site. It indicates that a browser at a particular IP address requested a particular Web page on the site at a particular time. Most of the interesting information provided by a Web reporting tool (time spent on site, number of visitors, etc.) is estimated, calculated and extrapolated from this raw log data using algorithms that include different criteria, depending on the reporting tool that an organization uses to view the data.

The consequence is that most of these calculated metrics are not comparable from one reporting tool to another. The only truly valid comparisons are relative between reports generated for the same site with the same tool. In other words, using the same tool and viewing the same metric, did the metric go up or down in a given period, and if so, what may have caused it?

Top pages

The "top pages viewed" report found in most Web reporting tools can show two different, but important, indicators. If it lists a page that an organization might not consider a particularly important one, it could be that visitors are interested in the page topic and that additional information about the topic may be warranted. Pages that an organization expects to see in this report may be high in the ranking because the content is interesting and relevant to visitors, or simply because the pages are well-promoted.

Visitors, sessions and exit pages

Most reporting tools aggregate hits from the same IP address within a certain period and assume it is a contiguous session from a single person— a metric often called a session or visit. The most common Web metrics are the number of visitors and session duration, but those are only general success indicators.

Much more useful is a report on top exit pages, the last page someone views before leaving the site, because it may indicate a need for a change in content. For example, if a top exit page includes a donation appeal, the call to action may need to be more compelling to hold visitors' attention and motivate them to respond.

Search engines and referring sites

The largest single source of unprompted visitors is public search engines such as Google. When someone goes to a site from a search engine, the "referrer" column in the Web server log will contain the search engine's URL, or Web address. Many Web log tools report on the top keywords and phrases used in the search engine to find a site— a useful guide for an organization to topics that may deserve additional site content.

More importantly, an organization should look at what keywords are missing— what words it associates with its organization that are not showing up— and take steps to make those keywords better represented in its site content. Those words represent a potential audience the organization is not reaching because search engines are sending people elsewhere.

Conclusion

Every Web site has Web server data that can and should be analyzed. Organizations should look closely at the metrics that matter most to them, and use that data to make informed decisions on how to improve their sites and ensure continued site traffic that can drive ongoing support.

Have a colleague who might be interested in this topic? Why not forward this article?

Return to Convio Connection Newsletter page