Analytics Talk

In Part 1 of this series I explained link tagging, the technology that Google Analytics uses to track on-line marketing campaigns. In Part 2 I discussed how to tag your links and posted a tool that I use to quickly tag large numbers of marketing URLs. Today, in Part 3, I’ll start to pull this whole thing together by walking through a very basic analysis.

How I Start

I like to start my marketing analysis using the Marketing Campaign Results reports. Using these reports I can immediately identify any campaigns that are under or over performing. They’re a great launching pad for further analysis. You can find them in the Marketing Optimization > Marketing Campaign Results section.

The reports segment the data based on where the visitor came from using the values from the campaign tracking variables. So, for each of the major campaign variables we discussed in Part 1 (utm_campaign, utm_source and utm_medium) we have a corresponding Google Analytics report. Here’s the mapping of campaign variable to GA report:

Campaign Variable	Report Name
utm_campaign	Campaign Conversion
utm_source	Source Conversion
utm_medium	Medium Conversion

This means that the values you used in the campaign variables will be pulled directly into the reports. Exciting stuff, huh? :)

Campaign Conversion Report

Let’s start with the Campaign Conversion report.

This report segments the traffic based on campaign name. It contains information from tagged URLs (using the utm_campaign variable) and un-tagged URLs. How does it get data for the un-tagged URLs? If you’re using the auto-tagging feature in AdWords then Google Analytics will automatically pull in the Campaign names you create in AdWords. All other un-tagged URLs get put into the following buckets:

1. (direct): visitors that entered your website address directly into the browser
2. (organic): visitors from an un-paid search engine listing
3. (referral) : visitors that clicked on an un-tagged link
4. (not set) visitors from links that were tagged but were missing some information. For example, if you are looking at the Campaign Conversion report, and see that there were 10 visits from ‘(not set)’ this means that the utm_campaign variable was missing from the tagged link.

Ok, so what does this report tell us? It helps us quickly understand how well our campaign is performing using some basic metrics:

• Visits: How much interest did the campaign generate?
• Goal Conversion Rate (G1/Visits): Did the visitors from this campaign do what we wanted them to do?
• Transaction Average (T/Visits): How many transactions were generated by this campaign?
• Revenue per Visit ($/Visits): How much money did we make from each visit in the campaign?

It’s important to realize that each metric gives you a bit more insight into what is going on. For example, let’s say a campaign has a very low conversion rate. Why? Look at the number of visits. Is the campaign generating a lot of traffic? If there are a high number of visits but a low conversion rate there may be a disconnect between the marketing message you’re sending and the content the visitor sees when they land on the site. Dig deeper, Try checking the bounce rate for the landing page.

Again, this is a good starting point for a deeper analysis. And analysis means segmenting the data to gain more insight.

Segmenting Campaign Data

Notice that the first column in the report is named ‘Campaign/Source’ and not just ‘Campaign’? The reason is that this report let’s us drill down into our campaign and view the sources associated with the campaign. If we click on the plus sign for the ‘Ongoing’ campaign we can drill into the data and see the associated sources.

This tells us is there were three sources of traffic in the ‘Ongoing’ campaign: Squidoo, Wikipedia and MySpace. This is real data from a company using social networking and viral sites to drive traffic. The value in the brackets (Social_Networking and viral) is the medium (which we’ll get to later).

Remember from Part 1 of this series, the source is the ‘who’ part of the campaign. ‘Who’ did we partner with to distribute our message? By drilling down into the data we can find out. Drilling from the campaign level to the source level revealed a lot about our campaign. It looks like this business should dump MySpace and focus more on Squidoo! Not only do they get more traffic, they get far better conversion.

But let’s take a different look. It could be that these sources are used in multiple campaigns. Maybe MySpace did really well in a different campaign. Let’s use the Source Conversion Report to get a different view.

Source Conversion Report

The source report shows us how all of our various sources are doing. It does not matter which campaign the source belongs to, they are all listed in this report. This report is very helpful because it shows, historically, how well a source performs. It may be that a source just under performed for a specific campaign. We can see above that Squidoo, which performed very well in the Ongoing campaign, does not crack the top 10 sources. MySpace is no where to be found. This probably means that MySpace performs poorly across the board, not just in the Ongoing campaign.

Next to each source we can see the medium associated with that source. Again, I like to think of the medium as the mechanism that we use to push our marketing message out. Was it email, CPC, banner, print, etc. Google Analytics has pulled the medium value from the utm_medium variable and placed it in the report.

Medium Conversion Report

Looking at the medium we can evaluate how well the mechanism is working for us. Let’s see how well the Social Networking ‘mechanism’ is working.

Interesting. We can see that the ‘Social Networking’ medium doesn’t get a lot of traffic, but it gets attentive traffic (high number of pageviews per visit) and what appears to be an average conversion rate (for this site) for Goal 1.

The medium report is good at identifying dependencies. Are you too dependent on a particular way of getting traffic? If all your conversions come from organic, and the search engines drastically change their ranking algorithms, then you could loose a lot of traffic and a lot of money!

Other Reports

In addition to the Conversion reports, there are also three ROI reports. These reports are very similar to the Conversion reports. They segment the data in the same way (based on campaign, medium and source). The difference is the metrics reported. Rather than conversion rates, these reports show cost, revenue and ROI. If you have an e-commerce site and are collecting revenue, or have monetized the values of your goals, then the revenue generated by each campaign will be displayed.

A warning about this report. GA will only pull in cost data from your AdWords campaigns. Do not be alarmed if you see no other cost data in this report. GA is a closed system, you can not import cost data from other sources. This means that the ROI calculations will be incottrct for non-AdWords campaigns.

The referral conversion report is another fantastic report :) This report lists all of the un-tagged, non-organic and non-direct links that drove traffic to the site.

Drilling down into this report will show you where on each referral domain, the visitor originated. I find the referral report enlightening. The web is a wacky place. And people reference content in so many different ways. This report will help you hunt down all the sources of your traffic.

Some Final Thoughts

You may notice that some of the reports above have multiple lines for the same items. For example, the Medium Conversion report has two line items for social networking:

• Social_Networking
• social_networking

The reason the item is listed twice is that the person tagging the links specified two different values for the utm_medium variable. That’s why it’s important to use a standard naming convention when tagging your links.

Well, that wraps up this Part 3 of our Campaign Tracking series. What did you think? My ultimate goal is to make you all marketing measurement wizards. Am I doing a good job?