Traffic—it’s something every company and inbound marketer lives for. For us, every site visitor is a potential business lead, which, one day, could translate into real business.
The problem is traffic is also anonymous. A web visitor doesn't say who they are, only where they come from, and—as you'll discover in this article—even that data can be confusing.
Today, we'll take a deep dive into one of the most confusing areas of website analytics: direct traffic. It's a subset of your website visitors that should be crystal clear—those who directly and intentionally visit your website, but it isn't that simple. Because of several unsolved technical problems, most websites today will see spikes in direct traffic or consistently high direct traffic metrics, and more marketers can't explain why.
In this article, I'll provide some insights on how to interpret direct traffic spikes and what to do to avoid unusual direct traffic metrics in the future.
Google Analytics breaks down website traffic into 5 major areas: organic search, direct, referral, social, and paid (Pay-per-click) traffic. Similarly, HubSpot shows you these 5 main types of traffic, plus the addition of traffic from email marketing. These can then be broken down into various sub components. Social media, for instance, can be split out into the main networks: Facebook, LinkedIn, Twitter, Google+, and YouTube, as well as other less trafficked online communities.
One way that Google, HubSpot, and other traffic analytic platforms know where you are coming from is using tags at the end of a URL, known as UTM parameters. If you have ever clicked on a link in a tweet that is tracked you may notice that the URL contains a string of extra text at the end of it. This is the tag that signals where you are coming from. By dissecting the UTM string you (or your analytics platform) can discover where the original source of that link was.
In the image below you can see the UTM tag for a Twitter link. Notice you can see that the medium was social, the source was Twitter, and the campaign was buffer—a social media automation tool used to post the tweet. All these elements of the UTM parameters help your analytics software understand where traffic is coming from.
In most situations, direct traffic is any type of traffic that is not tracked with a known source. For instance, if you type in www.wikipedia.org into your browser address bar that would be considered direct traffic, but if instead you search for Wikipedia on Google and click on the search result that would be considered Organic Search.
Now that we know about the different types of traffic and how they are tracked, we can dig into the issue with large spikes in direct traffic.
Because direct traffic is basically anything not tracked, it can be difficult to determine its true source. In other words, "direct traffic" is a misnomer because it's more of a catchall than an accurate analytic of intention. However, there are a few ways we can dig into the data to determine where the traffic is coming from.
Dark Social was first coined in 2012 by Alexis C. Madrigal in an article written for The Atlantic titled: "Dark Social: We Have The Whole History of the Web Wrong." The term “Dark Social” refers to traffic from social communities that is not tracked as being from social media.
Consider this. How often do you send someone a link you found in social media over a text message or in an email instead of sharing it with them on Facebook or Twitter? Because that link does not contain referral data it is lumped in with the rest of the direct traffic. This traffic is called Dark Social. It comes from a social source, but you do not know it.
Then, there is the issue of people sharing the links by copying the URL directly from their browser bar with the UTM parameters attached. Say someone finds a site via Twitter. They then share this URL with a friend via email and instead of their friend showing up as email traffic it groups them in with Twitter visitors. While it doesn't contribute to direct traffic numbers, it's a definite misattribution to social traffic.
A related source to Dark Social is mobile app traffic. More of our time—greater than 50% now according to some sources—is being spent online via mobile devices, and most of that time is not in browsers but in applications. Facebook, the biggest social platform of them all, has been shown to not always correctly attach referral information to track a visit when coming from their mobile app.
The increase in use of SSL—i.e. secured websites—has also become an issue. SSL is set to increase in usage as online security becomes more of a concern. Google is even starting to use SSL as an indicator to boost a site page ranking. When someone visits a secured link from a non-secure site (or vice versa), it strips the visit of all referral data, lumping them in with the rest of the direct traffic.
Generally, links from personally sent emails will be tracked as direct links. The same goes for emails from most automation platforms, except they're usually set to track by default. These settings add the necessary UTM parameters to the links in emails and give you accurate reporting of what emails are generating traffic. So, your marketing emails shouldn't cause mystery direct traffic. However, for emails sent by employees or by others, those links will likely be tracked as direct.
How do you tell if the direct traffic is from one of these problematic areas—dark social, mobile apps, or secure-non-secure traffic? One major method is to examine your social media traffic and see if there is a correlating spike in that traffic that hints at dark social traffic or irregular mobile activity. If there is, it's likely your website is experiencing some bleed over into your direct traffic as well.
Also, take a look at what pages your direct traffic is landing on. Are they coming to an easily typable URL like your homepage, or are they arriving on a blog post with a long keyword title that no one is going to type in? If it's the latter that can also probably be attributed to Dark Social.
Unfortunately, current problems in tracking secure-non-secure traffic and email-based traffic aren't as easily solved for. However, as more websites move to SSL, most marketers hope that the problem will gradually subside. And for email, as long as most of your marketing emails are tracked as such, you shouldn't worry too much about other email-based traffic infiltrating your direct traffic numbers—unless you have a highly successful forward chain.
A major source of traffic, particularly for large or remote workforces, can be your company's own employees. Because they are so familiar with your website they may simply type in the main URL or have a saved bookmark of the homepage in their browser. You don't want this kind of activity to disrupt your direct traffic metrics—or, for that matter, any traffic subset.
A tool to fight against recording this type of self-inflicted traffic is by using IP filter tools. By creating a list or range of domains that are known to be used by employees, such as the corporate office’s IP address, you can discard these visits and just record those from potential customers. IP filtering is available in Google Analytics as well as most marketing automation platforms, such as HubSpot.
Creating this IP list can be a challenge. Many internet service providers use dynamic IP addresses, meaning that the address of employees will change over time.
Larger companies tend to buy static IP addresses, which do not change, or a small range of addresses that can be filtered out, for their primary interest connection. Of course, this still leaves the issue of workers logging on after-hours from home or remote workers and teams collaborating from all over a country or the world. Unless you opt for total control and constant maintenance, most companies have to live with imperfect traffic data.
For situations with dynamic IP addresses and remote logins, it is best to do periodic updates to your IP list to help filter out the known addresses. It’s not a perfect solution, but it will help generate overall better traffic data.
At the end of the day, knowing exactly where every site visit came from is probably an impossibility. However, understanding your traffic and where it likely came from will help you optimize your site and content offerings to bring in the best traffic possible.
For marketers, the real question is: Is traffic a good indicator of my marketing activities' performance? In inbound marketing, we tend to advocate for less attention to traffic numbers and more focus on creating conversions—real contacts with whom you can develop a relationship. That doesn't mean traffic numbers aren't important to listen to, but having perfect data—especially when perfect data isn't possible—should not by a marketing analyst's priority.
Instead, opt for building a strong inbound marketing funnel, where traffic is converted into contacts, and contacts become real leads who your marketing team can move toward the sales process.
Topics: Website Design
