In my previous post, Reporting vs. Analysis, we established what the differences are between the two and hopefully you realized that analysis is the most important of the two. Reporting returns little to no ROI. Also, if reporting is done properly it will lead you right back to analysis. Because if an anomaly is detected in reporting you typically need to dive into data to understand what is going on and define the next steps.
So over the next months I’ll be sharing posts from a step by step perspective using a ‘real world’ example from an of an analysis that I did some time ago. I’ll be going through what I did in Excel and things I did in SiteCatalyst to give you some insights into the process invovled in doing analysis.
Why is analysis so difficult to learn?
I guess you can compare doing analysis with a bird learning to fly. Momma bird spends the first two weeks lecturing the baby birds the mechanics for flying. But it isn’t until she actually kicks thems out and they start flapping their wings that they begin understanding what she’s been talking about. It is easy to talk about, but you can’t learn it by talking. It has to be experienced. Flying is like doing analysis, we can present all day long best practices for doing analysis, but until you do it and have success, you really are not going to know how to do it.
Also, one of the things that makes analysis so hard to teach and so hard to talk about is that it is an extremely unstructured activity, that requires tremendous amount of creativity, critical thinking and proffesional judgement. One of the things that I have come to realize is that not two people will do an analysis the same. It is very individual and unique excersize.
When I’m doing analysis I follow a 4 step framework:
This is not an official framework for doing analysis and certainly not one you have to follow, but it is working well for me. And if you aren’t used to doing analysis it may be a good framework to start with. There are most likely better ways of doing this, but the above is working for me.
Create your list of rocks
The first step is a matter of brainstorming on what ‘rocks’ that are relevant for your site/data. It shouldn’t take too long. Never the less, it is an important step as it will help you keep focus on the right things as we move along.
There are several places you can look to get some ideas flowing to come up with your ‘rocks’.
Playbook or Business Requirements Document – If you have this document, it is the best place to get ideas.
If you had Adobe Consulting support you on your implementation, they delivered a PDF file named a Playbook. The Playbook contains a description of each of the solutions that you were recommended to implement on your site. All based on the requirements that was gathered during the project.
Solution Design Reference – With your playbook you have also received a solution design reference which is a map over all the variables enabled on the report suite. This is an excellent place to go cherry pick. Have a look at the different variables and what they are used for and pick out the ones that seems to be the most interesting ones.
Vertical/Industry expertise – You may be an expert yourself within the industry you work in. So use that to come up with ideas. If you’re not an expert yourself, I’m sure one of your colleagues are. Offer him coffee and go pick his brain. A search on Google wouldn’t hurt either.
Common rocks – There are always some common rocks that you can look under. You can consider these as the gifts that keep on giving as, regardless of the client and regardless of the situation, you can always peak under these and you will almost always find something. These are relevant across all clients and includes things like page fold analysis, bounce rate analysis and page participation.
Dashboards – Take a look at your dashboards being sent around internally. This should give you a good idea on what KPIs are important and may help you come up with some more rocks to look under
Copy from me – I came up with the below based on past experience, looking at their playbook and solution design reference.
The data I’ve based the analysis on is from a bank – let’s call it, Giant National Bank. Giant National Bank make money like every other bank:
- More saving and checking accounts
- More money in those accounts
- More loans and credit cards
Their KBR’s (Key Business Requirements) are:
- Increase account signups
- Decrease customer servicing costs
- Improve external campaign investment performance
The analysis was done prior to Adobe Analytics, which meant Giant National Bank didn’t have Discover (now called, Adhoc Analysis), so the only tools I used for the analysis was SiteCatalyst (now called Reports & Analytics) and Microsoft Excel. This only means that you can do more and cooler things as, if you’re a Adobe Analytics customer, have access to Discover and Data Warehouse.
This is the list of rocks I came up with:
Some of these aren’t relevant if I were to do the analysis today. Here I am specifically thinking about ‘Branded Terms Analysis’ as Google have removed search keywords. If this is news to you, it is worth taking a look at Brent Dykes blog post, How Google’s Expanded search encryption impacts Adobe Analytics.
It is now time for you to come up with your own list of rocks that you want to look under. In the next post we’ll begin ‘kicking the tyres’ which will help you understand which of the rocks here you should be focusing on.
Now, go and come up with your own list of rocks, so that you’re ready for the next post – Warm Up Analysis. And feel free to share your ‘rocks’ in the comments below.