Google Conversion Room Asia-Pacific Blog Tips on tracking and improving conversions online

Measuring and optimising Fairmont’s social media efforts with Google Analytics

Thursday, November 25, 2010 | 10:08 AM

Labels: , , , ,

This is a guest post by Barbara Pezzi, Director of Web Analytics and Search Optimization at Fairmont Raffles Hotels International. Barbara joined the company in 2001 and was, until recently, Swissotel’s Director of Web Marketing. She is so passionate about web analytics that she has decided to make it her full time job. - Ed.

More and more marketers are engaging with their customers through social media channels such as Twitter. In this post, I share how I use Google Analytics to measure and optimise Fairmont’s Twitter activities. These best practices can also be applied to other social media channels, such as Facebook, Google Buzz, and email campaigns.

The problem
We use Fairmont’s Twitter account to increase awareness of our offers and drive traffic to our sites. While we think it’s effective for generating online buzz, we have trouble accurately identifying traffic to our website generated by our tweets. By default, traffic from Twitter’s web interface will be tracked as referral traffic in Google Analytics. If a user clicks on a link to your site from a tweet, you will see ‘’ in the Referring Sites report.

A lot of traffic generated by our tweets, however, does not originate from Many users of Twitter do not use the web interface, but instead use one of the many applications available such as a desktop client or mobile app. Additionally, the link we have posted on Twitter might be forwarded via email or SMS. In all these cases, any traffic which in theory should be attributed to our Twitter activities will be reported as direct or other referral traffic.

The solution: tagging our links
We solved this problem by using Google Analytics’ campaign tracking variables within the links that we post on Twitter. Campaign tracking variables allow us to tag our links so that Google Analytics can recognise and measure non-AdWords campaigns that brought visitors to our site. You can use these tags to track paid search and display campaigns, e-newsletters, social media campaigns, and more.

When campaign tracking variables are applied to any link in our tweets, the traffic resulting from those tweets will be correctly attributed to the respective tweet regardless of where the visitor found and clicked on the link.

Generating these tags is very simple using the conveniently provided URL Builder tool.

Tagging can result in lengthy links, which use up valuable characters in a 140-character tweet. You can keep the links short by using one of the many free URL shortening services such as or

Campaign tracking variables and URL shorteners in action
Recently we used Twitter to drive traffic to this URL - - to highlight Labor Day specials for our property in Boston. When we distributed this link over Twitter, we tagged it with Google Analytics’ campaign tracking variables so that we could easily identify our visitors who came from our “TwooseCOP” campaign.

When users click the link, they’re redirected by the shortening service to:

Analysing our social media traffic
Tagging our links takes care of the mechanics of tracking our Twitter traffic. Once we’ve set that up, we wait for the traffic to start flowing in, and then start analysing and taking action on our incoming data.

The tracking parameters will classify the above visits from Twitter in the “social media” medium and the campaign parameter will contain a unique identifier for this particular Twitter campaign. Since we use the same naming convention for all our social media posts, we can easily monitor the aggregate growth and traffic patterns of our social media efforts by viewing our traffic sources by “medium”

We then compare the quality and growth rate of the Twitter traffic generated by our own activities compared to the “organic” traffic (e.g. a guest tweeting about their experience at one of our properties) we receive from Twitter:

Additionally, since each Twitter campaign is uniquely tagged, we can review how each one performed, by first clicking on our chosen traffic channel, in this case “twitter / socialmedia”, and then by selecting “Campaign” in the dropdown box.

We are now able to view how many visits each campaign tweet generated as well as additional data such as “Bounce Rate” and the number or percentage of “New Visitors”. We can also see how these campaigns contributed to room bookings, revenue and conversion objectives by clicking on the “Goal Set 1” and “Ecommerce” tabs. This allows us to understand better what our visitors are interested in and which stories resonated the most.

Segmenting our traffic
We can also take this a step further with my all time favourite feature: advanced segments. You can find out more about advanced segments in my previous post on the subject.

First, I create a segment for our Twitter traffic (you can get a copy of my advanced segment here):

Then I apply this segment to any report I might be interested in. For example, I might wish to see what content these visitors consume when they land on our site:

Or even which hotel they booked:

Marketers are starting to generate significant traffic from Twitter. Just by adding campaign parameters, we are now able to understand how many clicks are coming from Twitter beyond those from Twitter’s web interface. More importantly, we can measure user behavior and conversion data. With those numbers in hand, we are a step closer to measuring the effectiveness of our social media campaigns and better evaluate where to spend our time and effort. And it’s not just about Twitter. You can use these tips on other social media platforms such as Google Buzz, Facebook, email newsletters, and more.


aprillins said...

I'll study more about this especially about unique segmentation and purchases data.. thank you

Paul Tyler said...

Thanks Barbara, great to see advice and information from a real world example. This has really piqued my interest and will be looking at this in more depth. Thanks again for sharing.

Unknown said...

Honestly Barbara, thank you for sharing your wisdom. It's going to take some practice for me, but conceptually everything makes sense. It's almost beautiful (I'm a geek, what I can say) in the way everything is explained in a logical manner. THANK YOU!

mwailes said...

This comment has been removed by the author.

mwailes said...

We are doing something very similar with sharing posts from our company's blog site. I built a custom url shortener that builds in the google tags.

Within the tool employees select the blog post that they want to share via social media (twitter, linkedin, facebook, and email).

The tool then generates the shortened url with the social media outlet named in the source tag and the employees name in the content tag. This has allowed us the additional segmentation of seeing how our employees' peers are interacting with our blog site.

The only problem is getting people to use it consistently.

Natasha McEachron said...

Thank you Barbara!

I've been trying to find a way to measure/analyze my social media efforts across different networks. I don't think this method will fully solve the issue for me but it should take care of the most important aspect. Thanks again for sharing!