MentorBuilt is deliberately experimenting with tone, messaging, and timing

An announcement earlier this week revealed that MentorBuilt, the widely celebrated FRC digital media company, has undertaken an innovative human-subject research program which involves subtly varying the tone, messaging, and timing of its content to solicit responses from the readers.

We discussed this decision via a long chain of emails with Barley Kail, MentorBuilt Vice President of Development. "This new program is part of our ongoing experiments to deliver the most relevant and valuable content to our readers," Kail says. "However, we realize that some posts will fall flat. We don't expect them to resonate with all audiences."

By recording the amount of reader engagement with posts of different tones and style, we can gather insight about what our users want. "We hear your feedback loud and clear," says Kail. "We see the strongest passion about our site coming from FIRST alumni & volunteers, and we're always happy to get better ideas on how to nurture & engage."

But as the experiment proceeds and user engagement increases, the media company is changing in a way that is quite different from the "nurturing" environment described by Kail.

Preliminary results indicate that one class of MentorBuilt articles garners the most views and creates the most engagement: it's the content that is most divisive and malicious. When views and engagement are the only valuable metrics to stakeholders, then a high-performing website is one that's deeply upsetting to users.

Although the experiment is ongoing, MentorBuilt has already capitalized on this surprising result by posting a series of articles which are tone-deaf and controversial. The numbers of views and users are increasing in response, which in turn is generating more funding for the site. A perverse trend emerges: While these upsetting articles spur much more negative feedback, MentorBuilt's advertisers and sponsors provide a strong incentive to not listen to this feedback and instead double down on the inflammatory content.

At press time, IRB approval for the experiment was still pending.