What Is a CGM? A Plain-English Explainer for Non-Diabetics

A CGM is a quarter-sized sensor that sits on the back of your arm and reads your blood sugar every minute for 14 days.

You can't feel it. You don't bleed. The sensor pushes a tiny filament into the fluid between your skin cells, measures glucose there, and beams the number to your phone.

Think of it like a fitness tracker — but instead of counting steps, it tracks what every meal, every coffee, every late-night snack actually does to your body. The number you've been guessing at for years finally shows up on screen.

That's the whole pitch. The reason non-diabetics started wearing them around 2020. The reason most people who try one for two weeks change something about how they eat afterward. I did.

Here's what a CGM actually is, how it works, what's normal, whether it's worth $80–$200 for two weeks, and what it can't tell you.

A glucose molecule is sugar in your blood. Every carb you eat breaks down into glucose. Your body uses some of it for fuel, stores some as glycogen, and — if there's too much left over — converts the rest to fat.

The job of insulin is to clear glucose out of your bloodstream after a meal. When that system works well, your blood sugar rises a little after eating, then comes back down within 2–3 hours. When it doesn't work well, sugar stays high for longer, and your body gets the message to store more fat and inflame more tissue.

A CGM measures glucose in your interstitial fluid — the fluid surrounding the cells just under your skin. That number tracks your actual blood glucose with about a 5–15 minute lag.¹

Every minute, the sensor sends a reading to your phone. By the end of 14 days, you have ~20,000 data points showing exactly how your body handles every meal, every workout, every night of bad sleep.

Three things matter from a CGM trace, and none of them is the single number you'd see at a doctor's office.

Fasting glucose — your baseline first thing in the morning. Optimal is 70–85 mg/dL. Anything consistently above 100 is the first warning sign of metabolic dysfunction.²

Post-meal spike — how high your glucose climbs after eating, and how fast. A spike under 30 mg/dL above baseline that returns to normal within 2 hours is healthy. A spike of 60+ mg/dL that lingers is the pattern that drives insulin resistance over time.³

Glycemic variability — the up-and-down jaggedness of your trace across the day. Smooth, gentle waves are good. Sharp spikes and crashes — even if the average is normal — predict cardiovascular risk independently of average glucose.⁴

The average across all of these gets compressed into HbA1c at your annual blood draw. That single number is a 3-month average. It hides everything interesting. A CGM is the opposite — it shows the texture.

For a healthy non-diabetic adult, here's what a CGM trace should look like:

Fasting glucose: 70–90 mg/dL
Post-meal peak: under 140 mg/dL (and ideally under 120)
Time spent above 140: less than 30 minutes per day
Glycemic variability (standard deviation): under 15 mg/dL across the day
HbA1c equivalent: under 5.4%

Most "healthy" adults who wear one for the first time are shocked. Their fasting glucose is fine. But a bowl of oatmeal with banana spikes them to 180. Their afternoon energy crash at 3pm is on the trace — a sharp drop from 140 down to 65 after lunch. The Korean BBQ dinner with rice keeps them elevated all night and they wake up at 110 instead of 80.

This is the value of a CGM for non-diabetics. The averages look fine on paper. The texture tells the real story.

The honest answer: yes, for 14 days, once. Maybe twice — once now and once after a real intervention.

Here's why. By the time conventional bloodwork catches metabolic dysfunction, it's been compounding for 5–10 years. HbA1c rises slowly. Fasting glucose doesn't move until the system is genuinely strained. But glycemic variability and post-meal spikes show up early — and they're modifiable.

You'll learn three things from two weeks on a CGM:

Which "healthy" foods aren't, for you. Oatmeal spikes some people and barely moves others. Same with rice, sweet potato, even fruit. Your response is partly genetic, partly microbiome, partly muscle mass.⁵ The only way to know is to test.

What sequence and timing actually matter. Eating protein before carbs flattens the spike. A 10-minute walk after dinner cuts it in half. Eating the same meal at 9pm vs 6pm produces a dramatically different curve. These aren't theoretical — you watch it happen.

