Articles in this section

See more

Implementation 4 – Review of Data Collection

Avatar
Team Training
  • Updated
Follow

This training chapter shows ways for reviewing the data transmitted to Webtrekk and how the integration should be tested.
You can download this training chapter as PDF file here.

Content:

  1. Onsite-Test
    1.1 Tools
    1.2 Status Codes
    1.3 Parameter Overview
    1.4 Requests
    1.5 Checking Tag Integration
  2. Testing
    2.1 Crawlers
    2.2 Checking Requests
    2.3 Basic Tests in Webtrekk
    2.4 Extended Tests in Webtrekk
  3. Data comparison

1. Onsite Tests

With onsite tests you can directly check on your website which data is passed to Webtrekk.

1.1 Tools

Checking the source code of a page does often not suffice to know which data is passed to Webtrekk.
There are various tools available to analyze the data traffic. Webtrekk recommends the following solutions:

  • Browser extensions:
    • Firefox Web Developer Tools
    • Chrome Developer Tools
    • Webtrekk Tracker
  • Browser-independent Software:
    • Fiddler

 

1.1.1 Firefox Developer Tools

Firefox provides the possibility to examine requests with the help of an internal tool.

  • Open Firefox and Click on „ > Web Developer > Network“ or use the shortcut Ctrl+Shift+E.
  • Open the website you want to analyze and filter the data traffic on your individual track domain
  • Choose a request. The parameters passed to Webtrekk are shown at „Params“.



1.1.2 Chrome Developer Tools

Google Chrome provides the possibility to examine requests with the help of an internal tool.

  • Open Google Chrome and press „F12“ to open the Developer Tools. Choose the tab „Network“.
  • Open the website you want to analyze and filter on your track domain.
  • Choose a request.
    The parameters passed to Webtrekk are shown at „Headers > Query String Parameters“.



1.1.3 Webtrekk Tracker

The Pixel Check Plugin developed by the Webtrekk Community is available as Chrome extension and as Firefox Add-On.

  • It directly shows the kind of request (incl. separately showing „webtrekk ignore“).
  • The plugin is available in the Chrome Web Store and as Firefox Add-on.
  • Please note that custom track domains have to be entered at „Go to options“.

 

1.1.4 Fiddler

Fiddler is an HTTP debugging proxy server application. It allows for browser independent testing.

  • When inspecting https connections, „Decrypt HTTPS traffic“ has to be activated in the options.

1.2 Status Codes

The „Status“ or „Result“ codes provide information whether the data was transmitted correctly.

Example statuses:

  • „200“ = successfully tracked
  • „Cache“ = successfully tracked
  • „404“ = data transmission was not successful. TrackID or track domain do not exist.
  • „(Aborted)“ = in most cases the request was measured.
    • Usually occurs with heatmap or event requests.
      Technical background: The browser establishes a connection to the server (track request) in order to load the picture requested from there. As soon as the query reaches the server, the request is measured. But the server then has to return the picture, which in part is then no longer possible due to leaving the page.

1.3 Parameter Overview

There are several parameters that pass different information.

A description of all parameters can be found in the here.

  • Example: Information passed in the parameter „p“


1.4 Requests

Webtrekk tracks the following requests:

  • Page Requests
  • Event Requests
  • Form Requests
  • Media Requests
  • Heatmap Requests

1.4.1 Page Request

In a page request, information on a page view is passed. Thus, a page impression is counted.
With further parameters, other information can be passed.

The following information is tracked in the example:

 

1.4.2 Event Request

In an event request, information on an page view is passed; thus, a click is counted.
With further parameters, other information can be passed.

  • Event requests can be identified by the parameter “ct” (Click Tracking).


  • The following information is tracked in the example:
  • The event “themen/bei-tisch-202.htm”.
    The event is automatically assigned to the page that was tracked last within the visit (i.e., not necessarily to the page “themen” that you see in the request).

 

1.4.3 Form Request

In a form request, information on an form and the included form fields is passed;
thus, “Qty Form Started” is counted.

  • Form requests can be identified by the parameter „fn“.

The following information is tracked in the example:

  • The form “manieren_per_mausklick” with submit status “0” (not send).
  • The form field “catchword.text” with form field value “empty”.

