You've spent a fortune on wellness apps, push notifications, and free gym memberships. You've told your employees about the no-cost health coaching, the zero-deductible preventive care, and the shiny mobile app that lets them see their claims in seconds. And still, maybe one in ten employees with a chronic condition actually uses the program you paid for. Annual checkups? Same story. The most common feedback you hear is, "I don't know where to start."
Conventional wisdom says the problem is motivation. Or health literacy. Or maybe a clunky interface. So you throw more money at better design, bigger incentives, more frequent reminders. The real culprit is the architecture of your data, not the member's willpower.
The problem isn't the app. It's the pipes.
Most health plans are built on a transactional data model. Your claims system, your pharmacy system, your care management platform, your wellness vendor, your EAP, your navigation tool. They're all running on different technology stacks. Sometimes different companies you acquired over the years. They don't talk to each other in real time. When a member logs in or gets a message, the system doesn't see the full picture. It sees a single trigger, like a recent claim or a diagnosis code, and sends a message based on that one narrow thing.
This creates three engagement killers that almost nobody in the industry addresses directly.
Killer #1: The orphaned member
Consider an employee named Sarah.
- Her claims system shows an urgent care visit for chest pain.
- Her pharmacy system shows a nicotine patch prescription.
- Her wellness platform shows she's prediabetic from a recent biometric screening.
Because these systems don't share data in real time, the engagement engine sends Sarah a generic message: "Here's a tip sheet for chest pain." It completely misses the fact that Sarah has multiple risk factors that, seen together, scream for a coordinated intervention: a smoking cessation coach, a dietitian, a cardiology navigator. The system sees parts. Sarah feels unseen. She ignores the message. The data model itself prevents the kind of contextual engagement that might actually work.
Killer #2: The walled garden of point solutions
Most large employers now have eight or more point solutions: medical, dental, vision, EAP, fertility, musculoskeletal, diabetes, mental health. Each one has its own login, its own engagement algorithm, its own definition of a "healthy member." Each one sends its own notifications, often at the same time.
Now think about a member dealing with depression (through the EAP) and chronic back pain (through the MSK vendor). Two separate, uncoordinated campaigns hit their inbox on the same Tuesday morning:
- "Take this mental health screening."
- "Try this stretch for your back."
The member knows these conditions are connected. But the system doesn't. So the person has to manage multiple portals, remember different passwords, figure out which recommendation to follow first. It becomes a second job. The engagement strategy creates the very friction it was supposed to remove.
Killer #3: The missing signal
Everyone talks about personalization, but true personalization requires more than a diagnosis code. You need a behavioral signal, some evidence of intent or emotional state. Most systems have only billing events. They know you had a blood test. They don't know you're anxious about the results. They know you searched for a specialist. They don't know you abandoned the appointment booking halfway through.
The system isn't designed to meet the member where they are. It's designed to send a message based on where a claim code says they were. So the member gets a reminder for an annual checkup six days after they already had one. Or a nudge to schedule a colonoscopy when they're in the middle of a cancer scare. The engagement feels tone-deaf because, structurally, it is.
The invisible integration layer
The smartest health plans are quietly solving this problem, not by building a better app but by building an invisible integration layer. This isn't a new user interface. It's a core system re-architecture that does three things:
- Ingests passive data - phone movement (steps), smartwatch vitals, and, where members opt in, grocery purchase patterns (via retailer APIs) alongside claims. It creates a behavioral phenotype, not just a clinical diagnosis.
- Unifies the member record - EAP, MSK vendor, medical plan, all queried in a single, low-latency API call. The engagement engine sees one person with multiple needs, not three separate accounts.
- Enables predictive friction reduction - Instead of a generic reminder email, the system automatically syncs the member's calendar with available appointment slots across all their plans (vision, dental, medical). The engagement is the completed transaction. The notification is just an afterthought.
The consent and privacy layer
When an integration layer reaches past claims into steps, vitals, and purchase history, the data it pulls in often originates outside HIPAA's covered-entity rules. A grocery retailer and a smartwatch app are not health plans, so the federal privacy rule that protects clinical records doesn't cover what they collect. State law now fills that gap. Washington's My Health My Data Act, in effect since 2024, was the first state health-privacy law of its kind: it requires consent before regulated entities collect or share consumer health data and treats any violation as an automatic Consumer Protection Act offense. Other states have passed or proposed similar laws, and the FTC has used its Health Breach Notification Rule against health apps that shared user data without permission.
Consent has to be part of the architecture. Members who share steps or purchase history should know exactly what is collected, what it is used for, and how to turn it off. The plans doing this well make opting in easy to understand and easy to reverse.
The evidence also favors keeping purchase data optional. A 2025 Milliman study found that adding consumer purchasing data to its risk model added almost no predictive lift. If the data doesn't change the prediction, it isn't worth the trust you spend to collect it. The unified member record from your own vendors is the real fix. The grocery data is optional, and for many members it's a reason to opt out of the whole thing.
What you should ask your vendor tomorrow
You're probably measuring engagement by clicks. You should be measuring it by systemic efficiency: did the member accomplish what they needed with the least possible friction?
Next time you're in a meeting with your TPA or benefits platform vendor, ask them one question:
"When my member logs in, do you query their holistic health profile from a single, unified data lake, or are you hitting a claims database from 2004?"
The answer will tell you everything you need to know about why your engagement rates are stuck. The problem isn't the member. It's the pipes. Until we fix the architecture of our data, we're just building a better-looking mast on a ship that's taking on water. Let's fix the ship.
This article is for general information only and is not legal, tax, or medical advice. Employers should consult their own advisors.