Everyone is an Analyst

Posted on January 10, 2014

TL;DR; use your web analytics expertise to unlock the domain expertise of the client

As I’m sure anyone with a LinkedIn profile and web analytics as a skill knows there is a bit of a skills gap in organisations when it comes to analysing their data.

So businesses hire analysts and, depending on how well they can hire, reports or actionable insights (or possibly both) are produced. The assumption here is that it is only the analyst who has the skills and ability to use the data to produce these reports or insights; they have been hired because these skills are lacking in the organisation.

A few projects I’ve had recently and the strategy I’m developing for another client make me question the usefulness of this assumption.

I don’t think that most people are idiots when it comes to web analytics. Most idiotic behaviour arises from a complete lack of web analytics domain knowledge combined with an attempt to guess the solution whilst stating the problem.

On the flip side, new analysts completely lack of domain knowledge for the organisation they are working with. This leads to reporting on metrics that don’t matter and impractical solution suggestions to unimportant problems. I know this is true; I’ve done it myself.

As I mentioned earlier, the traditional solution to this is for the analyst to learn the problem domain and then to provide insight. The alternative, which I think is the way things are going, is for the web analytics expert to build tools so that the client’s lack of web analytics knowledge doesn’t block them from using their organisation domain knowledge to perform their own analysis.

Google Analytics custom reports are an example of this trend; the analyst uses their domain expertise enable to client to use their domain expertise without the client having to learn lots about Google Analytics.

An interesting corollary to this idea is that unless your client is interested only in what can be reported in Google Analytics you will have to learn how to integrate with other systems. This is part of the trend where I think web analytics will be absorbed by business intelligence within the next 10 years.

This is a big step up in complexity from making Custom Reports. The introduction of Universal Analytics might help plug the gap here, but all it really does is shift the work away from creating the report and onto getting the collectors implemented correctly.

1 Bonus: Creating simple analysis tools

It should maybe be a different blog post to talk about ways to enable analysis for a non-domain expert outside of custom reports but I blog so infrequently that I’m going to stick it in here whilst I’m in flow.

What is the difference between an analysis tool and a report? There isn’t a clear dividing line but I’d say the main differences are that a report is less flexible and less subtle. A report presents and justifies a conclusion that is already decided; a tool enables the user to find their own conclusion (although the sort of conclusion they might want to find can be built in).

Where to start with building tools? The first question is to figure out what analysis you want to enable. Then select the tools necessary to do this.

  • Excel. Whatever tool you make, the data is probably going to end up in Excel anyway so why not skip the middleman and put it directly in this format? A good Excel workbook for analysis is not the same as a good Excel report; for example, things like whether a field is in a row or a column suddenly become much more important when the end user is going to be trying to mung it about themselves. Probably the only option in many cases.
  • Third party tools. If you have the budget there are many, many awesome tools designed for exactly this sort of thing. I’ve only briefly used Qlikview but I was very impressed. A good place to get an idea of what kind of tools are out there and how they might be integrated is the Snowplow blog which frequently has posts on integrating new frontends and visualisations with their data backend.
  • Web Applications. For my client Keyfund I’ve increased the number of analysts by a factor of 10 (potentially a factor of 100 if we roll the tool out further) by presenting a small amount of data on top of Google Maps. This is a big deal because the two main bottlenecks with Keyfund are my lack of time and my lack of domain knowledge. I’m also finding that http://filter.eanalytica.com is saving me enough time that I’m ROI positive from making it with me as the only user.

2 End bonus

When you think about it this way there is actually a hierarchy of enablers:

  1. I create a custom report which enables the client
  2. In doing so I’m enabled by the Google Analytics web interface
  3. I’m also enabled by Doug Hall who has written some javascript that helps me track something I wouldn’t otherwise be able to
  4. Doug’s javascript uses jquery
  5. Turing came up with the idea for a universal computer

At level X something is created using domain specific knowledge of level X which can then empower those on the level X-1 to use their domain specific knowledge.

I am slightly reminded of Up and Down the Ladder of Abstraction which you should definitely have a look at if you haven’t seen it already.

So in conclusion, you are at your most powerful when you enable others to use your domain knowledge to better use their domain knowledge.