Over two years ago, I made a screencast about Web Analytics for Nonprofits that covered the basics of using google analytics, web metrics, and some nonprofit case studies featuring Laura White and the Idealist. That's just about when I discovered Avinash Kaushik's blog, Occam's Razor and his first book on Web analytics.
I reached out to Avinash and he was very generous with his time, helping us understand traditional web metrics as well as Google Analytics. Through our conversation, we also touched on the metrics for blogs. I've learned a lot about metrics from Avinash.
Avinash has just published his second book, Web Analytics 2.0. It's a desktop bible for anyone who has to gather web analytics data as part of their job. What I like best is that it takes the mystery out of social media metrics. It gets even better: Avinash is donates all proceeds from his books to two charities: The Smile Train and The Ekai Vidyalaya Foundation.
Harry and Sara say "Thank You Avinash" for the Google Schwag!
Avinash Kaushik sent me a copy of his new book and it's so good, I'm keeping it. But, I'm purchasing a copy to giveaway to the reader from a nonprofit who leaves a comment about how they might best use this book. (Don't forget to fill out the comment form completely, so I can track you down if you win.)
I had the pleasure having lunch with Avinash over at the Googleplex. The chapter that caught my curiosity was Chapter 7: Failing Faster - Unleashing the Power of Testing and Experimentation and we discussed it in the video above. He explains why experimentation is critical for success in using the web, particularly social media.
Avinash feels that in a world of finite resources, it is very important to experiment and fail fast. With social media and on the web, experiments are fast, cheap, and scalable. The learning that results is what brings your more success. Experimentation also helps an organization make decisions based on audience feedback and analytics data, not your own hunches. This try it, fix it approach leads to incremental improvements which in turn leads to better outcomes.
The F-word chapter (Failure) offers some really useful tips about creating and nurturing a "experiment culture." I was thrilled to discover this part of the book because I'm designing a learning community/technical project that is based on valuing experimentation. So, been working on a methodology for social media experiments.
While the advice in the book is geared for tests to improve a web site, these are translatable to social media experiments. I've summarized a couple of the tips he offers about methods for testing.
(1) The First Test: KISS: The first experiment should be simple from an idea, execution, and measurement and use A/B method. This is a technique for testing two or more versions.
(2) Just Get Started: Avoid spending so much time trying to design the perfect experiment with the perfect measurement tool that you don't actually implement. Learning means implementation - even if you fail.
(3) Test to Learn, Not Validate Your Gut: Don't think about testing as a way to support a decision that you're making based on your gut. Do it to learn what works or doesn't.
(4) Start with a Hypothesis: Your hypothesis should embed a success metric. For example, "My hypothesis is that our Facebook Fans are more likely to engage with us when we post links that have a question in the title."
(5) Make Goals Evaluation Criteria and Up-Front Decisions: It is important to not only identify your success metric, but also establish the criteria to judge a victory.
(6) Design Tests That Solve A Pain Point for Your Audience: Design your experiments so they address a point of pain for a customer or audience.
(7) Learn, Learn, Learn: If you're going to experiment, you need to analyze your data and learn from it. Even if your social media experiment was a miserable flop, there is a lot of valuable learning.
(8) Evangelism and Expertise: It is important to have someone who can preach and share why testing is important and someone who has the expertise in metrics and data.
Finally, Avinash suggests that testing should be fun!