Common Red Flags In Your Analytics

For most marketers, their biggest fear with analytics is that they’ll lose their tracking, turning sales into a black box that they can’t interpret. Corrupted, or invisible data can ruin your marketing efforts, but it’s typically noticeable. If analytics stops reporting sessions but you keep making sales, you know that something went wrong and can start working on a fix.

This is why proper data migration services are so important. Incorrect data is, in some ways, more insidious than your sales going dark because it’s harder to identify. An error can continue for weeks or months, causing you to develop a marketing strategy around flawed data, which can lead to spending resources where you shouldn’t, and ignoring places where you should.

Even after you fix the error, you’re left comparing the new data to statistics you know are inaccurate. Depending on how widespread the errors are, it can be difficult to know if you fixed the issue completely This is why it’s important to monitor your Google Analytics account closely, and to investigate anything unusual as soon as you notice it.

The key to maintaining correct analytics is to regularly test data integrity. The earlier you spot the issue and fix it, the sooner you can confidently measure your results. One of the difficult things about data errors is that they can appear differently depending on how you have your tracking configured. However, some types of errors tend to occur more often than others.

Here are a few early warning signs to look out for, and ways that you can troubleshoot the system to see what’s really going on.

 

Your Campaigns Have Mirrored Traffic

Larger eCommerce companies run multiple campaigns across platforms to capture revenue. This includes paid ads, SEO, social, and email. On top of this, a company might perform A/B tests to find a better layout for landing and conversion pages. UTM codes and other tracking parameters help track users in each campaign, but sometimes this data can get mixed.

If this is the case, AdWords traffic might appear as organic, or a successful social media campaign might mistakenly label all traffic as “direct” instead of coming from Facebook, Twitter, or Instagram. If this occurs, traffic will appear to drop from one source while increasing in another. For large organizations that have different specialists working on their campaigns, this change might not get noticed at first.

For example, if you’re monitoring organic and see a large drop in traffic over the course of several days, your first thought might be that Google updated their algorithm. It could look something like this:

Drop in organic traffic caused by analytics bug

 

This is why you should check the performance of all sources of data. In this case, there was an increase in direct traffic during the month that mirrored the dip seen in organic:

Organic traffic appearing as direct due to bug

 

After further investigation, it was discovered that an error in an A/B test caused all organic visitors who saw a test page to appear as direct traffic. Once discovered, the error was quickly corrected, but unless you monitor all sources of traffic, or the error only causes a handful of visits to shift every day, you might not notice anything is wrong until your traffic statistics are off by thousands of sessions.

 

Referral Spam Inflating Your Numbers

Shadow Brokers In a past article, we discussed the rise of traffics brokers. In short, websites that struggle to find advertisers (like porn and weapons dealers) sell their traffic to brokers who create redirects to safer sites. These blogs like “shoesforme.com” will redirect to retailers and brands that need traffic from display ads and referrals. The adult website can monetize while the advertiser gets referral traffic; however, to get the highest ad revenue, these brokers often send traffic to legitimate sites before redirecting them to their final destination.

If your website is one of those used by an ad broker, you might see a large spike in referral traffic from websites you never heard of, often to pages of little value. If you have an affiliate network, some of these secondary sites may be the ones that purchased traffic from the broker, causing these short sessions to appear under that channel instead.

Use the 80/20 rule when checking for fraud on your website and in your analytics. If you know your top 20% of traffic sources, then you should be able to easily spot any suspicious newcomers that quickly rise to the top. By keeping an eye on this 20%, you should keep 80% of your data clean.

Fraud can be found in a variety of metrics depending on your website. For eCommerce sites, you might see a jump in sales from your affiliate provider that doesn’t translate into sales on your website. For blogs and lead generation brands, fraud can be found in traffic spikes with minimal time on site and a small number of pages visited. A referral blog that drives 10,000 visitors in one day isn’t useful if none of them spend more than 30 seconds on your website before leaving.

This is another case where the corrupt data could be a blip on your radar for a few days or weeks. Google attempts to filter this data out automatically, but scammers will create additional pages as quickly as their old ones get shut down. The only way to keep your data as clean as possible is to closely monitor your own traffic sources. Once you weed out the fraudulent website, you can make a note in your analytics or adjust the numbers in your reporting to what they should have been.

 

You Suddenly See an Unexpected Sales Increase

Increasing Sales
If the data’s too good to be true, verify it.

What data is susceptible to corruption depends on what you’re tracking. The publishing world and the lead generation world are different from eCommerce, and the data they collect. The more detailed tracking you require of a particular data point, the more susceptible it is to corruption.

Advanced eCommerce sites will filter and tabulate this data to remove things their marketers don’t wish to track, such as sales tax, so that they can get a more accurate picture of revenue. However, if that filtering misfires, it can cause unusual changes in the data that might be difficult to pick up on, unless you’re looking for them.

In many cases, the data is typically corrupt after website maintenance or updates to the site’s tags. A tag might be misplaced or set twice, and the parameters could cause brands to capture double the information — or none at all.

If you see an unexpected jump or dip in your sales, be sure to double-check the data. One of the first places you should look is the Average Order Value. If the AOV is significantly higher — by whatever you happen to charge for shipping or your local tax rate, then you could be in trouble. The same is applied if your AOV is suspiciously lower. If this is the case, look at the top products that you sold and see if they’re any different than when your AOV was at a normal level.

In some cases, you may need to run a manual test to make sure the order values that Google Analytics (or whichever analytics provider you use) lines up with your actual website orders. For example, an eCommerce shop can discover an error where their analytics misfires only when someone has more than one product in their shopping cart, causing it not to track these potentially valuable sales.

Run the gamut of possible test orders — order, order with discount, two items ordered, etc. — and make sure the values match in both locations. You may find the data is only corrupt when specific issues occur.

 

The Solution: Run Regular Audits for Hidden Problems

These are just a few of the major issues that can cause inaccurate data, but there are dozens of smaller errors that could cause your tags to misfire. The best way to stop errors from corrupting your site is by running regular audits since this will make it easier to locate problems faster.

When we start working with clients, one of our first steps is to recommend a visibility audit. Our data specialists will review your tags and analytics to isolate any existing errors and then provide fixes to the code to ensure you have the correct data moving forward.

To make informed decisions, you need to have clean data. If you don’t know where to start, we can help. Contact us today for a free consultation or learn more about our web analytics consulting.