Posts Tagged ‘content grouping’

Apr 15
2013

leaking ecommerce funnel

Sometimes, the urls (and titles) of your pages are not conducive to web analytics reporting. For example, your ecommerce site’s payment, shipping, and order confirmation page may all have the same url for some reason – http://www.domain.com/checkout.aspx. To web analytics, all these funnel pages are reported as one page. You are now stuck, you can’t create ecommerce funnels and measuring shopping cart funnel abandonment is impossible. And there is a more subtle and serious issue as well, in your report, you may find hits (events, ecommerce, social, etc) associated with this url, but you won’t know what part of the order process these hits belong to.

checkout.aspx browser bar

Here’s the actual flow we want to track and understand:
ecommerce funnel

If you have a shipping calculator on your shipping page, a card type drop-down on the payment page, social buttons on all pages, you want to track each of these events on each page. However, it will show like this:
events-landing-page-bad-url

All the urls are the same! How do you know if these events happened on the shipping page, payment page or order confirmation page? You might be able to tell from the event names, but in some cases you may not be 100% sure, and this is definitely not clean and ideal.

While fixing the actual real urls and title tags (assigning unique urls and titles per page) would make things very organized, your content management system may not support this, or you might prefer not to spend that time or money on developers.

Luckily, there is a secret, undocumented method that allows you to actually set the page url and title of a visited page in Google Analytics. More importantly, it will actually associate the hits with these new, more meaningful page urls and titles. Your reports will be easier to read and will provide insights that may not have been available before.

The Issue With Virtual Pages

The traditional solution to this would be to use the _trackPageview method and trigger virtual pages for each “step” (i.e. /virtual-page/shipping.html, etc.).

_gaq.push(['_trackPageview', '/new-meaningful-url.html']);

The drawback here though is still, the actual events will not be associated with these virtual pages you’ve created. They will always be connected to the “real” page, which would be /checkout.aspx (as you can see in the screenshot above). You’re still lacking potentially valuable insights.

SECRET HACK! Setting the URL and TITLE in Google Analytics – _set method

With this new _set method, you can manually set the url and title of the page to whatever convenient name you want.

  _gaq.push(['_set', 'page', '/new-meaningful-url.html']);
  _gaq.push(['_set', 'title', 'New Meaningful Title']);

Simply push the new page’s title using _set method before calling your _trackPageview

<script type="text/javascript">

var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXXXXX-X']);
_gaq.push(['_setDomainName', 'e-nor.com']);
  _gaq.push(['_set', 'page', '/new-meaningful-url.html']);
  _gaq.push(['_set', 'title', 'New Meaningful Title']);
_gaq.push(['_trackPageview']);

(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>

pages-google-analytics

In the case of the ecommerce example mentioned earlier, for each page you’d like to rename, you can pass the preferred url and title so that it’s separated and meaningful in Google Analytics and associated WITH THE HITS:

_gaq.push(['_set', 'page', '/shipping.html']);
_gaq.push(['_set', 'title', 'shipping page']);

OR

_gaq.push(['_set', 'page', '/payment.html']);
_gaq.push(['_set', 'title', 'payment page']);

OR

_gaq.push(['_set', 'page', '/confirmation.html']);
_gaq.push(['_set', 'title', 'order confirmation page']);

events landing page good url

You’ll be able to see events and figure out things like “Did they abandon the cart at the payment page? At the shipping page? What are they clicking on the order confirmation page?” etc.

What do you think? Share what other use cases you might have in mind.

Jan 11
2010

It’s that time of the year to review 2009: “Top 5 Google Analytics Posts in 2009″! We want to thank our blog readers for their time, input and comments and we look forward to offering you more useful tips in 2010 and additional methods and strategies to leverage Google Analytics and take your marketing optimization efforts to the max. This post lists the top viewed Google Analytics blog posts, as well as a couple of bonus points related to measuring blogs.

Let’s get started!