Where your sleep, stress, and alcohol actually hit you. A bad night of sleep raises your fasting glucose the next morning by 10–20 mg/dL.⁶ Two glasses of wine destroys your overnight curve. A stressful call spikes you with no food at all. The CGM makes the invisible visible.

After 14 days, most people don't need to keep wearing one. The patterns are learned. The behavior changes stick.

In the US, a single 14-day Abbott Libre or Dexcom Stelo sensor runs $80–$100 without a prescription. Programs like Levels, Nutrisense, and Veri bundle two sensors plus app coaching for $200–$400.

In Korea, sensors run ₩50,000–₩80,000 each. No prescription needed for the Libre.

For non-diabetics, you don't need ongoing data. Two sensors, 28 days total, run once and again 3 months later after you've changed something — that's the protocol that produces insight without becoming another piece of optimization theater.

A CGM is honest about glucose. It's silent about everything else.

It doesn't measure insulin. Two people with the exact same glucose curve can have wildly different insulin levels driving them. A fasting insulin and HOMA-IR from a regular blood draw fills this in. We test it on every KINS panel.

It doesn't predict diabetes risk by itself. Genetic risk, body composition, family history, and inflammatory markers all matter. The CGM is one input.

It can't tell you what your gut microbiome is doing. Why oatmeal spikes you and not your spouse — that's microbiome territory, and a CGM can't see it.

It will lie to you in the first 24 hours. New sensors take a day to calibrate. Ignore everything from day one.

It will make you neurotic if you wear one too long. Most people get the insight in two weeks. After a month, the data turns into anxiety. Take it off.

If you're a high performer who hasn't worn one yet, here's the simplest path.

1. Order a Libre 3 sensor from Amazon (US) or Coupang (Korea). No prescription needed.
2. Apply it on a quiet Sunday morning. Skip the first 24 hours of data.
3. Eat normally for the first 7 days. Don't try to "win." You want to see what your current life actually does.
4. Days 8–14, run small experiments. Eat protein first. Walk after dinner. Try the same meal at two different times.
5. After 14 days, take it off. Write down three patterns you learned.

That's it. Two hundred dollars, one shocking week, and a permanent shift in how you think about food.

Does it hurt to put on?
You feel a small click. No real pain. The filament is thinner than a hair.

Can I shower and swim with it?
Yes. It's waterproof to about 3 feet for 30 minutes.

Does it work if I lift weights or run?
Yes. The sensor on the back of the upper arm stays put for the full 14 days through most workouts. Some people use a Skin Grip patch for extra security.

Will my insurance cover it?
In the US, only if you have a diabetes diagnosis. Non-diabetic CGM use is out-of-pocket. In Korea, the same — pay cash.

What's the difference between a Libre and a Dexcom?
Both measure interstitial glucose every minute. Libre is cheaper and the app is simpler. Dexcom Stelo (the over-the-counter version) is similar. For 14 days of self-experimentation, either is fine.

Can a CGM be wrong?
The sensor can drift 5–10% from a finger-stick reading. If a number looks shocking, confirm with a finger-stick before changing anything major.

• Comparison: Best CGM for non-diabetics (coming soon)
• Personal: What my CGM told me about my breakfast (coming soon)
• Related: How to test cortisol — and why every high performer should

1. Bailey TS, et al. (2015). The performance and usability of a factory-calibrated flash glucose monitoring system. Diabetes Technol Ther, 17(11), 787-794. PubMed
2. American Diabetes Association (2024). Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes. Diabetes Care, 47(Suppl 1), S20-S42. PubMed
3. Hall H, et al. (2018). Glucotypes reveal new patterns of glucose dysregulation. PLOS Biology, 16(7), e2005143. PubMed
4. Monnier L, et al. (2006). Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA, 295(14), 1681-1687. PubMed
5. Zeevi D, et al. (2015). Personalized nutrition by prediction of glycemic responses. Cell, 163(5), 1079-1094. PubMed
6. Spiegel K, et al. (1999). Impact of sleep debt on metabolic and endocrine function. Lancet, 354(9188), 1435-1439. PubMed

This is educational content. It's not medical advice. If you have diabetes, prediabetes, or a family history of either, work with a doctor before starting a CGM. The patterns on the trace need clinical context.