← Go back

How I Use Google Analytics

18 Sep 2013

Google Analytics has a massive number of features, but I found that in my day-to-day usage, I don’t really touch a lot of the functionality.

Especially after I’ve started applying Lean Startup Principles and measuring progress with more actionable metrics like A/B Tests or Cohort-Analysis, I spent a bit less time with GA and more time with analyzing user-data in my database or setting up A/B tests.

So, my GA usage pattern became more kind of a “Monitoring and Disovery”, i.e. to stay aware of any unsual changes in visits or conversion rates, and to gain some insights through GA data, e.g. by analyzing Keywords by which people found my site.

Below an overview of the functionality I use from GA. The metrics and functionality you need for your business will vary, but I think the following points are a good starting point for every web business.

Goals


If you haven’t yet setup goals, you should do it now. Goals not only let you measure how many of your visitors reached a specific page (e.g. ‘successfully purchased product’), but also visualize the visitor flow accross the different stages of the funnel.

In practice, things can become a bit complicated if you want to have clean and consistent data for each part of the funnel. An example from my site: users can request quotes from different handymen by publishing a job description. The funnel is as following:

start page → job description form → registration form → email confirmation page → job posted page (goal).

But users can also first register and then post a job when they are logged in. The funnel looks for this case a bit different (no registration form), but it has the same goal.

The problem here is that logged in users don’t visit the registration form, thereby skewing the funnel metrics. There’s even a third case, but I leave it out to make things not more complicated.

To have precise and consistent data, I created a separate funnel for new users only (= users who need to go trough the registration form) with following destination pages:

/jobs/new$
/users/new_employer\?pageflow=postjob
/users/confirm_registration\?pageflow=confirmsignupemployerpostjob

The $ sign in the first destination is a regular expression that excludes any URL with characters after ‘/jobs/new’. Since on my site, logged in users are redirected to ‘/jobs/new?pageflow=isloggedin’ when they want to visit /jobs/new, I can exclude visits from logged in users.

It sounds like a lot of effort just to define a funnel, but the use of regular expressions is just one option. The main goal is to have unique URLs for a specific funnel, and either you define them explicitely in your application, or you use regular expressions (or even both).

If you don’t want to build your own regular expressions, check out the Regexinator from Lunar Metrics.

Custom Segments

Often, you are more intersted in a specific group of the traffic, e.g. all users who visited your page for the first time. GA lets you define Custom Segments for all kind of criterias.

I’ve created a Custom Segment for longtail traffic (search terms with let’s say three or more keywords) by defining the following regular expression for the Keyword dimension (via vkistudios.com):

^\s*[^\s]+(\s+[^\s]+){2,}\s*$

You can then view all Google Analytics data through the lenses of your Custom Segement: Organic Search, User Behavior, Mobile, Browser, etc.

Campaigns


Also a pretty neat option is “Campaigns”. It let’s you define specific parameters that you can add to an URL, e.g. a Link in your Newsletter Campaign. Below an example link with the “utm” Campaign parameters (you can also create them with this tool):

www.mysite.com?utm_source=newsletter&utm_medium=email&utm_content=textlink&utm_campaign=spring_sale

Every visitor who clicks on this link is then counted as visitor under the Campaign ‘spring_sale’.

For my site, I use Custom Campaigns not only for newsletters, but also for transaction mails (e.g. notifications that are sent out to handymens whenever there’s a new job published on my site).

Campaigns provide for me an easy way to separate and aggregate traffic under a custom category instead of just having them list
ed under “Referral” or “Direct”.

Also good to know is that traffic from links with campaign parameters are excluded from the source type “Referral”.

Site Speed


Keeping an eye on your site speed is never wrong. According to a study of gomez.com and akamai.com, 40% of the users abandon a website that takes more than 3 seconds to load.

I’ve had a few very inefficient database queries for my site which took up to 8 seconds. I fixed them in the meantime, but there’s always something that pops up here, so I added an E-Mail-Notification (under ‘Intelligence Events’) if the loading time exceeds a certain value to keep me alerted.

I also recently found that GA has a tool PageSpeed Insights (under Speed Suggestions) which gives you a pretty decent report of what is wrong and what could be improved. You can actually test any site you want.

Custom Report

Sometimes it’s hard to get exactly that metrics out of GA you want. For example, I have a directory section where all the handymans of my site are listed. I wanted to monitor the Unique Visitors for this section. The regular Content Section on GA gave me only the ‘Visits’ which is not the same.

But the Custom Report let me compose exactly that combination of metrics (Visitor, Bounce Rate, …) and dimensions (Page, Keyword, …) I needed.

Funnel Visualisation


The visualisation is a great way to quickly identify problematic parts in the funnel. I usually check on which page users exit the funnel. This gives me hint about the reason users didn’t complete the process.

For example, if you have a signup funnel and a big chunk of users end up on the pricing page, you might not doing a good job of describing your pricing model (or you are too expensive). I would then run an A/B test with an alternative version of the concerned pages that does (hopefully) a better job.

If you want to dive deeper into funnels, there’s also an in-depth post about them from KISSmetrics.

Organic Search Traffic


The organic search can show you by which Keywords users found your site, which is particularly useful in combination with Goals. Given you have set them up, you can browse through all keywords and their conversion rates. There’s hardly a better way to do Content Marketing than knowing which Keywords convert.

Unfortunately, Google rolled out on the 18. October 2011 a secure search as the default setting for authenticated Google Account users. This means that for logged in users, the Keyword data is not available, and instead, the term “(not provided)” is displayed.

The percentage of (not provided) is steadily rising, I counted on my site for the last 30 days 59% for (not provided). There’s currently no sign that restriction will be reverted. A few SEO’s came up with a some workarounds here, here and here.

Event Tracking



There are actions for which no URL exists (e.g. modal views, watch a video, etc.), so they don’t show up in the Content section. To track them anyway, GA provides a method called Event Tracking.

Below a simple example from GA Guide how to use Event Tracking to track a user interaction with a video Play link:

<a href="#" onClick="_gaq.push(['_trackEvent', 'Videos', 'Play', 'Baby\'s First Birthday']);">Play</a>

Whenever this link is clicked, GA registers an event under the Category ‘Video’ with the Action ‘Play’ and the Label ’Baby’s First Birthday’.

So far, I’ve applied Event Tracking only in a few cases, mainly to validate demand in a lean way. In each case, I added the event method to a button/link, and after people clicked on it, I knew it’s ok to invest some more time in the corresponding feature.

There’s definitely a lot more GA has to offer, but I haven’t found a use case yet for them, and the current metrics I monitor give me plenty of work.

View Comments
blog comments powered by Disqus