Jennifer White on Web Analytics Wednesdays

I've had a fascination with all manner of data analysis since my days as a bookstore owner. So when I first heard news of a local web analytics group, I had more than a little interest. I asked Jennifer White, the chief instigator of Web Analytics Wednesdays Austin, to give me a little background on the group.

Lynn Bender:: Jennifer, tell me about the group.

Jennifer White: Web Analytics Wednesday is the world's only social networking event for web analytics professionals. Eric T. Peterson, an author and a global web analytics community leader, founded Web Analytics Wednesday as a global effort to put "faces with names" and to get local members of the web analytics community networking. I started the Austin group in 2 years ago in July.

The local group focuses mainly on web analytics, but we also talk about analytics for other media such as Twitter or search, but those tie back to web analytics. For example, how does Twitter influence bounce rate (which is when a visitor comes to only one of your web pages and leaves the site immediately)?

Lynn Bender:: What about non web data -- like email? Are e-mail response rates, direct mail campaign data, sales and lead information also part of web analytics?

Jennifer White: Since the web is so central to most calls to action within the tactics you mention, they do play a part in web analytics. So there may be informal discussions around the lunch table for tactical measurement, but generally we talk about analytics tools, problems with those tools that we need help solving or just sharing ideas.

We also have speakers that talk about various subjects. Last month, Ed Wu from Dell spoke about his recent election to the Web Analytics Association board and what he learned at the recent eMetrics Summit in San Jose. In June we'll be talking about search.

Lynn Bender:: The line between what is and is not the web seems to be coming increasingly blurred. Tools like tweetdeck and twhirl are browser-like for the end user, but the nature of the data is often closer to message traffic (aim/gtalk,etc). Would 'internet analytics' be a more appropriate term?

Jennifer White: Yes, the line is becoming blurred in that lots of things happen on the "internet". But I think in this case, "web analytics" and "internet analytics" are two different things. I would say that "web analytics" is about measuring your own or your company's website. You can't know what the metrics for the tools like tweetdeck and twhirl, in the same way that your website is (unless you are the owner of those sites). There are ways to measure how your brand is doing on Twitter, how things are tracking back to you, etc. Perhaps "internet analytics" could be an umbrella term for all things on the web that is not your website. Banner advertising and paid search could be lumped into "internet analytics" too.

Lynn Bender:: Would I be correct in saying that there are two parts to web analytics -- the collection and the analysis of the data? Does the Web Analytics Wednesday group discuss data collection methods as well as the analysis of that data?

Jennifer White: We talk about data collection in terms of the vendors in the space, what the experiences with the vendors are and new ways to maximize the tool in question. Some of the team is more technical than others so the complexity of those conversations varies.

"Don't be a reporting monkey!" is one of the rallying cries that one of the leaders in the industry uses. Unfortunately it is very easy to be a reporting monkey. One of the challenges analysts have, and we have discussed in the group, is how you sell the value of analytics to your organization and how to best use an analyst's time. Certainly collection and analysis is part of the job. There are a lot of analysts that also need to sell the value of the metrics and demonstrate why web metrics can change the business. Things like educating internally about why "10,000 hits this month" isn't a particularly meaningful metric. We have to change the question to ask what those 10,000 hits did after they arrived and what were the implications of those activities (purchase, registration, etc).

Analysts not only interpret the data but also make recommendations on changing customer behavior on the web to make the business more profitable. Actionable analysis is the name of the game.

Lynn Bender: Back when web analytics worked primarily with server logfiles, the web analyst was limited primarily to on-site data. Increasingly, content producers are concern with qualities such as "buzz" and visibility" -- metrics which require the collection of off-site data as well. Is this now considered part of web analytics?

Jennifer White: Offline data such as events, print advertising or brick and mortar store traffic can be correlated with web analytics, but it's not easy. Avanish Kaushik has a great blog post about numerous ways you can tie in offline activities to your website metrics. (http://www.kaushik.net/avinash/2008/07/tracking-offline-conversions-hope-seven-best-practices-bonus-tips.html) For example, you can see how many people are using your store locator functionality. If there is a large number of people from a particular area searching where the closest store is, you might infer that the store is not easy to find. Checking with your on-the-ground team, they might point out that they are hard to see from the road which might lead to some tactical changes such as more signage. Site exit surveys can also help marry offline and online data.

Lynn Bender: What kind of background is needed to be a successful web analytics ninja? Calculus? Statistics? IT? Marketing?

Jennifer White: It really varies depending on the company and the role of the individual. A larger company with ecommerce might need analysts that are skilled in things statistics and predictive analysis. A smaller nonprofit will probably need a "jack of all trades". Really, the most important skills for web analysts are critical thinking, being open to new ideas, and a hunger for knowledge. Business experience and organization agility are also helpful skills. Coding skills come in handy. I might be biased because my background is marketing, but I think my having a marketing background is helpful. I feel more confident in making marketing recommendations or interpreting data based on an understanding of the marketing tactics that drove the traffic.

Lynn Bender: Since the matter of analytics is data, is any kind of database knowledge needed? SQL? Excel? Pivot tables?

Jennifer White: All of those are useful, but not all are required because of the variety of tools available. For example, I have not had the need to learn SQL when using Google Analytics. I would also add PowerPoint since many organizations can't function without "a deck" and a sense of how to display information to non-analysts. HTML is also a helpful skill, but if you are teamed with a technical programmer, it might not be critical.

Lynn Bender: For the budding web analyst, other than joining your group -- which they should do, of course -- what recommendations do you have for developing their skill set? Any favorite books?

Jennifer White: There is a surprising large community of web analysts for a relatively new industry. Our group has almost 100 people. There is a Yahoo! Group (http://tech.groups.yahoo.com/group/webanalytics/) with a worldwide community that is very supportive. I would recommend Eric Peterson's book Web Analytics Demystified and Avinash Kaushik's book Web Analytics an Hour a Day as must reads when getting started. There are tons of blogs; my personal favorite is Occam's Razor by Avinash Kaushik. All of those sites have cross links to other blogs. If you are able to travel to a conference, The eMetrics Optimization Summit is a good one.

Lynn Bender: How can people connect with the local group?

Jennifer White: We have a LinkedIn group: Web Analytics Wednesday - Austin. I also have a mailing list I maintain if I speak with someone directly or they ask me to add them.

Financial support for GeekAustin interviews provided by Smart Bear Software.