My "web analytics" story

I thought this story is exceptionally nerdy/geeky, but that is who I am! I found this image on sxc.hu, so I thought it would make it a little more interesting. Onward!

Monitoring a large site, sometime it is difficult to figure out what to do next. Website traffic goes up (yeah!), or down (oh S*^%). What does it really mean though? If it goes up, what are you doing right? Is it because of your great design skills, or simply a seasonal shift that fluctuates based on holidays or the economy? Or more likely something else that you never thought of!

It is a question all good web designers struggle with and don’t really learn about in school. I knew I was young and naive, but boy I had no idea until recently.  It wasn’t until a couple of Decembers ago that I finally was a convert to taking web statistics seriously.

First Encounter

Back in December 2007, I flew out to Chicago by myself for a Search Engine Marketing Conference. Seth Godin was the keynote speaker. I didn’t know anything about him at the time, but I snatched about 8 copies of his book “Meatball Sundae” when I checked in to the uber big Hilton Hotel. I later found out he is pretty much the “god” of marketing and had a lot of views on marketing on the internet.

Going in, I was mainly focused on SEO (search engine optimization) tactics, but I was open to new things. By the time I left Tuesday night, I transformed into a new designer. The main focus wasn’t specifically about web analytics, but those sessions resonated with me so well that I knew that it was where the meat & potatoes are for being a web guru.

What I found out is that web analytics is more than just looking at numbers and seeing how you can improve sales. It is about truly understanding your audience and helping them do what they came to do. This hit me like a ton of bricks and I began doing more research into the subject. I searched around Amazon and found the book Web Analytics:An Hour a Day, so I thought I would see if it was good.

Enter Avinash Kaushik

The book says an hour a day, but I just couldn’t put it down when I started. I took out my trusty highlighter and blew through the whole book (480 pages) in about a month. This isn’t a book review, but I knew I would need to highlight it because there was no way I would be able to retain everything the first time through. It went from the ground up from implentation stategies to custom segmenting and multivariate testing. I enjoyed it so much, I subscribed to his RSS feed blog and make comments when applicable. He has even answered my questions on his blog posts. My question is #3 (go Avinash!)

Real World Applications

When I got to where I work, the web statistics were dismal. The KPIs (key performance indicators) they were using had no actionability at all. They were the basic aggregated metrics like unique visitors, average time on site, average page views, etc. I could tell the marketing director was frustrated at the data, but didn’t know what to do.

To make it worse, they were spending $30,000/year on a tool that they could not leverage very well. I knew something had to change, but I didn’t have the knowledge or experience at the time to come up with better solutions. After reading about the 10/90 rule (spend 10% on the tool, 90% on people), it was obvious that the tool was way too robust and we didn’t have the resources to implement it correctly.

Seeing that our existing metrics could be done by just about any analytics tool, we made the switch to Google Analytics (FREE!!) a few months later. To maintain the data, I did A/B comparison (data reconciliation)  and looked at the percentage changes to accomodate the variation in data.

When you learn more about web analytics, you realize that the exact numbers aren’t important. There are many ways to calculate visitors based off cookies, session timeouts, and IP filtering. Different tools will give different data, but they are close enough. The real power lies in segmenting and trending.

Recently (Q4 2008), Google released some great improvements to their analytics tool that does custom reporting, advanced segmentation, keyword motion charts, and supposedly an API that I haven’t heard much about.

We also linked up our PPC campaigns  to our Adwords account so we can finally see how our paid search is performing. It was all lumped in organic search (free) prior, so we couldn’t receive any data on our paid traffic.

Another thing is putting on an on-exit survey to get customer feedback. For the short couple weeks it was on, we got over 800 people sending suggestions and complaints to make the site better. How cool is that!? It will go back on after we have made all of the improvements.

Throughout this whole learning process (which is still in its infancy), I really try to make more customer-centric choices when doing changes or making desicions. Doing the web analytics is only a small part of my job, but it truly is the key to seeing through your customers’ eyes. I can only pray that I continue to grow at the same pace throughout the rest of my career.