Troubleshooting Tools for Web Analytics

I recently read an article by a friend of mine Neil Mason, called – Tackling the basics of web analytics: Getting the right numbers right . To summarize, Neil discusses how difficult it can be to install even the most simplest of tracking tags (data collector beacons) across an enterprise web site. That is, a site that is large (thousands if not hundreds of thousands of pages of content), uses multiple technologies and has multiple stakeholders – often in different countries and sometimes different companies.

However, this fundamental step of getting the data in using page tags, is the key to everything else i.e. getting good, solid, accurate data in*. There simply is no point investing in analysis if the data is flawed. After all, garbage in = garbage out. And of course the web is ever changing, so maintaining data integrity is also key. Page tagging is therefore not a one time, “set it and forget it” process. It requires careful deployment planning and regular maintenance checks to ensure data holes are not appearing.

*I discuss this specifically for Google Analytics in a series of articles .

Six tools to help you troubleshoot web analytics

As page tag deployment is so important, I wanted to share which tools I use for troubleshooting implementations on a day to day basis. These are all free(ish) to use and are not specific to Google Analytics – they can help you sanity check any web analytics implementation that uses page tags.

  • WASP

    Web Analytics Solution Profile (WASP) is a Firefox add-on developed by Stéphane Hamel based in Québec. I highly recommend this to anyone that needs to quickly check what web analytics tracking code is present on their pages. When I first tried this out it was able to track 20+ vendor page tags. Now its up to 121 and includes all the web analytics tools I ever knew of, plus Ad networks, behavioural targeting solutions and multivariate testing tools.WASP works by executing the page tag detecting the cookies your browser sets when you view a web page. It then matches this to a known list of providers. The downside was that it could only be done on a per page basis. However, now even this have been overcome with the latest release including a site crawl that automates the process by following links within your web site. WASP is free for scanning 10 web sites upto 100 pages each (free to look at any web site on a per page basis).

  • Web Developer toolkit

    This is a firefox add-on that adds a menu bar to your browser with a whole range of useful features for anyone that has an interest in creating web pages. It has an excellent browser error console and DOM inspector, as well as quick lookup tools for cookies, source code etc. With over 8 million downloads, it is one of the most popular add-ons ever and is free.

  • SiteScan GA

    This is a Google Analytics service developed by EpikOne from the US. It is used for checking the existence of the Google Analytics Tracking Code (GATC) on your pages. The methodology is to first establish what web site pages you have in the Google Index (via a Google Search API call). Then it retrieves these URLs and performs a pattern match on the HTML source code looking for the GATC “signature” code. Its a very neat way of avoiding the overhead of crawling your web site (fraught with traps that can result in an infinite loop that can bring servers down!) and it is free for 5 web sites, upto 1,000 pages.The slight downside is that if your pages are not in the Google Index, then it cannot scan your pages. This includes e-commerce checkout systems.

  • Firebug

    This is a free firefox add-on that adds debug capabilities for JavaScript, CSS and HTML live in your browser.

  • WebBug

    A windows application that allows you to enter a URL and see exactly what is sent to the web server and what response is sent back. This is the information that your browser takes care of when rendering a page. I mainly use this to check a web server’s status code response. It is very useful for tracking redirection issues – a common problem that can result in the redirection server loosing campaign variables required for your web analytics tool. WebBug is free to use.

  • Regex Coach

    If you are involved with the implementation of web analytics tools then you will often need to create regular expressions. These are snippets of pseudo code that match (i.e. find) patterns in text. They are used mainly to allow you to create filters for data cleansing and segmentation purposes. The problem is they look like algrebra, which means they are very prone to errors. Essentially, I never implement a regex without first running it through this excellent Windows application. Regex Coach is free to use though a donation always goes down well…

Also worth mentioning is Maxamine – a paid for enterprise class tool that is now part of the behemoth consulting firm Accenture. This works in a similar way to WASP, that is auditing web pages. However, as a web diagnostic tool it is capable of going way beyond looking for page tags. Maxamine is used to identify web performance issues such as site compliancy and design consistency issues etc. Maxamine is a complex and sophisticated tool with a price tag to match! However, if you are the type of organisation that uses services from companies such as Accenture, this can be a very powerful addition to your armory.