  1. Content Grouping in Google Analytics: this is our top viewed post in 2009. Marketers loved it and techies loved it too :) , it showed you the what and the how. This very handy method allows you to categorize pages into groups of related content and collect these pages together and treat them as a single entity (for further analysis as a group). For example, if you have an online store with women and men clothing categories, you can use this technique to group the women pages as one “content group” and the men pages as another “content group” and then, as Avinash mentioned in his comment on this post, “content grouping can really help make a complex site much easier to understand from a macro perspective”. You can also, apply the same content grouping concept to brand pages, or to a groups of landing pages.
  2. Monetize your SEO effort by Leveraging Google Analytics: this was one of my posts and it had to be marketing & analysis focused, since I am not the javascript guy :) . If you are running a Search Engine Optimization (SEO) program, you’d want to take a few minutes and read this post, if you haven’t already. The post uses a case study and real numbers to help you answer questions on how ranking, or lack thereof, impact the bottom line, and help you get decision makers to act and get the most out of your SEO program.
  3. Tracking Press Releases in Google Analytics: again, another method to enhance your measurement system and get a better sense of how your marketing initiatives are performing. Granted Press Releases fall under the “branding/awareness” marketing category, and we don’t just measure branding/awareness by immediate visits/outcomes, it’s still nice to have performance data for each press release. Check out this method, some coding is involved, but the implementation is detailed for you.
  4. The Cost of Misinformation: a popular post addressing the mis-information (by some fee-based web analytics vendors) about Google Analytics. In additional to the advanced capabilities and enterprise-level features that Google Analytics has been introducing, this post highlighted Google’s innovative, open and global eco-system for support, training and consulting, available around the globe by some of the brightest in the industry.
  5. Problems with Bounce Rates: this post was an answer to a lot of questions we get on how to correctly read and analyze one of the most useful metrics, the bounce rate. Hint: look at your top landing pages report.

And some from 2008!
And since we are talking about top posts, here are three posts that were published in 2008 but continue to be very popular, check them out and put them to use!

And another bonus – E-Nor’s Guest Posts on the official Google Analytics Blog

In addition to the top posts on the E-Nor site, here are few posts that were well received (based on the limited qualitative data we have) on the GA blog:

A couple of notes on measuring blogs and posts

Note #1- Normalize your data

If you really want to measure the most popular post in a year, the aggregate data might not tell the entire story. A post that was published in January will have a whole lot more time to get traffic/visits/comments/feed subscriptions/retweets than a blog that is published in December. This reminds me of what Malcolm Gladwell describes in his book “Outliers – The Story of Success” and how Canadian hockey players born early in the year all have a huge advantage and how this advantage compounded over time (he showed the stats and the numbers to back up his findings). So if you truly want to compare how each post did, you might want to normalize the data, add a weighing factor to compensate for the sequence of the month in the year, or simply measure stats for each post in X weeks after it has been posted.

Additionally, and for the visually inclined, you can use a Google Analytics’ Motion Chart to “play” the graph over time and watch how each post did and compare the various metrics concurrently over the span of the year.

E-Nor_Blog_KPI_Motion_Chart

For example, the chart above represents a number of blog posts (from the GA Top Content report) along with few metrics. The x-axis represents pageviews; y-axis: average time on site; size of the bubble represents $index, and each color represents a specific post.

You see how the blog post represented in dark blue behaved differently than the post represented in lighter blue. For example, you see a “big bubble” on the right hand side of the graph, ~550 unique pageviews with a relatively larger $index value, both are positive outcomes compared to other posts. One can then do a bit more digging and find out what led to this positive result and repeat it!

Note #2: Blog Engagement Metrics

When it comes to blogs, you don’t just want to measure visits & pageviews (that is so 2009! :) ), you want to have more meaningful metrics. Who cares if you are pumping out posts like there is no tomorrow and no one is engaged. I’d look for things like feed subscriber rates, comments per post, words per comment, posts per blogger, among other things.

