Refining Analysis

You can get meaningful reports from your web activity data without any additional configuration. However, Webtrends provides a number of features that can help you focus your analysis, target the correct data, and improve analysis performance. This article describes some of these features.

Using Profile-Level Filters

Because Webtrends Analytics is designed to analyze large volumes of web data, an important part of configuring Webtrends is identifying and isolating exactly the data you need to make business decisions. Using filters to narrow down the data Webtrends analyzes for profiles and custom reports can help you focus reporting much more effectively. Because Webtrends analyzes only the relevant data, the result of careful filtering is more meaningful, readable reports as well as improved performance. This section describes how to use each type of Webtrends filter and how to combine them to customize the data you analyze. For more information about using specific filter elements to isolate business data, see the Webtrends Guide to Web Analytics.

For the purposes of this article, we will refer to filters applied at the profile level as profile-level filters to distinguish them from custom report filters, which only filter data at the individual report level.

Using Hit and Visit Filters

You can filter Webtrends Analytics data by including or excluding either individual hits or entire visitor sessions.

Hit filters include or exclude raw data generated by individual actions on a Web site. A hit is an individual request for a page, file, or image. For example, when a user requests a Web page, the user can incidentally generate hits to the .htm page itself, to several .gif image files, and to the .css file that stores the formatting for the page. A Hit filter includes or excludes each of these hits individually. For example, if you create a Hit filter to exclude all .gif, .jpg, and .bmp files, Webtrends analyzes data about page requests, but ignores data about requests for the images embedded in pages.

Visit filters include or exclude data that has already been grouped into visits. A visit is a single session of activity identified with an individual user, including all the page requests and hits logged during that session.

Profile-Level and Custom Report-Level Visit Filters

When you use a profile-level visit filter, Webtrends checks the first hit of a visit against the filter and then includes or excludes all the data contained in the visit based on whether it matches the criteria. For example, if you create a profile-level filter to include all visits that use the entry page www.zedesco.com\sale.htm, Webtrends checks the first hit in the visit to determine the entry page, analyzes only data for visits that originated on the specified entry page, and ignores all data associated with visits that originated on other pages. This works because profile-level visit filters track criteria that occur on the first hit of the visit.

When filtering custom report visit filters, however, Webtrends checks all hits in a visit to match visit criteria. For example, to filter URL Viewed in Visit, Webtrends needs to check every hit to determine if the specified URL was accessed during the visit.

Combining Hit and Visit Filters

When you combine hit and visit filters in the same profile or custom report, Webtrends processes the hit filters first and then the visit filters.

Include and Exclude Filters

When you create a filter, you specify whether to include or exclude the filtered information. Include filters specify that analysis includes only the information specified in the filter. For example, if you analyze a profile using an Include filter that specifies domains in the UK, reports based on the profile include only data from domains that include the string *.UK. Exclude filters specify that analysis includes all data except the information specified in the filter. For example, if you analyze a profile using an Exclude filter that specifies domains in the UK, reports based on the profile include all the data in the Web activity data file except data logged for domains that include the string *.UK.

Combining Include and Exclude Filters

You can use multiple filters of the same type (include or exclude) or combine different types in the same custom report or profile. For example, if you want to determine how much of your domestic Web server activity comes from organizations other than the military, you can create an include filter for all U.S. user addresses and an exclude filter for all military user addresses. Webtrends processes the include filters first and then the exclude filters. For any given activity, if both an include filter and an exclude filter match data in the Web activity data file, the exclude filter prevails and the data is excluded.

When you use multiple filters of the same type (multiple include filters or multiple exclude filters), if the information meets the criteria of any of the filters, it is included. For example, if you combine an include filter definition of URL=\test.htm with an include filter definition of returncode=304, Webtrends includes all hits that have either a URL of \test.htm OR a return code of 304. The order in which you specify filters does not affect the way they are processed.

The following graphic shows how Webtrends applies include and exclude filters. The web data files in the left column pass through an include filter and then an exclude filter. The final data set for analysis is shown in the right-hand column.

Creating Profile-Level Filters

You create profile-level filters using the Web Analysis options in Webtrends Administration. You can then apply them to individual profiles using the Advanced profile settings, or make them global to ensure they are applied to all the profiles you create. For example, you may find it useful to exclude all traffic from internal domains from your reports by creating an internal address filter and making it global.

When you create a profile-level filter, you can combine multiple filter criteria in a single filter that includes only data meeting all the criteria you specify. For example, suppose you want to isolate traffic from government domains on Thursdays. Creating a single hit filter that specifies address=.gov and day of week=Thursday allows you to include or exclude only traffic that both occurs on Thursday AND originates from the .gov domain. On the other hand, if you create two separate hit filters, one specifying .gov addresses and one specifying traffic occurring on Thursday, and apply both of them to the same profile, Webtrends includes or excludes all traffic that occurs on Thursday OR originates from the .gov domain.

The following table shows where to configure and apply profile-level filters in Webtrends Administration.

