What is the 1st thing to do when considering a web analytics implementation?

[This article is part of a series entitled: GA Implementation ABCs]

What came first?I had an interesting conversation while in Seattle recently. On discussing the progress of the book (95% complete now..!) with a friend from the industry, I described how my last chapter is going to be a sum up all the things learnt plus my thoughts on what’s next for web analytics. For the sum up, I was asked, “so what do you think should be the first thing to do when considering a web analytics implementation?”.

“That’s a great question” I replied and gave the following response: “Tag all your pages i.e. Collect the data”.

The conventional wisdom for web analytics, has traditionally said that before you even choose your preferred vendor for an implementation, you should prepare your web analytics business objectives, map who your stakeholders are, canvas throughout your organisation for KPIs, business plans, marketing plans etc. – anything that relates to the success of the web site.

But, you don’t know what you don’t know…
So I disagree with that school of thought, particularly if you are implementing a web analytics tool for the first time. Adding page tags is more important as your first step because they collect your valuable data. Seeing raw, unfiltered, non-segmented data will help you understand how visitors interact with your web site at a fundamental level.

Reflecting on this during my flight home, the first thing to do when considering a web analytics implementation is actually a combination of two elementary things – tag all your pages, then view reports to understand the initial collected data.

Step 1: Tag all your pages
Ensure you tag ALL your web pages with your tracking code (usually a javascript snippet). The emphasis is on ALL as quite often by default tools will not track download files (PDFs, XLS, DOC, EXE) or external links (banners, affiliates, advertisements). Google Analytics for example uses a function call to urchinTracker to create ‘virtual pageviews’ which you need to manually add to your download links. Also, quite often page tags are simply missed out – usually by human or machine error. For sites with many web pages that are constantly updated/evolving, consider a regular auditing service to ensure all page and virtual pageview data is consistently tracked. In fact regular tag audits are an important aspect for the ongoing management of your web analytics solution and many GAACs offer this service for Google Analytics.

Step 2: Understand the initial data
This is the initial pass by of your data before mapping who in your organisation your stakeholders are and what KPIs are useful for benchmarking your web site against. With good data coming in (usually 2-4 weeks worth), simple questions can then be answered, for example:

  • How many daily visitors do I receive
  • What is my average conversion rate
  • What percentage of visitors are from search engines
  • What are my Top 5 visited pages
  • What is the average visitor time on site
  • What is the average visitor page depth
  • What is the geographic distribution of visitors
  • What is the average visitor bounce rate (single page visits)
  • etc.

So my approach is first to understand the make-up (distribution) of your visitors with some base metrics of what these visitors do. That simply requires the tagging of your pages and an initial understanding of what the reports are showing you. Armed with this information, it is then time to start thinking about your stakeholders, KPIs , and how to benchmark yourself in order to make improvements.

How have you approached the implementation of web analytics? What came first, the data or the stakeholder maps, business plan or KPIs? Please add your thoughts with a comment.

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…

Analytics Predictions 2019

Analytics Predictions 2019

Now and then I tease myself by trying to predict the future, and so in preparation for this years' attempt I found...


  1. Brian Clifton

    Thanks for the feedback Steve. I think we are on the same page on this. The difference is really about the timing – as per my latest post in this series.

    Leaving competitors of Google aside, my point is simply that the ‘serious’ planning required to build an e-metrics framework within a large enterprise requires initial data to support the case. For example, even experts can get it wrong and none of us from a consultancy background are fortune tellers.

    So before embarking on the mammoth task of building a case for web analytics (or any specific product), my approach and recommendation is to first collect as much supporting data as possible. That usually means web visitor data, but can also include server performance, CRM, call centre, support response times and off-line marketing data. This then builds a solid foundation from which expectations can be managed – both from the clients’ and suppliers’ perspective – as well as allowing clear project objectives to be set and potential results to be benchmarked against.

    If the client has an existing web analytics tool in place, then great, data can be pulled from that. But in most cases I have come across, the client has little faith in the numbers, so implementing GA to collect good quality, clean data is my first step. Yes I agree that should be done in conjunction with an expert/consultant who knows what they are doing with the tool and I am very happy to have you guys (SATAMA) providing this service in Finland as well as other regions 🙂

    If using GA takes the client to a higher level and they wish to bring in another tool – that’s great, and means Google Analytics has done its job. The basic law of web analytics is that no one tool can do everything that all potential clients may need. In fact most ‘sophisticated’ clients I have come across actually use multiple tools. That tends to be the case with US clients but I am seeing more occurrences in EMEA now (for example see: who-uses-google-analytics). As you know GA can happily run along side other web analytics tools – in fact I am using GA, IndexTools and Crazyegg all on this site.

