Creating a Webtrends report to see daily page views over a month for a specific page

Webtrends trend reports break down monthly data by daya, but you can’t see the results for a specific page without creating a custom report.

Having seen this request twice in a couple of months I thought I would share how to do it!

Unfortunately it is necessary to create a custom report as the trend reports available don’t break the metrics down by page.

Create the report with:

Dimensions: Pages, Time Period
Measures: Page Views, Visits, View Time, avg. View Time (calculation)

It isn’t perfect but at least you get the information.

Webtrends report Pages by Day

Problems are:

  • because the dimension is “Time period”, you have to select a specific month to get the days, (selecting e.g. 2015 gives you the data by month instead)
  • the order of the time period is not by day, but by the number of page views (or visits etc.) so you have to extract the report and reorder it
  • depending on the number of pages you have, you may exceed the report limits.  If this happens you would need to do one of the following:
    • increase the table limits
    • segment the site areas into different profiles (my client had already done that for other reasons)
    • add a Filter to the report for specific pages (if you are only interested in some of them)

P.S. If you can find a way to improve this, please share!

How to delete unused Webtrends reports!

1.    Login to Webtrends V9 with the Administrator role
2.    Click on Administration

Webtrends Administration

3.    Select Web Analytics > Report Configuration > Custom Reports > Reports

Webtrends reports

4.    Use the Search facility to find the report you wish to delete:

Webtrends Custom Reports

5.    Hover the mouse to the right of the name and pick Show Uses from the Dropdown Menu:

Show uses in Webtrends

6.    This will show you the list of profiles that use that report (or tell you that it is not used):

Show Uses of Webtrends Reports

7.    If you are still sure that you want to delete the report, you need to remove it from each profile in the above list first.  Click on Reports and Profiles and Edit each one.
8.    Select Reports from the Advanced tab, scroll down to the relevant report and uncheck the box.  Save the profile and repeat for all the others.
9.    When the report is no longer used in any profiles, or if it was not in use to start with, it can be deleted.  Go back to the list of reports and choose Delete from the Dropdown menu (you will be told if you still need to delete it from a profile):

Delete Webtrends report.png

10.    Confirm that you want to delete the report and add a Change Comment.
11.    That’s it!

How to insert the Webtrends tag into Drupal sites

Following a Twitter conversation on how to implement the Webtrends tag into Drupal sites, I contacted their HelpLine and received the following (thanks Jon Bijak!), which I thought worth sharing as I could find nothing in their knowledge base when searching for Drupal: Products Webtrends On Demand Webtrends Analytics 9.2 Webtrends Analytics 8.7 Webtrends Analytics 8.5 Webtrends Analytics 8.1 Webtrends Analytics 8.0 Drupal is a widely used theme-based content management system. While Drupal has the capability of putting footer code on every page, blocks are filtered out of the footer. While Drupal has contributed modules that may provide the ability to do this, Webtrends does not manage or support them, and therefore the Webtrends tag should be added to the theme itself (the HTML section of a PHP file). Since it is possible to apply a separate theme to the Administration section, this gives the customer the added bonus of not having to track the administrator. Keep in mind that if you use multiple themes, you will need to apply this to each. Since it is possible to apply a separate theme to the Administration section, this gives the customer the added bonus of not having to track the administrator. 1. Go to http://tagbuilder.webtrends.com to build the tag using V9. After selecting “Build Tag,” in the “Javascript Folder Name:” field enter the path “/sites/all/libraries/webtrends/”. 3. Save the tag, then extract its contents. 4. Copy the webtrends.js file to the “/sites/all/libraries/webtrends/” of your site, creating it if it doesn’t already exist. 5. Use a text editor to open the main template file for each theme (“page.tpl.php” by default) and webtrends.html (extracted from the Tag Builder archive). 6. Copy the contents of webtrends.html and paste it above the tag in the theme’s template file. 7. Save the template and close both files. 8. Repeat steps 6 and 7 for each theme. The theme should now be functioning and you’ll want to test it as you would any other site. More Information  There is currently a Drupal module for Webtrends. Using this module hasn’t been researched and this module is not supported or affiliated with Webtrends. Unfortunately, there does not appear to be anything that will handle commerce tagging. Any exploration into configuring Drupal, or any other CMS (beyond basic tagging) is strictly a Professional Services engagement.  N.B.  There is an error in the renumbering, but I don’t think it is important… Please let us know how you get on!

Webtrends IP-less Cookie

Data Concerns

With electronic surveillance projects like PRISM catapulting data intrusion and collection issues into the headlines, privacy is becoming more and more high profile in today’s Big Data era. It’™s a concern for consumers and providers of digital media alike. Consumers don’™t like the idea of their browsing history being available to Big Brother, yet this same information is key for website-owning organisations to monitor and improve their performance. Organisations need to balance the need for data with the need to have the trust of their customers.

