Ghosts may seem to be at work in your e-mail statistics, with some strange goings-on in. Understand what is really happening and learn how to correct the open rate errors and be smarter than your fellow-marketers.
Be smarter than the rest!
Don't make the very common e-mail marketing mistake caused by the "Ghost Click Effect"!
At The House of Marketing, we are constantly hiring young marketers. During such interviews, one of the questions we put to them is: “What would you use as key KPIs for an e-mail campaign, for instance?” The answer tends to be “open rate”, “click rate”. There is nothing wrong with that answer. The open rate is indeed a good KPI to check if (the subject line of) your e-mail is triggering enough interest to read it. And click rate is a good way to measure the engagement with the content of the e-mail.
Moreover, when designing automated campaigns, the open behavior and click behavior are used to send reminder e-mails or follow-up e-mails.
And yet, in both cases, I feel I need to send out a warning message. Many marketers are not aware that you need to be cautious with the open rate, as there is such a thing as the “Ghost Click Effect”.
“Ghost Click Effect” explained
What a typical marketer would expect is shown in the drawing: you have the clicks that are a subset of the opens, which are, in themselves, a subset of the sent or delivered e-mails. Sounds logical right?
But if you deep-dive into your e-mail statistics, and I encourage you to do the test on your last big e-mail campaign, and compare per user if (s)he has opened the e-mail and also if (s)he has clicked, you will find some users have “not opened” and yet “clicked”. That’s what I mean by the “Ghost Click Effect”, as, if it wasn’t opened by the user, then there must have been a ghost clicking on the e-mail… If you don’t believe in ghosts, the question popping up in your mind right now must be “But how could this have happened?”
We need to make a quick foray into the technical side of e-mail marketing to understand. The opening of an e-mail is done in the mailbox of the recipient, whether by a PC software, a mobile app or a web service. And these inboxes are not controlled by the e-mail sending tool. This has prompted developers of campaign tools to come up with a way of measuring an e-mail open by adding a pixel (a tiny image 1px by 1px in size, hence the name) and tracking if this image is downloaded. Not a bad solution, but it has its limitations. By the way, a click is more reliable, as it tracks if the user has actually clicked on the link.
If you use for outlook as e-mail client, for instance, you have very probably seen the message below.
Now that you understand that the tracking of an open is done by a tiny image, a mail where the “Images are not displayed” will not download the tracking pixel and thus not be measured as an open. Yet, you may have read the e-mail and even clicked. All the proof you need that there is no actual ghost involved in the “Ghost Click Effect” :-)
How bad is this “Ghost Click Effect”?
We have done the test on sample e-mails sent to some of our clients and the effect is not negligible.
- A B2C mail at a bank had a 0.4% “Ghost Click Effect” error (1)
- A B2B mail in the energy sector had a 2.5% “Ghost Click Effect” error
- A B2C mail in the energy sector had a 3.1% “Ghost Click Effect” error
(1) Calculation method: If you have 10 clicks without an open and 1000 opens, the “Ghost Click Error” is 10/1000 = 1%
How to solve the Ghost Click Effect error?
There are 2 areas where you will see the impact of the Ghost Click Effect.
- In automated campaigns when you branch on “opened”:
If you want to send, say, a follow-up mail when clicked and a reminder when opened and not clicked, then the simplest way is to inverse the order: first set your decision block on clicked vs non-clicked to send the follow-up mail and then decide on opened vs not opened for the reminder mail.
If you are not using opens and clicks at a split in your campaign flow, but you only want to decide on opens, then you will need to construct the “true opens” by the following rule:
- “If truly opened” would then become “if opened or clicked”
- “If truly not opened” becomes “if not opened and not clicked”
If you apply this, you will be sure that a customer who did open and even clicked will never get a reminder e-mail and could see it as bad customer experience.
- In your “open rate” reporting
The ‘solution of least effort’ is to accept that there is an error in your reports. And by analyzing some past campaigns, you can even define the order of magnitude of the error. As you will see in the next paragraph, there are also other forces at play…
If you want it to be as accurate as possible, then you’ll need to step away from the standard e-mail reports of your e-mail marketing tool and create your own reports based on the ‘raw send data’ and calculate the click rate by counting the clicks and dividing it by the amount of sent (or delivered) e-mails. The ‘true open rate’ is calculated by counting the recipients who have opened or clicked and dividing it by the amount of sent (or delivered) e-mails.
But it is not over yet…
Two extra side notes on other “false” cases:
- When the spam tool of a company or mail provider is very strict, it will open the mail and check all links before dropping the e-mail in the inbox of a user, and you will see false opens and false clicks (so the inverse of the “Ghost Clicks”)
- Users who use the preview pane in outlook, will potentially trigger false opens if the preview is done, including images downloads, but the user didn’t actually read the e-mail.
Let us know your “Ghost Click Effect Error” value!
A quick wrap up first. The “Ghost Clicks” happen when a person may have opened and even clicked, but it is not tracked as an open (pixel not loaded). This means that, when using automated flows, deciding on ‘opened’ needs to be corrected to ‘opened or clicked’ in order to offer a correct customer experience. The same kind of correction applies to your reporting.
We have added 3 examples of “Ghost Click Error Rate” we have seen at our clients but are also interested in having a broader view of how ‘big’ this error can be. So, we invite you to do the analysis and share your results or have us dive quickly into your e-mail stats and calculate the effect for you. Use the form to send us your results or request to analyze it for you.