Bots Account for the Majority of Clicks in Email

If there is one universal truth about tracking things on the internet, it’s that data is never as clean as you might expect. And vendors make a ton of money helping advertisers and publishers try to get clearer on that data. Just Google “Ad click fraud” or “Click fraud” and you’ll see half a dozen different companies “saying goodbye to click fraud.”

And the sad reality is, the same is true in email. While I won’t classify it as fraud, the data is dirtier than we might first think.

One operator posted in the AMO Slack earlier this week saying: “[We’re] experiencing a huge increase in robo clicks lately, impacting our advertising client results. We use Sailthru. Anyone else experiencing this and have workarounds or know the source of the uptick?”

I can’t tell you how common this is. The newsletter publisher will report one number of clicks and then the advertiser will say that they only received some percentage of those clicks. Who should be trusted? As someone once said to me, “I know what clicks I am driving, I can’t help if the partner has bad analytics.”

Here’s the reality that we all need to come to terms with. The majority of clicks on your newsletter are not people. We’re not talking about 10 or 20%. We’re talking about a significant percentage of the clicks are not people.

So, I reached out to my friends at Omeda to better understand just how bad the problem is. In Q2 2023, Omeda sent 1.8 billion emails on behalf of all its media clients.

According to Omeda’s data, approximately 65% of all clicks in email are from bots. That percentage goes up and down, but over a long enough time horizon, it comes right back to ~65%.

I said above that I won’t classify it as fraud because it’s not. By and large, a lot of these bot clicks are security products that are protecting employees from cybersecurity threats. One example is Proofpoint, which acts as a middleman between the email sender and the recipient.

Here’s how we think email works (with my beautiful artwork):

Here’s what actually happens:

That third box exists to protect you from viruses, phishing attempts, or any other sort of cybersecurity risk. IT departments couldn’t care less about publishers needing clean data; their incentive is to ensure someone doesn’t accidentally get hacked.

And here’s how the software does this: it clicks every single link in the email. Once it verifies that there’s nothing wrong, the email is passed to the recipient’s inbox. This happens pretty quickly. But along the way, less sophisticated email service providers record that as a click. And when an ad campaign expires, you download an ad click report, and tell the advertiser that they got far more clicks than they actually did.

To make matters worse, it’s not always the same every day. When I spoke with Tony Napoleone, VP of Client Experience at Omeda, he said:

If you dig into a specific client, even down to a single daily newsletter, you could have very few clickbot clicks on a Tuesday, <0.1%, and then 95% of clicks the next day. There are sometimes really wild swings.

But so far, I have only been looking at this through the lens of reporting on ad clicks for your partners. What about tracking your own internal marketing? What if you build journeys that use click as the primary determinant for moving a user down funnel?

Alright, so clicks are obviously useless, right? No. It’s just not the perfect data point. You need to use a variety of data points to better understand what a user is doing versus just click. That means open rates, on-site behavior, and conversions all matter. These, plus clicks, can give you a better idea of the true engagement of your email list.

Smart ESPs automatically remove this data from reporting so that you are able to make better decisions. For example, it’s very easy to see that a user clicking every link in a newsletter in one second is probably not actually clicking all those links. And so, these ESPs just block that click behavior so you’re only reporting on true clicks.

But many ESPs are not smart. So, how can you figure out what percentage of clicks are from bots?

A simple tactic would be to put a hidden link in the newsletter. For example, if the background of your email is white, put a link with white text somewhere in the email. Unless someone is chaotic with their mouse, no one but the bots should click that hidden link. You can then use that to determine what percentage of that specific sent email has bot clicks.

But this tactic isn’t perfect for one very big reason: people with security bots protecting their inbox still manually click on stuff in newsletters. And so, that email will show clicks from the bot and then also show clicks from the user. So, we’re right back to square one with dirty data.

Another thing to consider is that the percentage of bot clicks is likely different industry to industry. For example, if you’re writing about government or defense, I’d suspect every email subscriber has some sort of security bot versus early stage startups that don’t know the first thing about cybersecurity.

There is going to come a point—soon, I think—where there will need to be a major education moment for advertisers. If email as an ad channel continues to grow, there will need to be better reporting. While this is clearly not fraud, it’s still not helping the advertiser and inflated numbers makes a publisher out to be better than it actually is.

And so, if you’re filtering out bot clicks and an advertiser says that you are driving far fewer reported clicks on their creative than a different newsletter, I would encourage you to push back. Tell them that 65% of all clicks in email are bots and you are filtering that out so you are honest. Why aren’t their other partners?

Thanks for reading today’s newsletter. If you have thoughts or wish to join the discussion on bots, the thread on Slack can be found here.

Tickets are still available for the upcoming AMO Summit held here in NYC on October 26th. I was going through the list of people who bought tickets and we’ve got some great companies represented, including Aging Media, Brick Underground, Workweek, Leonis Partners, Hearst, Endpoints News, Axios, Questex, Bloomberg, Payload Space, Vinepair, MIT Tech Review, Oaklins DeSilva+Phillips, Housing Wire, Flying Magazine, FreightWaves, TCG, Endeavor Business Media, and the list goes on. If you want to join this unbelievable list of attendees for the inaugural AMO Summit, buy your ticket today.