Recent EU legislation attempted to resolve this issue by forcing websites to obtain consent from visitors before serving cookies, but studies have shown that only a tiny minority of people (less than 2%) actually block cookies. So what’s going on? Are people really concerned by cookies? I think the answer is a qualified yes, despite the lack of action actually taken by the consumer. A survey by eConsultancy showed that 89% of those surveyed thought the EU cookie legislation was a positive step and that only 23% of people would be happy to accept cookies. So, people are clearly concerned about cookies and a potential invasion of privacy but have little understanding of what a cookie is and how it is used. Cookies essentially have an image problem!

Cookie Make Over

It is left to the analytics providers to take a lead in addressing public concerns in order to protect their customers’™ (and their own) business interest. A cookie make-over is needed to increase trust and preserve loyalty in the face of fast changing public opinion. And that’s just what Webtrends has done by introducing their new IP-less cookie.

A step in the right direction

Traditionally Webtrends cookies have been constructed using the visitors IP address to generate a random value. In order to remove the IP address, Webtrends is phasing in the new cookie over the coming months. New visitors will receive the new IP-less cookie, while returning visitors will have the expiry date of their cookie (containing their IP address) set to 1st January 2014. Webtrends estimates that, as many visitors clear their cookies every month, by this date only 5% of visitors will have the old style cookie left on their machine. You don’™t need to do anything to make this happen, the process will be automatic.

However, an option is provided to remove the final 5% for organisations that want to totally eliminate IP storing cookies. This requires adding a JavaScript plugin to the new v10.4 tag which forces the overwriting of the old cookie value with the new. However, this is a value judgement each organisations has to weigh up independently. None-the-less this is a step in the right direction towards reassuring web visitors that analytics need not be overly intrusive and that, hopefully, will preserve the ability for organisations to collect and analyse the data they need to make essential business decisions.

For more information read Webtrends Knowledge Base Article: Improved Data Collection

Webtrends Visitor Paradox

How can I have more visitors in a week than a month?

Recently we had a client who pulled up two visitor reports. One for the month of December and one for a custom date range, 1st – 25th December. He was startled to find that there had been 25 visitors between 1st – 25th yet only 13 in the whole of December. How could this possibly be true? It simply did not make sense.

To understand what had happened we need to first understand how the Webtrends Analytics package works out how many visitors visit your site in a given time period.

Webtrends will typically have buckets that it lumps visitors into.

If I visit the website today, I will be put into each of the Daily (Monday), Weekly & Monthly buckets.

WebtrendsVisitorBuckets0.1.png

If I visit tomorrow (and it’s the same week and month), I am only put in the daily bucket for Tuesday as I’m already in the Weekly & Monthly ones.

WebtrendsVisitorBucketsDaily0.2.png

If I visit next week (and it’s the same month), I will be put in the daily and weekly bucket as I’m still in the Monthly one.

Webtrends Visitor Buckets Weekly0.3.png

So, if I want to view a report that shows how many visitors there were for last Tuesday, Webtrends simply looks in the Day bucket for last Tuesday. Similarly if I want to see last week or January it can look in the relevant bucket.

But what happens if I want to view a custom date range? Webtrends does not have buckets for this, so it improvises and creates an approximation by summing visitors from the buckets it does have.

Let’s look at an example. We want the custom date range 1st December to 10th December. Webtrends will use a combination of the day buckets and week buckets to give us an approximation.

Webtrends Visitor Buckets Custom Date0.4.png

The problem is that visitors may be in more than one bucket and so could be double counted.

So in our example below, one guy visits on the 1st and then again during Week 1 whereas the rest of the gang only appear in one bucket. If we count up the visitors there are clearly 7 of them. But if we count them by bucket there are 8!

Webtrends Visitor Buckets Custom Date Calculations0.5.png

               1                        4                         2                       1

And if we look at the month bucket we see there are only 7 visitors.

Webtrends Visitor Buckets Custom Date Monthly Calculations0.6.png

This is the reason that sometimes, if you look at a custom date range, you will see more visitors in that shorter period than you do in a longer default time period. Because of this double counting you should therefore always be cautious when using custom date ranges and avoid them where possible.

Note that this calculation applies only to visitors. Visits & Page Views can be added together from their different buckets as they are not mutually exclusive between dates!

Webtrends Roadmap for 2014

Yesterday I was at the Webtrends User Conference #Engage13, during which two exciting announcements were made in the keynote speeches that indicate the direction for Webtrends over the coming year:

Webtrends Explore

This has been a front end to the Data Warehouse (or Visitor Data Mart as it was renamed) for some time and allows users to slice and dice the granular level.  Effectively it is like having the functionality of a massive Excel pivot table on a database containing every hit.

So, nothing new there, you say!  But the big announcement is that this functionality is coming to Analytics next year.  No longer will we have those table size limits restricting the number of Pages that we can see or having users complain that they can’t see their new Page because it keeps being deleted off the end of the table.