 

1.4.4 Media Request

In a media request, information on a medium is passed. Thus, a media view is counted.

  • Media requests can be identified by the parameter „mi“.

The following information is tracked in the example:

  • The medium “arbeiten_in_reports”.
  • The medium is paused at second 40 and has a total duration of 1159 seconds.

1.4.5 Heatmap Request

In a heatmap request coordinates of a single click are tracked.
This data can be visualized in the heatmap representation (only available in Webtrekk Q3).

  • Heatmap requests can be identified by the parameter „hm_ts“.

The following information is tracked in the example:

  • The coordinates of the clicks.
    The click is automatically assigned to the page that was tracked last in Webtrekk.

Webtrekk needs a minimum number of clicks until the heatmap can be depicted in an aggregated view. Alternatively, you can choose the view type „detailed“ in the overlay center.

1.5 Checking Tag Integration

Is Webtrekk Tag Integration loaded?

  • The check is done based on the request.
  • When using the default setup filter on „responder.wt-safetag.com“.
    Alternatively, use the Tag Integration ID.
  • Status Codes
    • 200
      If the script is not in cache (e.g., at the first request of a user or after again publishing a container): Succesfully loaded and container is delivered.
      If the script is in cache: Account ID not found.
    • 204
      No container was found for the current domain.
    • 304
      Successfully loaded and container is delivered (only from the second request on).

2. Testing

In this chapter we describe how it can be checked whether tracking data is coming in correctly. This way, you can check whether the implementation concept is realized properly.

2.1 Crawlers

Is the Webtrekk tracking pixel integrated on all websites?

  • Automated audit using a crawler is advisable to check for pixel integration.
    E.g. Screaming Frog SEO Spider (subject to charge), Tagalyzer

2.2 Checking Requests

Use one of the tools presented in section 1.1 and check whether the corresponding requests are created when, e.g., a page is viewed or a link is clicked.

2.3 Basic Tests in Webtrekk

The following tests are performed directly in Webtrekk Analytics.

Generally, there’s a 60 minutes update interval of the system. The exact point of time of the last data update can be found in the lower right of the analysis window.

When using Live analytics, you can see the values of the chosen dimensions and metrics a few seconds after opening the analysis in Webtrekk.

If the test is performed in the live environment and several persons are using the test environment, you have to filter on your test data out of all data that was tracked. To do this, you have several options:

Restrict on:

  • End device Visitor Ids
  • Order ID
  • Campaign



  • End device Visitor Ids: Filtering using the End device Visitor Id. It can be read out by the parameter „eid“ (only available if first-party cookies are used).


    The analysis is not restricted to a single visit, but to a visitor. Use the calendar to restrict your analysis on test sessions of a single day.

    Example:  Filtering on auf eine End device Visitor Id


  • Order ID: Filtering on the basis of an order ID generated in the test.

    Using the scope in the filter engine, you can define whether you want to restrict your analysis on a single page, on a visit or on a visitor.

    Example:  Filter with an order ID.

  • Campaign: Filtering on the basis of a campaign used in the test (e.g. media code).

    The campaign can be configured in Webtrekk.
    Using the scope in the filter engine, you can define whether you want to restrict your analysis on a single page, on a visit or on a visitor.

    Example:  Filtering an advertising media “wt_mc = newsletter.test”


Are Visits and Page Impressions tracked correctly?

  • View some pages and check whether the number of pages that you viewed is the same as the number of Page Impressions.

    Analysis:  Visitors > Traffic

    If there are more than 30 minutes between two page views, another browser is used or differing login information is sent, a new visit is generated.


Are pages tracked correctly?

  • Is the number of viewed pages the same as the number of page impressions?
  • Is the naming of the pages the same as defined in the implementation concept?

    Analysis:  Navigation > Pages > Pages

Are events tracked correctly?

  • Click the links on your website and check whether the number of clicks is the same as shown in Webtrekk.
  • Is the naming of the pages the same as defined in the implementation concept?

    Analysis:  Navigation > Events

    Events with the name „webtrekk_ignore“ are not shown in the analysis.

