A connected glucose meter can transmit a reading immediately. It cannot, by itself, decide how a person will respond, whether the response fits a clinical plan, or who should pay for the support. Livongo built its diabetes offering around that larger sequence. The company paired devices and supplies with feedback and access to coaching. The product was therefore partly a measurement system and partly a service relationship, with an employer or other institutional customer often paying for an individual member's access.

Livongo's 2019 securities filing described a distribution strategy involving employers, health plans, government entities, and other organizations. Contemporaneous Axios reporting identified self-insured employers as a major source of revenue. This created two different tests of usefulness. Members needed help that fit their lives. Benefits purchasers needed an offer that could justify its price, integrate into an existing benefits portfolio, and reach enough eligible people. A compelling demonstration for one audience did not automatically settle the other audience's question.

The service design made a plausible connection between information and action. Automatic transmission could reduce the friction of sharing measurements. A coach could answer questions or respond to unusually high or low readings. Supplies could reduce a practical barrier to monitoring. Those mechanisms describe what the program attempted; they do not establish the size of an incremental benefit. The distinction matters because a person can improve during a program for reasons that include ordinary clinical care, changing treatment, or renewed attention to a condition.

A randomized crossover study published in JMIR Diabetes in 2021 provides a useful encounter with that distinction. It involved 119 patients with poorly controlled type 2 diabetes at a specialty center. The intervention included a cellular-connected meter, telephone coaching, and responses to abnormal readings. Both study groups improved initially. The investigators did not find a statistically significant overall treatment effect across the twelve-month analysis, and described improvements as similar to usual care at that center. That is a qualified result, not evidence that connected support can never help.

The comparator is central to the interpretation. Adding a service to well-resourced specialty care is a different proposition from adding it where ongoing support is scarce. The study's population, duration, implementation, and clinical setting limit what can be inferred elsewhere. A purchaser should ask whether a proposed deployment resembles the evaluated setting. A founder should ask which component is expected to change a decision or behavior, and whether the evaluation can distinguish that component's contribution from the care already being delivered.

Teladoc's 2020 merger materials document a later strategic move: combining Livongo's chronic-condition capabilities with a broader virtual-care business. The transaction expanded the ambition, but a transaction is not a clinical endpoint. Integration creates additional questions about member experience, referral pathways, overlapping services, and the evidence needed for bundled purchasing. A company can have distribution value, a distinctive workflow, and an important role in patients' lives without every outcome claim being established by the same study.

Follow a measurement through to a decision and then through to a justified payment. What response occurs because the reading arrived? Who has authority to act? What happens when the member does not engage? What outcome would have occurred without the service? Livongo is useful because these questions cannot be answered by looking at the device alone. Testing one complete feedback loop, with a relevant comparison, creates a stronger foundation for a similar business than treating more data as equivalent to better care. The unit of value is the useful response, not the transmitted reading.

Sources and further reading

Sources checked November 4, 2026. Events and findings are described as of the dates in the narrative. Teaching insights are interpretation, not causal proof.

Keep exploring

Babylon Health: Growth could not outrun the care model.Cityblock Health: What if the missing infrastructure is trust?Put the question to work with an AI prompt →Browse all cases →

For a discussion tailored to your team, explore a workshop with Julien.