UnitedHealthcare just expanded a behavioral coaching program aimed at youths, scaling something that was already working rather than launching a new pilot. That's worth sitting with for a second. A payer with enormous incentive to spend money only on what moves outcomes decided this was worth more investment. The interesting part isn't the healthcare angle. It's the model underneath it.
Behavioral coaching, by design, doesn't treat a 14-year-old and a 17-year-old the same way just because they both fall under "youth mental health." It looks at what a specific young person actually does: how they respond to stress, what triggers withdrawal, what kind of support they engage with versus tune out. Age and diagnosis are the label. Behavior is the thing that actually determines what happens next.
Most campaigns still sort people by label, not behavior
This is where the story stops being about healthcare and starts being about every campaign you've ever seen fall flat despite a clean-looking strategy deck. Most marketing segments an audience by demographic: age, income, job title, ZIP code. Those labels are easy to pull from a spreadsheet. They're also weak predictors of what someone actually does when they land on a page or open an email.
of visitors on average abandon a page during their first visit without converting, across most industries, regardless of how targeted the traffic was.
Source: Baymard InstituteThat number doesn't move much just because the targeting got sharper. A campaign can nail the demographic and still lose the person at the exact moment it matters, because the thing that determined whether they stayed was never their age bracket. It was a friction point on the page, a mismatch between what they expected and what they got, a CTA that showed up before they were ready for it.
What UnitedHealthcare's model gets right
A coaching program built around behavior instead of category has to answer a harder question up front: what does this specific young person actually do, and where does the current approach lose them? That's a behavioral read, not a demographic one. It's also the exact question most campaigns skip, because demographic targeting is faster to set up and looks like enough.
The label tells you who someone is on paper. The behavior tells you what they'll actually do next. Most campaigns are built for the first and lose money on the second.
It's worth being honest about what that costs. A campaign aimed at "parents of teens" instead of "parents whose teen disengages after the third email" isn't wrong, exactly. It's just imprecise in a way that shows up later as a drop-off nobody can quite explain. The traffic report looks fine. The conversion number doesn't.
Behavior-first thinking travels well beyond healthcare
None of this is unique to coaching programs or healthcare payers. Any campaign selling to more than one kind of person is making the same choice UnitedHealthcare made when it decided demographic targeting wasn't precise enough: either segment by what people do, or accept that some of the audience gets a message that wasn't built for them and will quietly leave.
- A landing page written for "small business owners" when the actual split is between owners who are price-sensitive and owners who are time-sensitive
- An email that assumes everyone opening it is at the same stage of a decision, when some are three weeks out and others are three minutes out
- A single ad set targeting a job title when the real behavioral difference is between people who research before buying and people who buy on impulse
This is the same read DayClerk runs before it generates anything. Instead of starting from who a segment is on paper, its audience simulation models what each segment actually does, producing the behavior paths, friction points, and drop-off risks that then decide the page's module order and the email's tone before a single line of copy gets written.
The organizations getting this right, whether they're a national payer or a small business running one campaign, share the same instinct: stop asking who the audience is, and start asking what they'll actually do.