 

2.4 Extended Tests in Webtrekk

The following analyses are suitable for in-depth data review: 

Entry pages

  • For which pages are entries measured?

    Possible causes for error: session interruption
    When using First-Party cookies different domains lead to several visits and accordingly several entry pages.
    Please note: After a Session timeout (no action for longer than 30 minutes) the next page is also counted as an entry page.

Path analysis

  • Are pages etc. called in the expected order?

    Sources of error: missing or multiple measurement

Browser Tests

  • Does the tracking work (e.g. pages, actions, orders) in all browsers?

    Source of Error: Tracking does not work in particular browsers

Events per page

  • Are events measured on the correct page?

    Source of Error:

    • Measurement of the events does not occur on all pages.

    • Events are assigned to the wrong page (Multitab browsing is not activated)

Parameters per dimension

  • On which pages are page parameters tracked?
  • During what actions are event parameters tracked?
  • For what products/orders are e-commerce parameters tracked?
  • For what sessions/visits are session parameters tracked?
  • For what advertising media are campaign parameters tracked?

    Example: On what pages is the page parameter "Error Message" recorded?

    Source of Error: Parameters are not set for all defined objects

Category per dimension

  • To what pages are content groups assigned?
  • To what products are product categories assigned?
  • To what media are media categories assigned?
  • To what Customer-IDs are URM categories assigned?
  • To what advertising media are campaign categories assigned?
  • To what times are time categories assigned?


    All dimensions not assigned to a category appear under "-".

    Source of Error: No assignment

  • Categories (e.g. content groups) are only assigned once, on the first track request (e.g. page).
    If the category was not configured in Webtrekk during the first request or not transferred in pixels, it is shown as "-" (not assigned).To assign these objects, you can carry out an import (Excel, SOAP interface…) or activate a renewed categorization (reset function). The latter function can be activated optionally by your supervising consultant.

    Function for renewed reading of categories ("Webtrekk Q3 > Configuration > Categories > Import-Export")

 

3. Data comparison

Data Comparison is another important building block to check validity of the data.

Web analysis will never measure 100 percent of all accesses. A key factor for a meaningful analysis is that deviations remain consistent.

Reasons for missing data in Webtrekk:

  • Page not completely loaded or interrupted before the pixel was sent.
    Occurs frequently in the case of large websites or slow Internet connections (e.g. mobile access).
  • IP address or domain excluded from measurement in system configuration.
  • Tracking blocker active and no custom track domain in use.
  • User has excluded himself from measurement by opt-out.
  • Access occurs without activated JavaScript and no NoScript is integrated.
  • Technical Errors
    • No pixel integrated, wrong track domain / TrackID, JS file not linked correctly, missing function call, …

You could compare the data from Webtrekk with internal data, other Web Analytics tools or Marketing tools.

  • Comparison with internal data (e.g. DWH/CRM)
    • number of online orders VS. orders in Webtrekk
    • newsletter registrations according to newsletter tool VS. newsletter registrations in Webtrekk
    • recorded logins VS. Customer-IDs in Webtrekk

Hint: Import key figures (e.g., number of online orders) on a daily basis and use the alert function in Webtrekk in order to have insights into data validity at all times.

  • Comparison with other web analytics tools
    In general, other web analysis systems should provide similar data. Deviations of up to 5 percent are to be regarded as normal.
    For an initial comparison, an individual page should be considered for a limited period and the metric page impressions should be used.

    Example: Comparison of the page impressions of the homepage for the last seven days on a daily basis.
    • Are the deviations between the tools per day consistent?
    • In the next step the comparison can also be extended to other pages/content groups.

  • Comparison with marketing tools
    Two methods that measurw at different places of the site using different techniques will, in general, never show the exact same numbers. It is important that the deviation is consistent and within reasonable limits.

    Usually different tools measure at different “spots”. While – for example – your marketing tool measures clicks on an advertisement, Webtrekk measures users who viewed or loaded the landing page completely (until the pixel was sent). Users who click back and forth very fast cannot be tracked by Webtrekk.

    If the marketing tool aggregates clicks done quickly after each other, use the metric “Visits” or alternatively “Campaign Clicks” in Webtrekk.

     

 

 

 

Related articles