CGM data is not a diagnostic tool. This guide explains how to read device data — it is not medical guidance. Always consult your healthcare provider to interpret your glucose patterns.
This guide distinguishes manufacturer-documented device features, provider-described app functions and clinical definitions of CGM metrics. Device and app instructions apply to their named model and version; proprietary platform scores are not clinical measurements of overall health.
Reading the glucose graph
A glucose graph commonly places time on the horizontal axis and glucose concentration on the vertical axis. Available time windows, units and views depend on the model and app. Check the device instructions rather than assuming every app has the same views.
The graph connects model-specific glucose updates. Original Libre 3 reports every minute and G7 every five minutes; Stelo’s app display is fifteen-minute despite five-minute sensor transmission. A graph does not guarantee the next value.
X-axis = time. Each point on the line is a glucose reading at that moment. Y-axis = glucose level. Higher on the graph = higher glucose. Lower = lower glucose.
Most apps overlay horizontal bands or dashed lines showing your target glucose range. Any part of the line inside the bands is "in range." Parts above are elevated; parts below are low.
Compare costs across all major devices for any monitoring period.
What the trend arrows mean
The arrow examples below describe original Dexcom G7 trend categories. Do not transfer these thresholds to every CGM or app. Arrows summarize direction and rate; they do not remove interstitial lag or guarantee the next value.
Trend arrows provide device-specific context about glucose movement. Follow your device instructions when interpreting them; an arrow is not a guarantee of the next glucose value.
The metrics your app reports
Time in Range
The percentage of readings that fell within your target glucose range over a selected period. Reported as a percentage — higher is generally considered better for most users. See our full Time in Range guide.
Glucose Management Indicator
GMI is calculated from mean CGM glucose over the actual report window, not necessarily 90 days. The original paper recommends at least 10, preferably 14 days. GMI is not laboratory HbA1c and is not a diagnostic test.
Standard Deviation
A measure of how much glucose varies around the average. A higher standard deviation means more variability in readings. Lower SD is generally associated with more stable glucose patterns.
Coefficient of Variation
Standard deviation expressed as a percentage of the mean glucose. CV is a normalized variability measure that allows comparison across different average glucose levels. CV is expressed as a percentage; availability and reporting windows depend on the app.
What a typical day looks like on a CGM graph
Fasting (morning)
An overnight or morning trace can have several patterns. An early-morning rise may be consistent with a dawn pattern, but the graph alone does not prove a hormonal cause.
After breakfast
Post-meal curve timing varies by meal, individual and measurement method. No fixed rise, peak or recovery time on a personal CGM graph establishes health or disease.
Exercise
Exercise responses vary, and sensor glucose may differ from blood glucose during rapid changes. Follow the model’s confirmation instructions and your care plan rather than treating a trace as an exact or causal description.
Sleep
An overnight low may be real or an artifact. Do not assume pressure explains it; follow the device’s confirmation instructions and your care plan for symptoms or low readings.
Compare all CGM options and their costs in our free calculator.
Common patterns and what they look like
Post-meal rise: a sensor graph may rise and fall after eating. Timing and height vary and are not diagnostic cutoffs.
Overnight decline: the graph may slope down. A low or symptoms still require the device’s instructions and individualized care plan; the shape alone does not establish that it is harmless.
Exercise-induced rise: A rapid upswing during or immediately after high-intensity activity, typically followed by a decline during recovery. The graph can look sharp and then quickly reverse.
A low reading may be real or an artifact, including pressure on a sensor. Do not dismiss an isolated low or assume a position change resolves it. Follow your device instructions and care plan for low readings, symptoms or a mismatch with a blood glucose measurement.
CGM data shows patterns — not diagnoses. If you observe patterns that concern you, discuss them with your healthcare provider. CGM data is not a substitute for clinical evaluation.
How different apps display the same data
| App | Graph style | Key metric featured | Data export |
|---|---|---|---|
| Dexcom (G7 / Stelo) | Continuous graph + trend arrows | Time in Range | CSV + Clarity |
| LibreView (Libre 3) | AGP report + daily graph | Time in Range | PDF + CSV |
| Levels Health | Graph + Metabolic Score | Metabolic Score | Limited |
| Nutrisense | Graph + meal tags | Glucose variability | Full CSV |
| Abbott Lingo | Graph plus Lingo Count | Lingo Count | Check current export features |
The underlying sensor data is the same — interstitial glucose measured at regular intervals. What differs is how each app processes, displays, and summarizes that data. Reporting tools and proprietary scores differ between apps. Check the provider’s current feature definitions and distinguish a proprietary score from a clinical metric.
Tips for getting more from your data
Where supported, meal and activity tags help compare logged events with glucose observations. Check your app’s actual logging features; the association does not prove cause.
Use the longer time views: The 24-hour view shows the full day's pattern. The 7-day and 14-day views reveal whether patterns are consistent or variable across days. Check the selected report period when comparing patterns across days.
Review patterns, but do not dismiss an isolated low, symptoms or a reading that differs from a blood glucose check. Follow the device instructions and individualized care plan.
Warm-up is model-specific: original G7 about 30 minutes, G7 15 Day and original Libre 3 about 60 minutes, Eversense about 24 hours. Follow the exact device instructions while readings are unavailable.
Compare device costs for any monitoring period with our free calculator.