You probably have inaccurate Traffic Sources in your Analytics. Discover why.
A few days ago, my last article was featured on GrowthHackers.com. It generated valuable traffic and helped promote the article.
Several days later, I logged into Google Analytics for a routine check. I couldn’t see any referral traffic from GH.com, although it was clear that the traffic came from there.
Why didn’t Google Analytics track the referrals and show the correct traffic?
Chances are, Google Analytics doesn’t correctly attribute your traffic source and shows more direct traffic than there is. Keep on reading to know why.
Sometimes I see websites with a lot of direct traffic. In some cases, it accounts for more than 70% of the total traffic. How is that possible?
Do you believe it’s possible that 70% of visitors are directly typing a URL in their browser to reach your website? It depends on what you do, but it’s highly unlikely.
What happened?
I couldn’t see any traffic from GH.com. All the traffic was attributed to the “Direct” channel.
According to Google Analytics, the link never existed. I knew it was sending traffic. I could see it coming in real time.
Why it Matters?
The first issue is related to conversion. If you don’t know where visitors are coming from, you can’t tell which channels are effective and which ones aren’t.
The second issue is that you have no clue about who is linking to your content, which is very valuable. Maybe you want to thank them?
Getting the right sources of traffic is primary to understanding your visitors’ behaviors. Without clear data, your analysis will be void of depth.
Explanation
According to the RFC 2616:
Clients SHOULD NOT include a Referrer header field in a (non-secure) HTTP request if the referral page was transferred with a secure protocol.
In simple words: if you go from a secure (HTTPS) website to a website without SSL, browsers strip referral data.
Therefore, when visitors were coming from GH.com, a secure website, to my website, browsers weren’t sending any referral data.
Since browsers were removing the referral link, Google Analytics wasn’t able to collect the appropriate data.
Solution
Installing an SSL seems to be the only solution to get browsers to send the accurate data.
Since SSL also has a positive impact on SEO rankings, it can be a good idea to install one when you’re getting started.
SSLs are now quite cheap. With this in mind, the decision is a no-brainer.
Conclusion
TL;DR: Browsers remove referral data when being transferred from a secure to an insecure website. Installing an SSL can therefore ensure the appropriate attribution of all traffic sources.
I never expected to discover this. I was actually very surprised, especially because I had never heard about anything like this before.
Installing an SSL on each new website now makes a lot of sense. More and more websites are going to use them. If you don’t, you’ll soon be in the dark about those sending you traffic.
Overall, it shows that technical issues can affect Marketing and Business. This type of “tricky” answer can even make it hard to find out what’s happening.
Ah, that explains it! Buying an SSL now – thanks Pierre 🙂
You’re welcome Ryan! Just note that you’ll need to change your inbound links for correct attribution (a simple redirection doesn’t work).
Pierre, What do you mean by changing “inbounds links for correct attribution”?
If you install an SSL, you’ll need to change the inbound links to HTTPS if you want to see them attributed as referral.
Old links (non-HTTPS) will still be attributed as Direct Traffic, even with a 301 redirection.
Ah! Makes sense. Thanks for the clarification.
Analytics used to be great, but ever since Google started to allow people to opt out of stats being collected, the data set is very flawed. There are ways to extrapolate out additional data, but I recommend also look are packages like AWStats which although are not as comprehensive, do run straight from the server logs. This has the benefit of a complete set of stats… But this is also not 100% accurate so I normally take Analytics and compare with AWStats to get a composite data set…
Yeah. It’s getting harder and harder I guess. The main issue with server logs is that you have plenty of missing information about your users. We’ll probably need to look at both in the long-run…
Finally I understand!!! Thanks
Brilliant find! Just ran into a situation just like this, that started over a year ago. Thanks for sharing these insights Pierre.
Glad it helped! Hope you were able to solve your problem.
I’m having this problem at the moment with a site I’m working on. I’ll get an SSL added – but it’s a relatively new site so I’m not sure what could be referring that much traffic! We’ll see thanks for the post.
Hi Simon, getting an SSL should help in any way since Google started to recommend using them everywhere. You’ll also need to change inbound links from your HTTP to HTTPS website. Hope it helps!