Analytics

[DennisD8]

Here’s my question: How can I do accurate A/B testing when I’m finding discrepancies in Constant Contact's metrics and my analytics?

I frequently see hundreds of emails have been sent, hundreds opened and many, many links clicked. 

In a recent email, Constant Contact reports 800 opens and 226 clicks to two of my web pages. Yet my website's (Wix) own analytics and a third party analytics (Clicky) reported only about 10 clicks to these pages.

CC's online agent admits BOTS are an issue and apologized. I suspect the opens being recorded are actually spam filters checking the mail first before allowing it thru. I suspect something similar on the web link clicks. 

This seriously questions CC's entire metric reporting premise. 

Anyone have any insight on this issue?  

[William_A · Accepted solution]

Hello @DennisD8 ,

If these contacts' networks / security programs are using bots to scan for malware, or they're distribution lists / auto-forwarding to larger groups of people, then the numbers are going to be skewed regardless.

To be frank, the reporting info for these contacts simply won't be viable ever. If you're wanting to keep them separate for the sake for your reporting, then I'd advise isolating these contacts to a new list meant for contacts that auto-forward or use bot security checks, and send separate email copies to that list. That way your main emails to your direct, human contacts will have more accurate reporting. 

If you'd like the step-by-step process of what I'm describing:

1. Identify the contacts that are causing these bloated open and click statistics. Consider tagging them for quick referral when you identify them, as this can quicken the list creation.
2. Create a new list, call it something obvious like "Suspected Bot / Auto-Forward Contacts," then add the suspected contacts to that list. If you need to be particularly granular with the list memberships, then I'd advise instead making "suspected..." copies of each of your lists, and applying these suspected contacts to the applicable copied lists.
   • Consider making a note in the suspect contacts of what lists they were previously on, or use the tags mentioned earlier, in case you decide later you don't like this setup and want to put the suspect contacts back into the main lists.
3. Once the suspect contacts are on the suspect list(s), remove them from the normal list(s) they're on.
4. From here on, you'll need to send two versions of the emails you'd normally be sending to these lists - one for the seemingly normal contacts, and a copy that you send to the suspect lists.

Beyond that, there's not much else you can do from your end as you're already self-authenticated. Our devs are constantly trying to identify the various programs that utilize this functionality, so they can set our system to ignore the behavior and triggers from these particular programs. If, after setting your account up for self-auth, the click rates continue to be inflated or worsen, then it may be worth calling in. That way our general support or higher level technical team can see if there's any further info that can be gleamed from the content of the emails with you live on the phone, or to see if the domains of the contacts that are seeming to bot-click them show any other particular consistencies we can track.

[DennisD8]

If my website and 3rd party analytics can tell the difference between an actual open and a bot open - why can't Constant Contact? 

[William_A]

Because they're wholly different systems, ways of tracking clicks / links and access, and a myriad of other factors.