All data restrictions are going away to create “Webtrends Unlimited”!  What is more, Bruce Kenny (EVP of Products, Webtrends) said that this would be free for Webtrends OnDemand Analytics users.  I look forward to the presentations appearing online to check that I heard that one right as it could have massive implications for customers with large number of profiles (that may not be required any longer?) or the VDM where it is used simply as a staging database to get the data captured by Webtrends into their Enterprise Data Warehouse.

One of the key advantages of Webtrends over Google Analytics has been the ability to capture many different metrics and create a wide variety of custom reports based on them.  This advantage has been eroded significantly by Universal Analytics which will become ubiquitous (see more at http://elignum.co.uk/google-analytics-is-dead-all-hail-universal-analytics), however the power of Explore will take Webtrends to a new level in terms of being able to get exactly the data you want at any level of granularity without having to build new custom reports and reanalyse data.

Webtrends Lifetime Streams

At last year’s Engage conference, I sat and watched the demonstration of Streams and thought “what a beautiful visualisation of real-time data – there is a solution looking for a problem”.

However, this year, it became clear that there is at least one scenario where Streams fits and that is where basket abandoners can be emailed immediately, where “immediately” can be not just after the usual 30 minutes of inactivity defining the end of the session, but could be, say, 15 minutes after they look at a specific page, significantly improving conversion.

This happens through the Action Centre (https://help.webtrends.com/legacy/en/actioncenter/) which now has a Universal link in addition to the pre-integrated ones with specific email service providers, this means that you can now take the data and feed it anywhere you want in real-time.

Finally, the announcement of Lifetime Streams means that additional information on the Visitor such as a Webtrends Segment or Lifetime Value can be added to the event data and be available within those same milliseconds as before, enabling additional filtering to take place before deciding what action to take.

All in all, I left feeling that Webtrends have moved forwards again significantly in the Analytics Arena – excelling at collecting the data and making it available to their customers to use in whatever way suits them best…

More numbers, same cost – Webtrends OnDemand includes more of those “hidden extras” for free in 2014

Webtrends have just issued their new price list for 2014 and there is welcome news for those customers who have been constrained in their analysis by the cost of add-ons. We were greatly encouraged in the recent release (V10.7) with the ability to reanalyse the data back for 3 months at no cost (not to mention how easy it is to do!).  This has always been a thorn in the side for both development, waiting to see data in that report to test it, and the users, who often request a new report because they are interested in finding out what caused an anomaly in the recent past. The passing on to their customers of the presumed decrease in resource costs is continuing, and Webtrends have announced that OnDemand will include 100 profiles rather than the current  5 in 2014 (current customers will have to wait until renewal, although additional packs can still be purchased beforehand).  This will enable many customers to segment their data far more than they do currently and set up profiles to get all their reports by e.g. area of the site, so that each of the Webtrends Report Viewers will be able to see their reports based on just their own data rather than having to make educated guesses on e.g. the number of Visitors to their area of responsibility. Other announcements include unlimited Custom Reports – also welcome since the arrival of all the V10 ones. So, as we move into December and start preparing for next year, think about what you can use all those extra profiles for next year…!

Set of exploded 3d numbers

Clickthroughs, Visits and Page Views in Webtrends

These 3 metrics can provide what appear to be irrational numbers, so this article explains how they are derived.

First some definitions:

To tell Webtrends that the Visitor has come from a campaign, you will have added the WT.mc_id parameter to the link in your email, tweet or banner ad (or put it in the meta data of a landing page for off-line campaigns). As you would expect, clickthroughs are simply the number of times that someone clicks on the link (i.e. the number of times that Webtrends sees a hit with the WT.mc_id parameter present).

Conceptually, a Visit is when someone looks at your site, views a few pages and then goes away again. The difficulty is that we can’t know when they stop looking! In line with general practice, Webtrends assumes that the Visit ends after 30 minutes of inactivity. So that means that if someone goes to lunch in the middle of looking at your site and comes back after an hour to resume, that will count as a second Visit.

A Page View is the easy one! Someone looks at a Page, that is a Page View!

However, looking at all these together can result in some apparently strange results:

Page Views higher than Visits

If someone looks at the Home Page, the Products page and then goes back to the Home Page, then you would see 3 Page Views but only 1 Visit in your your Campaign report. In the Pages report for the Home Page you would see 2 Page Views and 1 Visit because that person has looked at the same page twice within a single visit.

Clickthroughs higher than Visits

When someone clicks through from an email, looks at a page on the site, returns to the email and clicks through on either the same or another link, that will be 2 clickthroughs. If that behaviour is seen by Webtrends as all being a part of the same Visit (i.e. without 30 minutes of inactivity at any point between the 2 clickthroughs), then there will be 2 Clickthroughs, but only 1 Visit.

Visits higher than Clickthroughs

A number of Campaign reports attribute future Visits to the last (or “Most Recent”) Campaign. So, if someone clicks through from an email on Monday, bookmarks the site and returns via the bookmark on Tuesday, then these reports would attribute the second visit to the campaign as well, even though the Visitor had come to the site directly. So, the metrics for the week would appear as 1 Clickthrough and 2 Visits.