Once a website benefits from visitors in some way, it’s time to worry about where those visitors are coming from. Doing so lets you:
Encourage sites that send you traffic, either by contacting them, sponsoring them, or inviting them to become an affiliate of some sort.
Advertise on sites that send you visitors who convert, since visitors they send your way are more likely to do what you want once they reach your site.
Measure affiliate referrals as part of an affiliate compensation program.
Understand the organic search terms people use to find you and adjust your marketing, positioning, and search engine optimization accordingly.
Verify that paid search results have a good return on investment.
Find the places where your customers, your competitors, and the Internet as a whole are talking about you so you can join the conversation.
The science of getting the right visitors to your site is a combination of Affiliate Marketing, Search Engine Marketing, and Search Engine Optimization, which are beyond the scope of this book.
When a web browser visits a website, it sends a request for a page. If the user linked to that page from elsewhere, the browser includes a referring URL. This lets you know who’s sending you traffic.
Note
The HTTP standard actually calls a referring URL a “referer,” which may have been a typo on the part of the standard’s authors. We’ll use the more common spelling of “referrer” here.
If you know the page that referred visitors, you can track those visits back to the site that sent them and see what’s driving them to you. Remember, however, that you need to look not only at who’s sending you visitors, but also at who’s sending you the ones that convert.
Referring URLs, once a mainstay of analytics, are becoming less common in web requests. JavaScript within web pages or from Flash plug-ins may not include it, and desktop clients may not preserve the referring URL. In other words, you don’t always know where visitors came from.
What to watch: Traffic volume by referring URL; goal conversion by URL.
An increasing number of visitors come to you from social networks. If your media site breaks a news story or offers popular content, social communities will often link to it. This includes not only social news aggregators like reddit or Digg, but also bloggers, comment threads, and sites like Twitter.
These traffic sources present their own challenges for monitoring.
The links that sent you visitors may appear in comment threads or transient conversations that you can’t link back to and examine because they’ve expired.
The traffic may come from desktop clients (Twitter users, for example, employ Tweetdeck and Twhirl to read messages without a web browser). These clients omit referring URLs, making visit sources hard to track.
The social network may require you to be a member, or at the very least, to log in to see the referring content.
The author of the link may not be affiliated with the operator of the social network, so you may have no recourse if you're misrepresented.
Most of the challenges of social network traffic come from difficulties in tracking down the original source of a message. Sometimes, you simply won’t see a referrer. But other times, you’ll still see the referring URL but will need to interpret it differently.
For example, a referral from www.reddit.com/new/ means that the link came from the list of new stories submitted to reddit. Over time, that link will move off the New Stories section of reddit, so you won’t be able to find the source submission there. But you do know that the referral was the result of a reddit submission. Other social network referrals contain similar clues:
Microblogging websites, such as Twitter, FriendFeed, or Identi.ca may tell you which person mentioned you, but also whether the inbound link came from someone’s own Twitter page (
/home) or a page they were reading that belongs to someone else.URL-shortening services, such as tinyurl, is.gd, bit.ly, or snipurl may hide some of the referral traffic. These usually rely on an HTTP redirect and won’t show up in the analytics data, but some providers such as bit.ly offer their own analytics.
Referrals from microblogging search show that a visitor learned about your site when searching Twitter feeds. Links from other microblog aggregation (such as hash tag sites) are signs that you’re a topic online.
Referrals from mail clients, such as Yahoo! Mail, Hotmail, or Gmail mean someone forwarded your URL via email.
Referrals from web-based chat clients, like meebo.com, are a sign that people are discussing your site in instant messages.
Homepage portal referral URLs can be confusing, and you need to look within the URL to understand the traffic source. For example, a URL ending in
/ig/came from an iGoogle home page;/reader/viewcame from Google Reader; and/notebook/came from Google Notebook. Some analytics packages break down these referrals automatically.
In other words, all referring sites aren’t equal. It’s not enough to differentiate and analyze referring sites by name; you have to determine the nature of the referrer. Different types of referring sites require different forms of analysis.
Since social networks and communities contain UGC, referrals may also show you when people aren’t just linking to you, but are instead presenting your content as their own without proper attribution. Tracking social network referrals is an important part of protecting your intellectual property.
People will often mention your content elsewhere but not link to you directly. This will generally result in users searching for the name of your site and the content in question, which will make referral traffic look less significant while overinflating the amount of search and direct traffic, particularly navigational search.
Community managers need to identify the source of the traffic so they can engage the people who brought them the attention and mitigate the inevitable comment battles. And site designers need to make sure that new visitors become returning visitors.
What to watch: Referring sites and tools by group; sudden changes in unique visitors and enrollments from those groups; search results from social aggregators and microblogs.
Knowing how visitors got to you doesn’t always tell the whole story. Sometimes the only way to get inside a visitor’s head is to ask her, using surveys and questions on the site. Such approaches are generally lumped into the broader category of voice of the customer (VOC).
Asking visitors what they think can yield surprising results. In the early days of travel sites, for example, site owners noticed an extremely high rate of abandonment just before paying for hotels. The sites tried many things to improve conversion—new page layouts, special offers, and so on, but it was only when they asked visitors, through pop-up surveys, why they were leaving that they realized the problem: many users were just checking room availability, with no intention of buying.
Use any travel site today and you’ll likely see messages encouraging visitors to sign up to be notified of specials or changes in availability. This is a direct result of the insights gained through VOC approaches.
What to watch: What were users trying to accomplish? Did they plan to make a purchase? Where did they first hear about the site? What other products or services are they considering? What demographic do they fit into?