Why Bot-Driven Email Data Solutions are Problematic

By LoriBeth Blair On Apr 19, 2019

Google will stop all third-party use of Gmail inbox data to inform an email data staple called “panel data” on March 31. For the uninitiated, much to the frustration of privacy-minded individuals, email panel data broke down behaviors of real people and provide information about email engagement, activity and success.

The loss of Gmail’s data is especially problematic for vendors who rely on panel data for recipient and engagement information, despite its known shortcomings. They need a new way to provide comparable insights to their customers.

Companies (such as Return Path and eDataSource) are introducing computer-generated alternatives to create simulated inbox engagement meant to mimic human behavior and appear as real, engaged recipients. This approach is problematic for a number of reasons.

Companies such as Google, Microsoft and Verizon have never taken kindly to any attempt to circumvent spam filtering, and there is no indication they will treat simulated inbox engagement any differently. In the past, they acted with extreme prejudice against those they perceived to be bad actors, so marketers using these data networks could see a direct negative impact on their reputation.

Industry experts are expecting the worst because this method of creating data is not new. Boris Mizhen, a US-based spammer, was sued by Microsoft for creating fake accounts and using bots to move hundreds of thousands of his messages at Hotmail from spam into inboxes in an attempt to exploit spam filters.

Simulated inbox engagement is simply the latest iteration of employing inauthentic human behavior. Using past attempts as examples, the industry has a good idea of how such forced engagement will cause more problems than solutions.

Valuable Insights Will Be Skewed

Marketers track metrics based on real users and engagements. For example, a savvy marketer might keep their eye on message open rate. If a bot opens up one message but does not open another because of an algorithm, rather than an emotional (or unemotional) response to a subject line, that data isn’t necessarily representative of the email’s resonance with the audience. Any metric driven by human intentions and emotions, which is basically all of them, will be skewed by bots without the capability to react authentically.

Data from Robo-seeds Won’t Accurately Approximate Real User Inboxes

You can’t fool Gmail. For example, marketers still try to engineer their way into the inbox tab of their choice even though Gmail sorts mail appropriately for a reason (because recipients expect promotions to be in the promotions tab). Beyond that, bot-driven data solutions don’t take into account another major human component of email: recipient-level filtering. Once Gmail has an engagement signal from a user, either positive or negative, they are likely to prefer that data point over their general algorithm. At its most basic, it’s trying to reverse-engineer Gmail filters, and that’s rarely a successful endeavor.

Innocent Senders May Be Harmed

Simulated inbox engagement could likely have one major casualty: senders impacted by bot sign-ups who don’t even benefit from these bot-data providers. Nothing is stopping simulated inboxes from subscribing to email of non-customers, which means the bot can mark the sender’s messages as spam, regardless of how a human might interact with the message. These innocent bystanders could be penalized as a result of imitated inbox engagement without having the benefit of access to the data provided by the vendors. 250ok customers are already reporting dozens of bot sign-ups on their email lists, and they certainly are not paying for the insight provided by them. Is that ethical?

It’s Just Plain Different

The major appeal of email panel data was the fact it was real users. Real users provided true insight into behavior, what works, what doesn’t, and so on. These were authentic lists with actual humans. Replacing the value of panel data with a simulated solution just doesn’t cut it. You’re not comparing apples to apples. In fact, you’re comparing humans to computers. Do those things provide the same value to marketers? Probably not.

While Google’s shift in data policy is sending waves through the email industry, there are safe and effective alternatives to bot-driven email data solutions. By combining an optimized seedlist testing strategy with advanced analytics, a company can keep gathering valuable insights and steer clear of the inevitable bot crackdown.