    As I mention in my latest post, I recommend discussing KPIs with clients after an initial overview of the visitor data i.e. once there is a basic understand of what is actually happening on the web site from a visitor’s stand point. This is simply because, apart from eCommerce web sites (the vast majority of commercial web sites are not eCommerce enabled), most clients do not know where to begin with KPIs. So initial visitor data is key to this process.

    Anyhow, sorry if I am preaching – got carried away there! There are lots of ways to skin a cat – this is what I have found to be successful with the clients I work with…

  2. Steve Jackson

    Hi Brian,

    For anyone reading this I am part of Google Analytics’ accredited partners (GAAC) supporting Finland. This means I have a great deal of knowledge about the GA system and understand what it can and can’t do. I love GA and I tell and still will continue to tell people that. Google did a great thing when they give GA away. We also do what Brian says in terms of audits and tagging pages as he suggests, as well as training folk to understand the reports. This however is not the point I wanted to make.

    Brian, Sorry mate, but I have to disagree and go with conventional wisdom on this.

    You would agree that Google Analytics is not right for everyone? Competitors of Google for instance (a quite large group) don’t want Google to have access to their data for quitre obvious reasons. So that said it would mean they would have to go elsewhere. It means that standard implementation of a tool comes at a high cost.

    If I am talking about minimum 5 figure numbers, but often 6 or 7 figure numbers with our clients using paid systems, good configuration and implementation is a must. Companies are not going to risk that kind of payment on something they aren’t sure will give them a ROI.

    You can’t do that without serious planning and ideally you want to understand what limitations you have so a KPI workshop at least is imperative. Secondly if there is no business process in the larger enterprises to fan the data out to the people that need it then what is the point? A process is imperative.

    Even with clients who use GA, when we’re talking about tagging complex sites implementation comes at a high cost in time to get it right. Even with GA which is essentially free to collect the data with, there is an associated cost of not doing this right first time.

    I would agree had you said experiment with GA first as an entry point into web analytics, I’ve worked with a lot of clients that have done that and they now have one or two people in the organisation that know the tool. But it doesn’t solve the bigger problem of bringing the web analytics culture into the organization.

    When you say “you don’t know what you don’t know” you’re absolutely right. You don’t know if you’re wasting your time or not until you have as you say “prepared your web analytics business objectives, map who your stakeholders are, canvas throughout your organisation for KPIs, business plans, marketing plans etc. – anything that relates to the success of the web site.”

    To bring my comment to an end, I have to say I have had a lot of success with clients using Google Analytics successfully, primarily because they have gone through the process of developing KPI’s and getting the culture started in the organization.

    I have also worked with a lot more clients who started collecting data with their tool and it took a lot of education just to teach them what they had done wrong in their implementation and set-ups after we had done a KPI workshop with them.

    This stuff is not easy for an average business, even the amazons of the world struggle with this so to me I always advise my clients to be clear why they’re doing what they’re doing before they do it.

  3. Brian Clifton

    Steen: “the same reason it is also very often necessary to clarify what the numbers can’t tell.”

    You are very right! I often come across clients that think web analytics tools are going to perform the analysis (and subsequent optimisation) for them. Web analytics is great for telling you the what and the when, but it cannot tell you why – why did visitors make that behaviour? That’s where the human element comes in.

    I usually illustrate this for clients with the following example:

    If the average time-on-site metric is increasing, is that a good thing? Actually without further investigation there is no way of knowing. It could be that visitors are more engaged with your web site content and therefore spend more time reading it – that is obviously a good thing as your site builds up the relationship with the visitor. But it could also mean that visitors are lost in your navigation system or confused by your content – that is a bad thing.

    Trying to automate that ‘analysis’ process could result in completely the wrong type of optimisation…

  4. Steen Rasmussen

    In my experience the approach very much depends on the knowledge level of the client.

    And as you mention the knowledge level very often requires additional basic input.

    For many companies the the road to professional webanalytics is a building based on knowledge. And often telling about the various numbers is far from enough. You often have to take the next step too. For a lot of first times users the numbers are basically just numbers.

    In most cases it is therefor needed to “translate” the connection between website numbers and the business activities -symptom and cause- before any talk about stakeholders and KPI can commence.

    And for the same reason it is also very often necessary to clarify what the numbers can’t tell.


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