Wednesday, July 27, 2016

Corrupted Data: How to Detect and Remove Referral Spam

There's a rampant issue impacting the Google Analytics account of nearly every business. It's called referral spam (a.k.a. ghost spam) and any marketer tasked with reporting online metrics should know how to detect and remove it from their data. Using Google Analytics to make data-driven decisions is important for any small business, but if you're only looking at the top level metrics, then you may not be getting an accurate picture of how your website is truly performing. 

What is referral spam?
Referral spam is a black hat SEO tactic, in which bots are used to send send multiple fake visits to your website. These visits are recorded in your data and falsely inflate your site session metric. What do spammers get out of this? Extra visits to their website. Every curious data collector who clicks on a link automatically provides traffic to the offending site. Sometime this means they can charge more for advertising on their site due to the falsely inflated site visitation, and other times they can serve up Malware so be cautious about clicking on unfamiliar links in your data!

How to spot referral spam?
The majority of these top 10 referral sources are spam. 
1. Login to your Google analytics account and navigate to the reporting tab.

2. On the left hand menu click on Acquisition.

3. Then select All Traffic.

4. In the dropdown select Referrals. Here you'll see a list of any site that has sent traffic to your website during the time period selected. 

What to look for:
- Visits with a 100% bounce rate.
- Sites with an average session duration under 1 second.
- Unfamiliar URLs, often worded to entice a click. 

What's a small business marketer to do?
Google is well aware of the issue and has been silent on any pending solutions. There are some great blogs that discuss advanced solutions for creating valid hostname and spam crawler filters. These strategies are a bit complex for the average marketer with no coding experience, but they are helpful in preventing spammer data from invading your metrics. Essentially you tell Analytics what constitutes a real referral for your site, and which ones should be blocked. This is a solid option for larger sites managing a ton of data. Need help setting it up? CC Communications has experience so contact us to learn how we can assist.

If your site only receives a modest amount of traffic each month, then it may be easier to filter the data and do the math yourself. It's a manual process, but once you get the hang of it, you'll be able to do this rather quickly. Here are the steps:

1. In the referral section of analytics make note of the total number of session being reported.

2. Above the data and to the right is the word advanced. Click that and a new filter section appears.

3. Select the following criteria.
- In the first dropdown choose Exclude.
- In the second dropdown choose Site Usage and then Average Session Duration.
- In the third dropdown select Less Than.
- In the field next to that type 1 (one). 

It should now look like this:

This filters out any site visits that were less than one second and provides you with a cleaner list to review. Keep in mind that this scrubs a lot of the spam, but not everything.

4. Next go through and identify which referrals are real. Add those session numbers together to get your actual number of legitimate referrals and then adjust your overall website session metric accordingly.

Hopefully this type of data corruption will soon be eradicated, but until it is, marketers must remain diligent to ensure accurate data is being reported. If you have any questions or need assistance contact our team at marketing@cccom.com.