5 Web Analytics New Year Resolutions

It’s that time of year again. Well-meaning people all over the world are proclaiming that they will quit smoking, drink less, eat fewer bacon sandwiches and of course, make more of an effort to analyze the performance of their websites. I am afraid I can’t help anyone with the first three having repeatedly failed with two of them myself. However, people who fall into the latter category have come to the right place.

To take the pain out of the process for you I have compiled our top five web analytics resolutions for 2010. These were based partly on some of the recurring analytics sins we see over and over and partly on what I feel can be underused or undervalued features. They all have the common theme of making your website data more useful. I make no apologies for the heavy bias towards Google Analytics here; it’s what we use most often, and for a reason. However, most of the below features are available to some extent in most serious web analytics packages.

Fix site search (and use it)

Tracking user searches on your website is the closest most marketers will get to having a conversation with their customers. If you have a site search function on your site, it will be getting use from your visitors, and probably more than you realize. Omniture, Webtrends and Nedstats to name but a few all support this. In Google Analytics, I’ve always felt access to this data is hidden away unnecessarily and this actually limits the number of users who set this up and are actively using it. In most cases it’s incredibly easy to activate. If your website search function doesn’t transmit a search term to the url in a Google Analytics friendly fashion, fear not. You can fix this by creating a virtual page view with the page URL “q=keyword” where “keyword” is the search term taken from the source code of the page.

For starters, take a look at some of the below once you’ve got the data coming through:

• Which pages are most searched from?
• Which pages do most searches land on?
• Which search terms are the most popular?
• Which of these most often lead to exits from the site?

Clean up content reports

This is another simple trick but is nevertheless crucial to getting accurate and useful content reports. Many (particularly ecommerce) websites have multiple session IDs which automatically append to the page URL as users browse through the site. By default Google Analytics tracks these as page entries. The result is often thousands of pages with single visits, essentially rendering your content reports useless. If you haven’t done this before it’s easy-peasy, just go to the settings page and look for “Exclude URL Query Parameters.” Compile a list of all the session IDs you’ve seen in your content report and put them in there.

Tag marketing initiatives (and then analyse the results)

Web analytics can be clever, really clever. So often though, you have to put the effort in proactively to reap the rewards later. In order to track the performance of your various marketing efforts you have to first tag them. That means giving email campaigns meaningful names and dates and tagging all paid search and ad display channels individually. Web analytics tools have different ways of doing this. Google provides a handy tool meaning there really is no excuse for not doing this. We marketers and web analysts have to constantly prove our worth. Here’s where we can prove and improve the return on investment we provide for our businesses.

Annotate reports and charts

It sometimes seems that at least half of the time we spend on web analytics involves working out what happened to make a certain trend chart shoot up or down. I’m going to admit that occasionally, this has to be carried out a second time several months down the line to refresh the old memory banks. With the new Google Analytics annotation feature you can add notes to any event or interesting trend shift. This can be done pro- or retro-actively and by any member of your team with access to a profile. Now, use your new found spare time to play with the next tip.

Segment, segment, segment!

Borrowing from political speech writing 101, I wrote “segment” three times. The intention was to give it more gravity and that’s because I really think this should be high on the lists of all DIY web analysts for 2010. Segmentation is simply selecting individual portions of your data to look at separately. If it helps, think of your data as a big juicy steak; it would perhaps be more efficient and quicker to just shove it down your throat in one. As we all know however, steaks are more pleasurable (and safer) if we chop them up into little portions which we then eat one-by-one. With that in mind and with Google doing an excellent job of providing easy (free) access to segmentation there really is no excuse not to spend the whole year dicing your data up into little, more palatable portions. It really does digest easier that way. For some excellent segmentation ideas check Avinash Kaushik’s post on segmentation or you could even use advanced segments to convert more new users

All the best for 2010 and happy segmenting!