Here are a couple of snapshots from two bloggers that are active authors on the E-Nor blog, you’ll notice completely different patterns and user interaction.

A couple of observations

  • Blogger A is more active in blogging 1.2 posts per month compared with 0.6 per month for Blogger B
  • Blogger A gets fewer comments, 0.8 per post while Blogger B gets 8.4 per post
  • One conclusion is that while Blogger A can write, his posts are not as engaging (ouch!) but Blogger B has a knack for getting people’s attention and input. Both bloggers can learn from this quick analysis and improve their posts in 2010 (Blogger A do something to get your audience attention, and Blogger B, charm us with more posts).

I hope you have found our 2009 posts useful! We’d love to hear from you for ideas, issues, questions and areas you like us to address in 2010. Leave us a comment below or email us directly at info @ e-nor.com

Related Posts

Mar 06
2009

The executive team at E-Nor is quite greedy! Provide them with a neat trick and instead of thanking you, they ask for more! :)

A few days after my colleague and I wrote about content grouping in Google Analytics, E-Nor president Feras Alhlou asked if it is possible to apply the same concept to referring sites.

Our objective is to group all domains and subdomains of related referring sites as one referring entity. For example, nextag.com, nextag.co.uk, and affiliates.nextag.com should appear as a single Nextag entity.

1) Create an advanced filter that renames all domains and subdomains of a particular referring site to one entity.

2) Repeat step 1 for every group of referring sites that send you significant traffic.

Another example of related sites: cnet.com, zdnet.com, download.com, and shopper.com. Affiliates and dealers could also be grouped this way.

3) Apply the filters you just created to a new profile.

New profile – my colleague Rehan Asif cannot stress this enough!

Congratulation, we have grouped related referring sites as entities!
Now we can look at the traffic from those referring sites at an aggregate level.
Happy, Feras? :)

Stay tuned for the next post on how to group pages based on their functionality. It is actually Avinash’s idea from the previous content grouping post and I promised him that I will write about it.

Finally, do not forget to adjust your clocks this coming Sunday and analyze your performance before and after the change :)

From now until the next blog post, I wish you a happy March and an enjoyable month of analysis :)

Jan 30
2009

The other day I was doing my daily reading and I came across the following paragraph: “Analytics people who like to cull patterns from massive amounts of data like to aggregate rather than split data. In web analytics this means treating several pages as one unit in order to know about visits that saw one or more of a certain set of pages that the analyst thinks belong together. In WebTrends and other software this is done with “content grouping” and Google has no parallel to it.” Chris Grant, Got Analytics? blog

“Google has no parallel to it!!” I have to admit that I took this statement personally as I consider Google Analytics my baby. :) So I went to my colleague, Rehan Asif, to discuss this and in less than twenty minutes we came up with the following concept:

  • Categorize pages into groups of related content.
  • Collect these pages together on one page and treat them as a single entity.
  • Specify the URLs that you want to include in each group by defining URL patterns.
  • Create a filter for each group.  Each filter will search for the group identifier and replace the entire URL with a new URL.

Here is a real example on an online shoe store where we want to take all pages that focus on specific brands (for example, Converse, Timberland, Vans, and Reebok) and treat them as one content group.

1) First, we studied the URLs and found that they contain the brand name.

http://www.domain.com/authentic-vans-shoes-satain-blackpink.html
http://www.domain.com/adidas-bg-superstar-whtblk.html
http://www.domain.com/puma-big-kids-drift-cat-jr-blkwht.html

2) Using an advanced filter, all pages with “vans” in their URL will be renamed to “/vans.html”

3) Now create filters for each brand and apply the filters to a new profile called “Content Groups”

4) Now we have created content groups that allow us look at all pages for any brand as a single entity. We can now study the links where people exit, the entrance keywords, the entrance sources, other pages they visit on the site, and more.

Now, as I like to say, the real analysis begins! :)