Drop everything and buy this book.
If you are struggling to master the art of web analytics (like I am right now), you must use the link above and purchase a copy of Avinash Kaushik's "Web Analytics: An Hour of Day."
As many of my subscribers know, I am working on a screencast about Web Analtyics for nonprofits in general and Google Analytics in particular for NTEN.
Since I am neither an analytics geek or techie, I've been researching the topic and collecting examples. Avanish Kaushik's blog has been invaluable. So have all the stories that people who work in nonprofits have shared with me.
Since I don't work for a nonprofit organization, I needed to get my hands dirty with some real life nonprofit web site data. Ami Dar of the idealist.org graciously offered to share their web site data ("we have nothing to hide") and introduced me to the idealist's web guy, Dave Amos. Today, I was able to spend an hour on the phone on a web conference call with Dave Amos and Avinash Kaushik looking at idealist data . I asked dumb questions and got really smart answers.
I learned a huge amount and I've only been able to transcribe the first 30 minutes!
Overview:
As the web person or IT person in your organization, you are most likely responsible for running web analytics. It is important that you are not the only person in the organization looking at the data and applying it. You have to change your organization's culture to a data-driven culture -- one that looks at data to inform decisions. You have shift people from feeling overloaded by analytics data to encouraging them to consume it. You are no longer just the web guy, you are web analytics steward or translator.
Whether or not you like this new role, you also have to become a marketing analyst. That is, the person in the organization who thinks about the needs of the users and understands the purpose of the web site well enough to translate what information management and program staff need to make decisions and dig into Google Analytics to extract it for them and present them with just the pretty report they need.
In the past, the web manager would round up huge amounts of analytics data and
"throw it over the fence." That's web analytics 1.0. Web analytics 2.0 requires a different approach - to be a bridge between the sea of data and decision-making.
Data must be connected to the purpose of web site and not viewed in a vacuum.
The very first words that came out of Avanish Kaushik's mouth was a question to Dave Amos "What is the purpose of the idealist.org web site?" It wasn't a PH.D dissertation on what various metrics mean or a whirlwind tour of Google Analytics cool features. This is important because what will make us get excited about analytics is connecting the data to mission.
It is important to build a data driven culture in the organization and get other staff members to appreciate and look at web analytics data. Kaushik notes that many nonprofits and small businesses don't have a culture of looking at web data to inform decisions. As Kaushik says, "It's hard and the data can be overwhelming."
He asked Dave if he was using the default web analytics report on the dashboard to share the reports with staff members. This default report is a six-page pdf document. Kaushik observes that the best strategy for changing the organizational culture around using data is to customize the report for staff. He showed us how this can be done easily in google analytics.
He also recommends that the report should only be one page long. Notes Kaushik, "In my experience, people tend to only read or look at one page before their eyes glaze over. Google Analytics can generate reports at the right level of prettiness so that people will actually take a look at it. So, don't overload them."
Finally, Kaushik recommends using the email report function (which can be scheduled for weekly or monthly or "however much you want to spam them.") The important point for web managers is don't be a gatekeeper for the web data. Don't make other staff members ask you for it.
Kaushik shared a step-by-step process for helping to make this culture shift:
A. Identify the key metrics! You need to help staff identify what information they need to inform decisions. You need to go through an upfront inquiry process before you generate a single report. When you talk to staff about web analytics data, refrain from using the word "data" because most people will not have a clear idea what to ask for. It isn't that people are stupid, it is just that they aren't aware of the possibilities that analytics data offers. You, as the IT person or Web Manager, need to show them. It goes beyond knowing which buttons to click.
You need to ask questions like ....
How do you spend money or your time to drive people to our web site?
How are you spending time or money in improving the content, design, or navigation of the web site?
How are you making decisions?
These questions always lead to data or suggest a report.
B. Context is everything! The reports should not show data for a single time period, but show a comparison between two time periods. (This week and last week for example). That is actionable information! The schedule reports function has a "check box" for "send data comparison" if that is checked, and it will automatically send a comparison report comparing the weekly data with last week's data. Notes Kaushik, "This is really important for people who don't understand data. The reports show "red" and "green" (red decrease, green increase) and everyone immediately understands what is good news and what isn't."
C. Don't give them all report views, only the most important. The new Google Analytics interface allows you to easily drill down and explore your web site data. Don't just hand people the top view. Begin with generating a report that answers the common questions from staff. Next create a customized report for them by drilling down into the most important view. You can add that report to the dashboard so it is automated. Kaushik recommends that you zero in on the clearest visual or stat that answers the question. Suggests Kaushik, "Edit out all the extra junk in the default report that people don't
care about. Once you overload them more information, they stop paying attention."
D. Use the schedule feature. Once you have a good understanding of the information that people really want, create that specific report for them and schedule the report to be sent weekly or monthly. Don't be a gate keeper.
E. Give them more detail or additional reports when they ask for it! Kaushik suggests that over time people will begin to ask more questions about the data or request additional reports. You are teaching them to interpret and apply analytics data one step at a time.
Kaushik says that by following the steps above, the culture shift starts to happen. In some organizations it happens faster and in other organizations it happens slower.
Kaushik's Favorite Metric (for Web Sites)
Bounce Rate
"It's a sexy metric!"
It is leading indicator of the quality of content on your web site and whether or not you are attracting the right traffic to your site.
Bounce rate measures people who come to your web site and leave instantly. There are different definitions of "instantly" based what tool you are using. Google Analytics defines it as "anyone who has come to your web
site and seen one page and left." Bounce rate is the best indicator of whether or not you are getting the right people to visit your site.
The bounce rate are the people who are coming to your web site, giving you one glance, and they leave. Why? The page they landed on sucked! The page didn't connect to them. Maybe the wrong keywords?
This is a great way to know that you're getting the right kind of traffic. If you're getting a high bounce rate, take a look at the page or the keywords you're using or whatever, and ask why. Keep
drilling down.
Be careful using it on blogs. The reason is that with blogs people can visit the top page of your blogs and read everything about you - the last five posts, your about page, and more. So you might come to my blog, spend a lot of time on my blog reading everything and then leave. But, you would be counted as part of the bounce rate (on Google Analytics) because you only visited one page.
What is the best metric for blogs?
Well, it isn't time on site, either. That's because of the way that analytics programs compute time on page. Watch the number of unique visitors over time and the percent of new visitors to my blog. I watch for every day and week, percent of new referring urls. I want to see if my blog is being linked from different places.
If I were to pick one metric, I would recommend to measure the number of RSS subscribers. I'm trying to build an audience that will consume my content. The RSS subscriber is a vote of confidence that I am writing the type of content for which they are going to extra pain to sign up. Not the absolute number, but the growth. RSS subscription is hurdle is a barrier to cross.
I'll spending quite a bit more than an hour an day on topic in the coming weeks. So stay tuned.
Beth, I think you'd be even more focused on web analytics if you tried to make money out of your blog using Google Ads. I've been experimenting with Google Ads on my blog. It's addictive -- kinda like slow-motion slot machines. I made $50 on my blog this month so far but I spent $80+ to do it. It's a cheap lesson when you think about it.
Posted by: abenamer | June 19, 2007 at 11:45 AM