14: Staffing and Tuning Your Web Team
(EN: At the time this book was written, Web analytics was still a young profession, so the author's approach is to seek out individuals who have no background or training in analytics. It might be argued that this is no longer needed, and that you can find trained and experienced professionals in the discipline. However, it remains true that no t everyone who has experience doing a job is capable of doing it well.)
Primarily, a Web analyst should have interpretive expertise. The author notes that the job-boards are filled with candidates whose experience is installing, configuring, and running Web analytics software - but this doesn't mean they understand how the software works, what it measures, and can assess whether it's measuring the right things.
Software skills are needed, but of significantly less importance than the ability to collect and interpret information. A major cause of dissatisfaction with analytics is that companies pay a great deal of money for a solution, hire an "analyst" to operate it, and get a monthly report full of data that does nothing for their business.
Some of the key skills for an analyst include the understanding of statistics - particularly, the ability to understand correlation and cause and effect (not just to do the calculations, but to interpret what they mean. An understanding of strategy and tactics is also valuable - in being able to know how to define success, measure progress toward it, and disregard information taht is not germane to it. Depending on the specific site or project, a person with a background in marketing statistics of operations statistics may be particularly well-suited.
Also, keep in mind that the job of a Web analyst is not all heads-down spreadsheet work. In addition to running calculations and developing reports, an analyst must be able to interact with people within an organization, to discuss the needs that analytics can address, present the outcome of an analysis, and "sell" the value of analysis within an organization.
Internal and External Teams
The author repeats the need for human analysts to interpret information and make it meaningful to a company - that software alone is not enough. He cites research ((Jupiter) that indicates companies that have at least one dedicated web analyst are "at least twice as likely" as companies that use software alone to undertake initiatives to improve their conversion rate, search engine marketing, and other processes to incrementally improve their sites.
The author weighs the merits of an in-house team versus an outside agency, implying that most firms should seek to use a combination of the two. What follows is a fairly standard comparison of the benefits of internal or external resources that is common to any such decision. Essentially ...
An external resource may be less expensive if they are needed only periodically. They are usually broadly experienced and can set to work immediately. Any insight they gain may be shared with the industry, including your competitors, regardless of non-disclosure agreements. While they are not immune to internal politics, they are more likely to be able to speak freely without fear of the consequences if they are the bearer of bad news.
An internal resource may be less expensive if they are needed often. They may not have a lot of experience to start with and, over time, their focus will be narrowed to your company and industry. Any insight they gain remains within the organization. They are also subject, over time, to the wiles of internal politics and may be reluctant to reveal information that will not be happily received
Training for Web Analysts
Given that web analytics is a new field, may analysts are self-educated in applying their existing skills to the task and learning new ones as needed. There are a few trade books, some blogs, and online communities where people struggling to learn the new profession can find information and support.
There is also a professional association (the Web Analytics Association) that offers educational resources, organizes conferences, and supports the professional community.
The author mentions vendor training, but notes that this is mostly involved with technical matters related to a particular tool - how to install and configure the vendor's software. This is good training for the IT resources that will maintain the software, but of little value to the analysts who will design and interpret analysis. (EN: In fact, vendor training is often counterproductive - the analyst is trained to rely on software rather than exercise independent judgment. When you have a hammer, every problem is treated as a nail.)
He also mentions that agency partners can help- train your internal team by working in conjunction with them. (EN: possibly, but the agency is not interested in training your people, as it means less work for them in future, so their interest is in keeping you from becoming independent. Few will say that outright, but it will color the relationship.)