How often should I check my implementation?

The main factor to consider here, is how often is your content changing? If 10% of your web site content changes each month, then by half way through the year the majority of your web site will have changed. The greater the change the higher the possibility of errors. Even non-humans such as CMS, CRM systems and web servers can, and do, make errors. And because page tags are a hidden piece of code, errors are not visible by simply visiting the page in your browser. The result is that page tag errors easily go unnoticed and build up rapidly on your web site.

As a rule of thumb, I suggest you sanity check your page tags at least on a quarterly basis. If you find significant errors with your first audit, move to a monthly or even weekly maintenance frequency until you solve the problem, then revert back to quarterly.

Do you have implementation anecdotes or horror stories? Please share these and any other tools you have found useful for web analytics implementation troubleshooting.

Looking for a keynote speaker, or wish to hire Brian…?

If you are an organisation wishing to hire me and my team, please view the Contact page. I am based in Sweden and advise organisations in Europe as well as North America.

You May Also Like…

Sayonara Universal Analytics

Sayonara Universal Analytics

My first Google Analytics data point was 15th May 2005 for UA-20024. If you are of a certain age, that may sound off...

8 Comments

  1. Kevin

    Not sure if I have a horror story on tracking code being missing, but I did have a situation where while upgrading to a new website.

    Our developers had done something screwy to rediret all 404’s to the home page, a practice I’m against, but hey, I’m just the analyst.

    In the same fashion, the script they were using to display SEO friendly URLs was causing the Google Analytics code to not recognize Organic vs Paid traffic.

    Long story short, I have 3 weeks of data from March that shows all my traffic was Organic, even though 50% of it was paid.

    Wonder if one of the tools you mentioned would have helped us solve that faster.

    Reply
    • Brian Clifton

      Kevin: If its any consolation, that type of error is very common. WebBug is a tool that could help you here. However, the simplest way is to test a random selection of your PPC ads on a regular basis. Essentially you click through and check the landing page URL to ensure the campaign tracking parameters are present. For Adwords this is of the form gclid=ASDDNCJDR123. For other networks it will be utm_source=ysm&utm_medium=ppc&utm_campaign=xmas for example.

      If these are not present in the landing page URL when you click through, then GA will not recognise visits as paid search – treating them as organic, as you know. Note, that it is possible for your web team to reassign the utm_name variables to something else (though not common). HTH.

      Reply
  2. Greg Tyree

    Charles is not Mac only, it is available for Windows as well. Also there is no mention of Fiddler, a tool similar to Charles and available on Windows as well.

    Both of these are pretty rich tools but with configuration can be used to track and verify metrics calls.

    Reply
  3. Big J

    Very cool tools, I just had sitescan crawl my site and I’m looking forward to getting the results. Many thanks!

    Reply
  4. Stephan Winter

    Nice list! In addition to most of the tools you mention, one of our Flex guys tipped me about Charles (http://www.charlesproxy.com/) for those hard to catch requests, like document downloads and other OnClicks where Firebug looses context. It´s Mac only, if you´re on Windows the “Live HTTP headers” Firefox addon does essentially the same thing, although not quite so smoothly.

    Reply
  5. Leevi Kokko

    Brian,

    thanks for pointing me to the Regex Coach! Since I’m not a native regex speaker, building complicated filters e.g. for GA reporting often tends to take too long – any help in cutting that time down is a lifesaver.

    Cheers,

    -Leevi

    Reply
  6. Andy Dawson

    Thanks for the information Brian. I think Wasp maybe what I’m looking for: I’ve had a few instances of a client updating pages across several International sites and either forgetting or (worse still) adding the wrong GATC!

    I also got your book recently – certainly recommended reading for getting the best out of GA.

    Andy

    Reply

Submit a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Share This