Feedback loops compound toward whoever closes them.
This is the mechanical claim underneath everything that went wrong in digital health over the last fifteen years. A product does not get better in the abstract. It gets better at whatever the person paying for it is measuring. Whoever closes the loop shapes the product, quietly and relentlessly, over every iteration. The team can have the best intentions in the world, but the loop, much like the house, always wins.
The loop shapes the product
When a benefits team is the customer, the loop closes on engagement rates, enrollment numbers, utilization, HEDIS lift, per-member-per-month math that pencils out against a budget. The product gets very good at those things. It has to, because those are the things that get measured at renewal. Dashboards get polished. Onboarding gets streamlined. The surface area that procurement evaluates gets beautiful.
None of those things are what the person with diabetes wants. That person wants their A1c to come down and to stop feeling like garbage. The person with a sinus infection wants to know if it is bacterial. The person with chest tightness wants someone to figure out what is actually going on. These are different outcomes than the ones the loop is compounding toward, which is why Livongo can sell a connected glucometer to self-insured employers for an $18.5 billion exit while the actual clinical benefit to people with diabetes remains contested. The Peterson Health Technology Institute reviewed eight major platforms representing $58 billion in investment and found average A1c reductions of 0.23 to 0.60 percentage points, below the threshold for clinical significance, with a net increase in healthcare spending. The product did what it was optimized to do. Teladoc has historically maintained 94% client retention while actual member utilization has sat in the single digits to low teens across its enterprise base. If employee benefits disappeared overnight and Teladoc had to survive based only on payments from patients who used it, let's just say it wouldn’t be pretty.
What patients want
What do patients actually want from a doctor? Not bedside manner, not accessibility, not convenience. Those are proxies. When researchers have forced patients to choose between a doctor who is thorough and a doctor who is friendly, thoroughness wins at nearly five to one. Patients want someone who figures out what is actually going on with them. Everything else they say they value, they value because they believe it makes that more likely.
Medical ground truth
Call the thing patients actually want medical ground truth. The real clinical picture of what is happening with a person, updated over time, acted on with judgment. No payer-built product has ever been designed to deliver it, because payers cannot buy ground truth directly. Patients buy proxies for ground truth: bedside manner, access, responsiveness. Payers buy proxies for the proxies: engagement rates, utilization, HEDIS scores.
The honest part
Now the honest part. Making the patient the customer does not automatically close the loop on A1c. A patient paying out of pocket can churn for cost, for friction, for life circumstances, for reasons that have nothing to do with whether their care is working. The signal is noisy. What changes is the shape of the noise. Under a payer contract, a patient who is not getting better keeps showing up in the engagement dashboard and the product learns nothing. Under direct pay, a patient who is not getting better leaves, and the product either figures out why or dies. The loop is not clean. It is just pointed at the right thing.
The bet
That is the bet. Not that direct pay is a magic incentive. That it is the first arrangement in modern healthcare where a product has to actually move the thing the patient came to us to move, or it does not survive. Over enough iterations, with enough patients, a product built against that constraint compounds toward ground truth in a way a product built against procurement never will.
This article is Part 1 of a three-part series. Next: why insurance is the wrong financial instrument for chronic care, and why it may not matter, because for most Americans the insurance wrapper has already been hollowed out anyway. Follow us on LinkedIn for more.