Principles of Web Analytics: #PCampLDN

Peter O’Neill
Reading time: 5 minutes
19th January 2012

Note: This post has moved from to Ayima as part of the 2018 acquisition.

ProductCampLondon Logo

I attended my first Product Camp on Sat 26th Nov, an unconference for Product Managers.  Obviously this is not my field but I like hearing what other people are talking about and telling everyone about the wonders of web analytics.  Thanks to the sponsors and organisers – kudos for a very smoothly run event.

As an unconference, everyone should be prepared to run a session although it doesn’t need to be a formal presentation.  I had previously put together presentations introducing people to web analytics but decided to aim for a more relaxed approach this time just making some notes on my iPad.

I wish I had been braver and gone for the big room before it was fully booked out but still had around 20 ppl cram into a room to learn about web analytics.  It was a good discussion and I hope everyone understand at least a little more by the end.  Following are some of the points I made.


I commenced the session by asking some questions of the crowd.  These questions are listed below along with an idea of the responses (from memory)

  • Who uses web analytics?
    • Had about two thirds of the room say they were using web analytics
  • Which tools do you use?
    • While GA was most popular, also had SiteCatalyst, WebTrends and Coremetrics users in the room – even had people whose companies were using SiteCatalyst Insight
  • What do you use it for?
    • Basic reporting was sadly common but there were a few people in the room using web analytics to identify issues and evaluate the impact of changes
  • What are your questions/issues around web analytics?
    • Range of responses including:
      • Data accuracy
      • Usability of web analytics tools
      • Making sense of the data
      • Website features not tracked

To get everyone in the right frame of mind, I then showed the recent Google Analytics Checkout video – it demonstrates the experience many people receive on retail websites and it is this experience that everyone wants to fix.

Purpose of Web Analytics

Possibly inspired by the recent web analytics discussion on the definition of web analytics, I wanted to make clear the purpose of web analytics.  My definition for this is:

To provide intelligence that informs business decisions leading to an improvement in performance for online organisations. 

The key element in that line for me is business decisions – I keep on returning to this with everything I talk about, that web analytics data has to be used to make better decisions.

Requirements to Receive Value from Web Analytics

Moving on from there, I felt the need to emphasise that if you want to receive value from web analytics, you needed to invest time into it (time obviously costing money).  My list of tasks involved in setting up and then using web analytics (not a complete list) was:

  • Define your business objectives
  • Translate business objectives into KPIs, metrics and reporting requirements
  • Translate these into data capture requirements
  • Set up the web analytics tool – implementation and configuration
  • Develop dashboard templates for easy access to an overview of performance
  • Review and interpret data
  •  Integrate web analytics into business processes
  • Take actions based on intelligence gained

Nobody said it would be easy…


At some point, caveats need to be provided around web analytics.  I am not a salesperson so like to get these potential surprises out of the way early on so people can do things right from the start.

  • Data is not 100% accurate
    • Visitors without JavaScript are not counted, implementations are flawed, people use multiple devices, etc
    • Check out Brian Clifton’s White Paper on Data Accuracy for more details
    • Solution – get over it – the data sample is large enough for the data to be useful and inform your business decisions.
  • Requirements need to be defined in advance and your web analytics tool set up to reflect this
    • So data is available when you need to answer questions
  • Different skills are required to define requirements, implement tracking code and interpret reports
    • There are very very few people with all of these skills so likely to need multiple resources

Using the Data

Sticky notes saying Make Things Happen

All this may suggest that once you have the right data in your web analytics tool, the work is done.  I felt there was a need to emphasise that this is only the start, that if the data is not being used, there is no point to web analytics.

  • Data can provide the “what” of website/business issues
  • It rarely provides the answers or solutions to these issues
  • Once the issues have been identified, move into a testing phase of potential solutions
  • Where you will be using the data again to evaluate the best solution

Impact on Product Managers

Tying all this back to the audience of Product Managers

  • Everything on a website can be tracked and reported on
    • Identifying what and how requires creative thinking based on experience and product knowledge
  • Product Managers should be using web analytics data to understand and evaluate the performance of websites and product features
    • To inform the decisions they are making ensuring these are the best options to improve performance


The session was quite well received and I even received some votes for best session.  There was a spare slot at the last session and so I ran an Introduction to Google Analytics training session which also appeared to be appreciated.  I felt like a bit of an interloper at this sort of event as it is not about web analytics.  Reassurance came from various people that web analytics is getting more and more notice from the Product Management community and that any insights into using it are appreciated.

I have been to a few unconferences now and intend to attend more in the future.  I think it would work for the web analytics community once people got their head around the approach and indeed, Stephane Hamel used an unconference approach at an event in Quebec in Jun 2011 (event write-up from iPerceptions).  I have some ideas germinating.

Technorati blog claim code – 67X6PG8YC2FQ

Written By Peter O’Neill
Asset 1 Asset 1 Asset 3