Creating Custom Report Filters

You create custom report filters using the Report Configuration options for custom reports, and apply them to individual custom report tables. Although custom report filters can be more complex than profile-level filters, because they use highly customized elements rather than pre-defined selections, they use similar logic. To combine multiple filter elements in an AND statement, you can combine multiple match criteria in a single filter. For example, if you want to track visits that both resulted in a purchase AND originated from the Google search engine, create a filter where the match criteria are Revenue=>0 and Visits resulting from specific search engines=Google and select Filter must match ALL criteria in the Filters dialog. If you want to track all the visits that resulted from Google searches whether or not they resulted in purchases, and all the visits with purchases, you can select Filter must match any one criteria, or specify each criterion in a separate filter and apply both filters to the custom report.

The following table shows where to configure and apply custom report filters in Webtrends Administration.

Understanding Filter Processing

Webtrends Analytics applies filters at different stages of analysis. When you apply a filter at the profile level, your selections affect the data available to all the report tables for the profile. When you apply a custom report filter, your selections affect only the data for a single report table. Information you filter out of that table is still available for other reports.

From within the Webtrends Analytics Reports, you can also apply queries or searches to the data within a report. Although queries and searches determine the data displayed in your report, and you can export a subset of data based on a query or search, Webtrends never discards any data based on queries and searches. Thus, they are not true filters.

The following graphic shows an overview of how Webtrends processes profile-level filters, custom report filters, queries, and searches.

Using URL Rebuilding

The goal of URL Rebuilding is only to track URLs as pages if they are truly unique pages. For example, many web sites add a unique user session ID or a time stamp to a page URL each time a visitor accesses it. Although these query parameters may be important for internal processing on your web site, they can make every URL appear to be a unique page, increasing your page count. Also, if you pass Webtrends query parameters in your URLs, you may want to ensure that they are not counted as part of the page URL.

URL Rebuilding affects the Pages reports and reports that are based on the Pages report. Effective URL Rebuilding definitions enable Webtrends to generate the Pages report and reports that are based on the Pages report using only the necessary hardware resources. By tracking only the relevant query parameters, you can minimize the disk space and memory required to generate these resource-intensive reports.

Understanding URL Rebuilding

When planning to use URL Rebuilding, consider which parameters are meaningful in reporting page views and which parameters are less meaningful or likely to result in a falsely unique URL. The following sample URL includes a session ID and several Webtrends query parameters, but it also includes parameters that provide valuable information about page content:

default.asp?type=domestic&div=news&article=104&WT.cg=local&WT.ti="Local%News"&sessionid=155428642

In this example, you would probably want to include the parameters that show which division and type of articles users requested. However, you would probably want to exclude the Webtrends query parameters (beginning with WT) and the session ID, because these parameters make each URL seem unique even when it requests the same content as other URLs. These extra parameters also make the URL longer and harder to understand when viewing page reports. If you created a URL Rebuilding definition that included only the div and type parameters, the Pages report would show the long URL in the previous example as follows:
default.asp?div=news&type=domestic

Configuring URL Rebuilding

In order to apply URL Rebuilding to your profile’s analysis, you need to create a URL Rebuilding definition in Webtrends Administration and enable it for profiles you want to use URL Rebuilding.

To create a URL rebuilding definition:

1. In the left pane, click Administration > Web Analysis > Options > URL Rebuilding.
2. Click New.
3. In the Name text box, specify a name to identify this URL rebuilding definition in Webtrends Administration.
4. Determine whether you want to rebuild URLs by specifying the query parameters to be included or excluded from the URL.
   • If you want Webtrends to analyze all query parameters for page view tracking except those you specify, leave Include all parameters except those specified in the exception list selected and click New to specify the query parameters that should not be analyzed. By default, Webtrends query parameters are excluded.
   • If you want Webtrends to analyze only the query parameters that you specify, select Exclude all parameters except those specified in the exception list and click New to specify the query parameters that should be included in page view analysis.
5. Click Save.
6. In the left pane, click Web Analysis > Reports & Profiles, mouse over a profile, and click Edit on the Action menu.
7. Click Advanced > URL Rebuilding.
8. Click the URL Rebuilding definition to apply to this profile.
9. If you want reports to include the unmodified URL rather than the rebuilt URL, select the Retain original URL check box.
10. Click Save.

Using URL Search and Replace

Each entry in the web activity file contains a field that references the URL of the page or file that was accessed. Webtrends Analytics uses this field to track activity on the web site and counts each unique URL as a separate page or file. However, some web server plug-ins or shopping cart software add a unique string to the page name in order to track visitors and their shopping carts. In these situations, every hit to such a page appears in the file as a unique URL (containing the ID of the visitor), which prevents Webtrends from counting of hits to these pages accurately.

URL search and replace allows you to delete specific identifiers from URLs or replace all unique identifiers with a common string. Doing so restores accuracy to page counts, makes your reports more informative, and avoids some out of memory conditions when the number of unique IDs is extremely high.

You can also use URL search and replace to analyze and report on URLs differently than they were captured in the web activity file. You might do this to report on all pages from a certain directory as hits to the same page, forcing the pages into a “group.” Grouping pages using URL search and replace differs from content groups because content group definitions only apply to content group reports, whereas a URL search and replace definition applies to all reports.

