Commentary

Sniffing Clicks: Detecting Email Activity Generated by Non-Humans

You may have a highly segmented list. But how sure are you your emails are reaching your audience—that is, human beings?

Research by Omeda found that 87.3% of all recorded clicks across 1.77 emails in Q1 of this year came from nonhuman sources. Even newsletters recorded 82.4% of clicks as bot-driven. 

What can you, the poor email sender, do about that?

Breaker is offering a solution for B2B brands: RealClick, a tool that filters automated click activity from campaign reporting. Tied to this are Breaker’s new Model Context Protocol (MCP), a server and an application programming interface (API) that lets compatible AI tools work newsletter data and workflows, all the while maintaining approval. 

“You should be able to ask what worked, pull the relevant data, and prepare the next campaign," says Ben Billups, Founder and CEO of Breaker. "That is where AI becomes genuinely useful in the newsletter workflow.” 

advertisement

advertisement

Is this the fault of some evildoer who wants to hack your system? Actually, it has to do with automation—and email security systems. 

For instance, Microsoft’s Safe Links scans URLs—a necessary safeguard, but the Safe Links activity can show up in campaign reporting. 

This distorts metrics—and the results that are reported to advertisers.

"Security scanners will click on links several times, and those machine-generated clicks are not real engagement," Billups says. "Separating those automated interactions from reader activity provides a more accurate picture.”

Billups adds that RealClick, MCP server and API were designed around data marketers rely on, and the actions software is allowed to take—two areas where automation can create problems:

RealClick evaluates clicks for signs of automated activity, and separates that information from the engagement metrics B2B rely on.

In addition, total clicks are reported alongside RealClick data, providing a basis for evaluating content, identifying engaged prospects and conducting A/B testing. 

Next story loading loading..