Understanding Google Analytics Multi-Channel Funnels
Goals - What does Google Analytics measure?
Beyond measuring pageviews, visits, visitors, traffic sources, campaigns, devices, countries, etc. Google Analytics allows us to measure actions. What do we mean by actions? We consider it to be Goals, Transactions and Events.
Google Analytics Multi Channel Funnels (MCF)
As we stated in a previous article, at the end of august 2011, Google Analytics announced the Multi-Channel Funnels reporting suite a “...set of five new reports in Google Analytics gives marketers insight into the full path to conversion over a 30 day period, not simply the last click”. This is where the future of conversion tracking is heading to.
Multi Channel Funnels Conversions
The Multi Channel Funnels Reporting uses Goals and Ecommerce transactions. So either you need to have set up the Ecommerce tracking on your site or to have set up at least one Goal, to be able to have the Multi Channels Funnels reports to work for you. If you have both Ecommerce and Goals activated, you’ll be able to select in a drop down for which Goal or Transaction you want the stats to be displayed.
Nevertheless there is an important difference the way MCF reporting attributes the conversions to the different sources of traffic.
Conversion Attribution, Last Click and Direct Traffic Exception
In Google Analytics, as well as in many other web analytics or advertising measuring tools, conversions are attributed to the last click. There is an exception within Google Analytics though: Direct traffic. Direct traffic is visits from people typing directly the url of the website into their web browsers or people having the website in their bookmarks. Note that Microsoft Outlook traffic will be considered as Direct traffic (if not tagged correctly). In that particular case, the conversion will be attributed to the previous source of traffic. Let’s illustrate that by three short examples:
1. I’m reading the review of a laptop on a blog. There is a link to the product page on Amazon. I click on the link, land to Amazon website, check the product price and availability. Then later in the day I come back to amazon by typing directly the url in my web browser and I order the laptop.
Google Analytics will attribute the conversion to the blog.
2. Let’s say that I heard speaking about ultrabooks at work. Coming back home, I do a search on Google and click on an AdWords ad of Amazon. Then later in the evening, I read a product review on a blog that link to Amazon. I click on the link, land on Amazon website and check the product price and availability. Finally at the end of the evening I come back to amazon by typing directly the url in my web browser and I order the laptop. Again Google Analytics will attribute the conversion to the blog.
3. Let’s now say that I heard speaking about ultrabooks at work. Coming back home, I go on Amazon website by typing directly the url in my web browser and I order the laptop. Google Analytics will attribute the conversion to Direct traffic.
Multi Channel Funnels and the Last Click Conversion attribution
So far, Goals and Transactions attribution has been based on the last click with the exception for Direct traffic (see above). The Multi Channel Funnels are now attributing conversions slightly differently. MCF does not take the Direct exception into consideration. MCF does attribute the conversion to the last traffic source.
Google Analytics Conversions Attribution: MCF vs Ecommerce Report
Here’s a real example that will illustrate how we should be careful when it comes to interpret the data. Hereunder the breakdown of the Ecommerce transactions by medium:
Now let’s look at the stats for the MCF:
As you can see above, conversions per medium breakdown is quite different between the two reports:
cpc (paid search) from 9 to 4,
Organic (organic search) from 3 to 2,
Direct (none) from 2 to 8.
Top Conversion Paths
In order to have a better understanding of what is actually happens, we need to have a closer look and to check the Top Conversion Paths report:
How to interpret the above Top Conversion Paths report?
First we notice that there were 12 different conversion paths. Conversion Paths report shows all of the unique conversion paths (i.e. sequences of channel interactions) that led to conversions.
MCF Conversions Attribution
If we look at the last interactions, the paths 2, 3, 4, 5, 6, 10, 11 and 12 had Direct as last interaction (= 8 attributed conversions). The paths 1 (with 3 conversions) and 8 had Paid Search as last interaction (= 4 attributed conversions). The paths 7 and 9 had Organic Search as last interaction (= 2 attributed conversions).
Ecommerce Conversions Attribution
If now we consider the Direct exception, only the paths 11 and 12 had Direct as last interaction (= 2 attributed conversions). The paths 1 to 6 and 8 had Paid Search as last interaction (= 9 attributed conversions). The paths 7, 9 and 10 had Organic Search as last interaction (= 3 attributed conversions).
Towards a better understanding
The Multi Channel Funnels reports open new possibilities and a better understanding at your client journey to a conversion. Keeping the last click model in mind, one of the challenges will be to better attribute the revenue to the channel that generated it. This is particularly important for traffic acquisition sources that we need to pay for such as Banners, Paid Search, Affiliates, etc.
For instance, in the above example, the MCF model gave us 4 conversions for AdWords (Paid Search, cpc) while the Ecommerce model gave us 9 conversions. Let’s say that we invested 1.000€. The MCF model gives a cost per conversion of 250€. On the other hand, the Ecommerce model gives a cost per conversion of 111€. That’s big enough to consider it twice!
Written by Nicolas Debray
Tagcloudwebsite iphone viral images video privacy facebook business linkbuilding news search engines blog indonesia tools microsoft russia bing research london search engine strategies spain youtube adwords google tools matt cutts baidu android social ses gmail europe advertising a4uexpo sea ppc sem search engine yahoo yandex analytics