How URL Search and Replace Affects Analysis

Because URL replacement is performed before analysis, your URL search and replace definitions affect the entire analysis for the profiles in which URL search and replace is used. Unlike URL Rebuilding, URL search and replace replaces any kind of string in the URL, not just parameters, and affects all your reports for the profile, not just reports on pages. Because it affects all reports, make sure you use this feature carefully. Do not replace data in a URL if you want to use that data in any Webtrends Analytics reports.

Configuring URL Search and Replace

To apply URL search and replace to your profile’s analysis, you need to create URL Search and Replace definitions in Webtrends Administration and enable them for global use or configure certain profiles to use them.

To create URL Search and Replace definitions:

1. In the left pane, click Administration > Web Analysis > Report Configuration > URL Search/Replace.

2. In the Name text box, specify a name to identify this URL search and replace definition in Webtrends Administration.

3. Select a replacement method from the Replace from and Up to lists.

4. If you want Webtrends to replace to the end of the query string if the text you specify is not found, select the Replace to End of String if pattern is not found check box.

5. In the Up to text box, specify the text that identifies the string to replace.

6. In the With text box, specify the text that you want Webtrends to use when the string you specified is found.

7. If you want Webtrends to replace the string only when certain conditions are true, select the Perform replacement only if the URL contains check box and specify the content that must be included in the URL for the replacement to be made.

8. In the Test Area text box, type a URL string that will test your definition and click Test. Webtrends shows the results of your definition in the Resulting URL and Text that was replaced text boxes.

9. If you want to use this URL search and replace definition in all profiles, select the Global: Include in all profiles check box.

10. Click Save.

11. If you did not select Global: Include in all profiles to apply this definition to all profiles, edit your profile and enable your URL search and replace definition to include it in reports.

    1. In the left pane click Administration > Web Analysis > Reports & Profiles.
    2. Mouse over a profile and click Edit on the Action menu.
    3. Click Advanced > URL Search And Replace.
    4. Select the URL search and replace definition to apply to the selected profile.
    5. Click Save.

Tracking Page Views and Downloads

As Web 2.0 and other advanced web technologies have become standard in page design, the idea of what should be considered a “page view” or a “download” has grown more complex. When there are many interactive elements on a single page, customers may now engage with a web site for long periods of time without ever loading a new HTM, PHP, or ASP page.

While page views and downloads can still measure engagement with your web site, you may want to tailor these measures to reflect specific elements of your page design. For example, if you use mouse-over animations on your home page, you may decide to treat hits to these animations as page views in their own right. On the other hand, you may decide to exclude them from page view reporting altogether and measure this activity using a different method. How you choose to identify page views and downloads depends on your site design and your web analytics goals.

Webtrends provides several methods of page view and download determination, including the simple method of classifying hits to specific file types as page views or downloads. However, we recommend using the wt.dl parameter to identify page views and downloads. For more information, see “Why Use wt.dl?” below.

You can access the Page View and Download definitions for each profile by editing the profile and selecting the Analysis settings. You can access the global Page View and Download definitions by clicking Administration > Web Analysis > Options > Analysis.

Why Use wt.dl?

The wt.dl parameter is the standard Webtrends parameter used to identify different types of web activity, including clicks to web forms, dynamic links, and other types of on-page clicks, by unique numeric values. If you use the Webtrends tag with Webtrends On Demand or SDC, you can use wt.dl as part of a larger strategy for precisely identifying and tracking your site activity. Using wt.dl simplifies Webtrends configuration and allows for more comprehensive reporting on page interactions.

For information about implementing wt.dl, see “Implementing wt.dl,” below.

As shown in the following table, wt.dl typically uses the following scheme for identifying page views and downloads:

For a complete list of supported wt.dl events and event IDs, see the "Query Parameter Reference."

Implementing wt.dl

If you use Webtrends On Demand or the SmartSource Data Collector, the wt.dl parameter can be passed to the data collection server using a META tag or ActionScript. However, the recommended method for enabling wt.dl tracking is to implement Event Tracking using the Webtrends Tag Builder at tagbuilder.webtrends.com. wt.dl can be passed to dcsMultiTrack.

If you did not create your tags using the Tag Builder, see “”.

Understanding URL Truncation

If you define your page views or downloads using file extensions, you can choose to shorten the URLs in page view and download reports by removing all query parameters from the URL string.

When you click File extensions as your identification method in the Download Definitions or Page View Definitions dialog, you specify the file types in a space-delimited list. To truncate URLs for each file type, include the file extension followed by a comma and a 0, and enclose the expression in curly brackets. For example, to specify that URLs should be truncated for hits to PDF and Excel files, type:

{pdf,0} {xlsx,0}

File types that should not be truncated can be entered simply by typing the extension. You can choose to truncate URLs for some file types and not others, as shown in the following example:

pdf xlsx {gif,0}

In this example, the query parameters in URLs requesting PDF and Excel files are not truncated, but URLs requesting GIF files will be truncated.