A couple of weeks ago I asked why web analytics is so hard – why we never seem to find enough time for it. In the meantime, I had the chance to discuss things with Jim Sterne, internationally renowned web analytics expert. Jim is the author of the book Web Metrics: Proven Methods for Measuring Web Site Success, and masters the task of breaking web analytics down into manageable parts.
I quizzed Jim on why web analytics seems to be so difficult for most and what you can do to get started:
Why is it so hard to know what to do with the data you collect?
Measuring is a chore. It requires a commitment to tools and technology and takes some time to get right. Web pages and events have to be tagged, data has to be verified and reports need to be validated. It’s a chore, but it is tactical. That means there is a beginning, a middle and an end. It requires some training. It is building a bridge or completing a course of study. It is the work of a skilled technician.
Analysis is what one does with the results. This is not a clearly defined series of tasks but a job of applying intuition, imagination and ingenuity. An analyst’s job is to come up with a hypothesis, create a test for it and come to a conclusion. This requires talent, resources and the political freedom to exercise them .
What are some typical traps that we fall into?
The first trap is assuming that the tool is doing the work. In fact, the tool is only a tool. Lumber is only lumber. Once you bring together the skills of an architect, a designer, a builder and a craftsman, then you can end up with a house. There are some tools that do very specific things like multivariate testing. That’s very valuable as a point-solution but it does not provide the process and infrastructure necessary to optimize a company’s website, marketing or business. That takes talent as well as skill.
The next trap is trusting the reports without understanding how the figures have been derived. Understanding the data elements themselves is critical for the creative process that follows. Knowing that steel has a certain tensile strength is useful to an engineer building a bridge, but if that engineer understands the circumstances of that material’s manufacture and the unique conditions of its use he can use it in more creative ways.
Many organisations, that are not selling a product, don’t know what to measure. What is your take?
For any organisation, it’s all about identifying specific goals. A direct-sales organisation will want to measure the number and quality of leads that come in through the website and where they came from. A government agency will want to improve constituent services, improve citizen satisfaction and lower costs. Each organisation (and even each project) within each organization will have its own set of goals to measure against.
You have an emetrics maturity model – how does that work?
A maturity model simply lists out the categories of how advanced a company is at a given discipline. For web intelligence, the first level is simple reporting, the next is website optimization and then the levels move up through behavioral targeting, dynamic promotions and all the way to business intelligence. This is an embryonic framework. Also, more work is being done on this by independent consultant Stephane Hamel as well as Webtrends.
If you want to act on analytics today, where should you start?
Goals! It’s all about knowing what you are trying to accomplish. Then work backwards – if high quality leads are the goal, then start with a specific definition of a high quality lead. Next, look at your current high quality leads and see how they navigated through your website and how they came to it in the first place. That’s the first step in figuring out how to find more of them.
