Summary

Web Analytics Tutorial

 

Appendix A – Making Reports More Usable

IN THIS APPENDIX
* Matching Expectations
   Web Server Configuration
   Adjusting Definitions to Improve Metrics
* Matching Business Organization
* Interfacing to Content Delivery Systems

Matching Expectations

In presenting your web analytics reports, it is important that the people reading them can understand what they mean. Much of web analytics is built on concepts, terms and units that are not familiar to people in other domains of expertise. Often, when presenting statistics to business, finance or marketing managers, you will need to organize and phrase your conclusions in terms of the business model or design model of your web site, which may not be the same as the request model that is in your web logs.

Some of the simplest things you may need to change are terms. For example, when presenting advertising results to a marketing manager, you may think of the number of ‘hits’ each ad received, while the marketing manager is concerned about ‘impressions.’ As we discussed in Lesson 1 - What is ‘Web Analytics?’ these are about the same thing (with variance for caches). Similarly, business managers may want to know how many ‘page views’ a given page has had, while Summary will list it as Pages or Hits (depending on the report.)

Summary produces many reports covering a broad range of web analytics concepts. This deluge of information can be confusing, especially to readers who are only interested in particular aspects of the performance of a web site. A simple solution in Summary Plus or Summary SP is to disable those reports that are not important to the audience. If you are building subreports you can limit each subreport to only contain the necessary reports. This makes the presentation much easier for the reader to manage.

Figure 1. A Sample Dashboard
Figure 1. Sometimes your readers just want
a “Dashboard” of important quantities.
Sometimes the only solution is to take the data from Summary’s reports and create a new report or presentation from it, specifically targeted to the people who are reading it. For example, maybe your CEO has asked you for a ‘Dashboard,’ like Figure 1, on his desk every Monday morning. This Dashboard is supposed to include the total number of visits for the previous week, the number of new registrations on the site and the number of purchases. In order to produce this, you would need to gather information from several reports. The Weekly Report provides the number of visits for the previous week. For the other two quantities you would have to configure subreports with a Goal defined for the registrations page and receipt page respectively. In each of these subreports, you can get the count from the Weekly Metrics report, Goals column. If you are handy with Perl (or another text-manipulation or scripting language) you could even write a script to automate this kind of extraction and reporting task.

Web Server Configuration

Sometimes, you can simplify your task by configuring Summary to reflect the expectations of the report readers. That way, you can just provide access to reports, without having to make custom aggregate ones. One common example of this is when a web server is configured to allow different requests for a give item. The widgetmanager.com site has a /products/ directory that contains information about their products, naturally. To simplify links and reduce log file size, they have set up the web server to map all request to /p/ to the products directory. So a promotion might say to click on http://www.widgetmanager.com/products/new_tool.html, but the site’s links all point to http://www.widgetmanager.com/p/new_tool.html. The log file is full of requests to both locations, but most of the people at widgetmanager.com only think about the pages in the /products directory. In order to make the reports clearer, the site manager can create an Request Alias, in Summary’s configuration to convert all requests like /p/* to /products/$1 (the ‘*’ matches anything and the ‘$1’ puts that into the result – see the Summary manual for details on setting up Aliases.)

There are many more web server configurations that you can use to make up for the difference between what shows up in the logs and how your readers think of the web site. One more example of this is the canonical name of your web site host. In our example company, most people think of the site as ‘www.widgetmanager.com;’ however, the server also responds to ‘widgetmanager.com.’ To distinguish local referrers from remote ones, Summary needs to be told that these are both valid domains for the web site. Because this is very common, Summary asks for the primary domain and then allows others to be listed, separately. If the site manager configures Summary to use ‘www.widgetmanager.com;’ as the primary domain, then all the requests in reports with be linked to this host name, which is what the people at widgetmanager.com expect.

Adjusting Definitions to Improve Metrics

Figure 2. Request Types Configuration
Figure 2. By changing the Request Types
Configuration you can match analytics
units to your web site design.
When Summary counts hits, it looks at the name of the file and decides whether it is a page, download, graphic or other hit. In the Summary configuration, you can change these definitions under Request Types (see Figure 2). Files are categorized based on the extension (the part of the filename after the last ‘.’). If you have implemented some custom graphics or animation types (like SVG) then you would want to add those to ‘Graphics file name extensions.’ Summary counts cascading stylesheets (CSS) files as graphics too, but you might want to count those as page views, or maybe as ‘other’ (in which case they would not be in any of the listed types). If you have PDF files on your site that are available for download, Summary’s default settings will work well. However, if a large portion of your site is PDF files, you might want to move that type to ‘Page file name extensions.’ This can be especially useful when your visitors all have newer versions of Adobe Acrobat Reader, as the reader will make a separate request for each page (or group of pages) within the PDF file. This makes the PDF look more like a series of page requests to your web site than a download.



Table of Contents | 1: What is Web Analytics? | 2: Where are My Visitors Coming From? | 3: Search Engines | 4: Advertising | 5: Revenue Modeling | 6: Design Considerations | 7: Determining Visitor Behavior Patterns | 8: Examining Subsets of Traffic  | 9: Incorporating Business Goals | 10: Bandwidth Management | 11: Site and Server Diagnostics | 12: Investigating Troublemakers | Appendix A: Making Reports More Usable | Appendix B: Technical Details of Metric Accuracy

Copyright 2002 by Summary.Net - Updated 16.Apr.2002