Website Analytics September 8th, 2009
Last month we looked at 'Ensuring Data Accuracy' within your Google Analytics profile. To summarize this post, there is little to no point reporting on data which doesn’t accurately reflect what your true visitor is doing on your site. I cannot stress enough the importance of continually reviewing your data quality to find ways to improve.
This month’s tip is the next important item in the analyst toolbox and will hopefully reinforce the whole purpose of web analysis — improving online performance! We should always be mindful of this overall objective when we analyze. It is all too easy to spend hours looking at the technicalities of analytics or sifting through impressive data sets but achieve very little in the way of providing business benefit.
My point here is to “Provide actionable reporting so that website/marketing executives and business owners can make informed decisions.
Executives usually want the high level information. They need to make decisions that improve their business and most will do so quickly and easily if there is clarity in what they are reviewing. This is the real challenge for analysts, what we are paid for, and the difference between a decent and great service.
Here is the process I follow in providing actionable reporting for yourself or your client.
Talk the talk, walk the walk, but truly understand why you are talking and walking!
In order to understand why results show up as they do and speak with authority and credibility you need to grasp how data is recorded into profiles and the nuances of why it is not always as expected. Additionally, you really need to be working within a process or framework to be able to data pattern properly, find insight and report this effectively over time.
I don’t want to get bogged down in technical descriptions of analytics nor processes in this post. I’m keeping it high level and hopefully actionable 🙂 There are a plethora of resources available online to review. For understanding technicalities, the Brian Clifton and Avinash Kaushik books are great starting points for learning the nuances of data collection and reporting.
Needless to say there is no need to re-invent the wheel for developing your process, just find a framework to use that suits your environment and revise as necessary.
Here are a couple to review:
1) From Alex Cohen, even though he calls it a BASIC analysis process, it provides great detail into the individual tasks that allow collection and reporting. It has a team focus and works well in consultative environments.
2) Eric Peterson has a wealth of information on his blog and is a great Twitterer. His white paper on Web Analytics Business Process is a corporate top down approach to analytics success.
We have a high customer to analyst ratio in house and we instinctively pattern to show insights and tailor results towards the needs of the client. We use a more generic framework that allows the flexibility we need to deal with varied customer requirements.
Step 1 – Examine the site and data and identify where to target analysis. Ex. Maximizing revenue for your top sources
Step 2 – Create a strategy for measuring and reporting around this goal. Identify the KPI’s that show trends and map out the metrics to examine.
Step 3 – Create a format to represent this data and draw insights
Step 4 – Review the data alongside the site and create actions for improved performance
Step 5 – Recommend / act on these actions and measure the results
Google Analytics presents data in a very intuitive format. In order to find certain insights or pattern a specific metric (especially over a period of time) it is recommended using a custom report to present data sets together. The Google Analytics custom report tool is great for selecting and presenting the relevant information quickly and easily. Once the report is created to your liking, you can save it and refer back to it when the next review comes around.
Lets look at an example of why and how we would do this.
We wish to analyze the performance of a set of corporate Pay Per Click campaigns to be able to determine the affects of campaign management on performance.
We really need to pull various metrics together to understand how effective some campaigns are being. So lets build a custom report:
Step 1 – Click on “Custom Reporting”
Step 2- Create a new report
Step 3 – Enter a title – e.g. Campaign Ecommerce Performance
Step 4 – Add the “Campaign” Dimensions
Step 5 – Add the following Metrics – “visits/impressions/CTR/transactions/cost/cost per transaction/revenue/average value”
Step 6 – Click Create Report
Here is the result:
Hey presto! A more useful set of data to report on. But we are not finished yet!
Our next requirement is to display this information in the same language for the executive only interested in high level data.
No one will understand the data in the detail that you (the analyst) does. Putting it into a format for executives and owners is really the key to being able to show the value of an analytics service. They are paying for your time and effort and at some point they will ask the question, “How is this analysis affecting my bottom line”. We need to get data into the best possible indicators of change.
Step 1) Speak to your customer to understand their needs.
Step 2) Export the relevant data from analytics
Step 3) Use your powerful analyst grey-matter eye to manipulate data into scorecards
Step 4) Track this data over time and monitor the affects of suggested changes to see positive results.
During these steps of analysis and scorecard creation if you have a mindset of dogged curiosity, then we find that insights naturally occur. Experience helps, but the enthusiastic determination of a Jack Russell chasing a bone flavoured ball help even more.
Continually ask “Why am I seeing this rise or fall in this data”. Segment and/or drill into it to get the answer. Recreate reports, link different metrics together. You will find answers. Keep going until things make sense!
When a light bulb goes off, you’ve hit an insight. Record it fast with enough detail to decipher at a later time, and get on with the next one!
Here is where we have been trying to get to. All this effort has led us to this point. Up until now we’re simply reporting what is going on. Business generally invests in analytics to make more money, rather than establish why they are making the money that they currently are. We need to take all our efforts and channel them into making recommendations that have real business benefit.
When I’ve spent a while immersed in data and firmly engaged in curiosity mode, I find it necessary to have a break to change mindsets. We are moving from a micro analysis of what is going on to big picture mode. So grab a cup of green tea, open the site and your list of insights and get going.
Look at improvement from the perspective of website content, usability, campaign targeting, budgeting just for starters. You should naturally have some things to A/B or multi variate test. Really get stuck in and enjoy yourself! Not many jobs allow this creative thinking phase that can bring such benefit to organizations.
Think inside and outside the box. Or create a new box! Refine your ideas into those that you have data to support the case for improvement and understand how you are going to measure the results!
Now sit back and relax, light a metaphorical cigarette and ask whether it was good for yourself.
Enjoy this moment because next we have to clean it all up and make it presentable!
Now that we’ve poured blood sweat and tears into collecting clean data, establishing amazing insights, and developing a set of actions that, it all counts for naught if we don’t get the information in front of the right decision maker(s) and in the right format.
Here is how:
Step 1) Understand the role and technical level of your audience. Gear your presentation to this level and adjust as you go along.
Step 2) Educate your audience as you present so that they can “get it”.
Step 3) Provide data in the simplest format possible to make your point (but have the detail to back it up!). Executive summaries are ESSENTIAL!
Step 4) Highlight your insights but focus a great deal on the business benefit of suggested actions.
Step 5) Ensure the meeting does not end with out owners and deadlines for the agreed actions, and a next meeting date to review the results of the changes.
The aim of an analyst is really quite simple – “to provide business benefit”. In practice it takes a skill, structure and the right approach to consistently meet this aim at the highest level . The above recommendations will hopefully help, whether you are a small site owner, dabbling with analytics yourself, or an expert that consults to a customer base.
Google calls it 'Data Driven Decisions', Yahoo! refers to it as 'Data Driven Insights', I like to refer to it as “Actionable Reporting” and it is fundamental for customers to be able make instinctive, informed and quick decisions.