Reading your own glucose graph: what time-in-range, spikes and 'wobble' are actually telling you
My CGM graph looks like a mountain range, everyone else’s looks smooth: how do I read time-in-range and calm the spikes?
A glucose graph that looks like a mountain range is not a mark against you. The most useful number on a continuous glucose monitor (a CGM, the sensor that reads your glucose every few minutes) is time in range: the share of the day your glucose spends inside a target band. That tells you more than any single reading, and often more than an average blood test, because two people with the same average can have very different graphs. Nobody’s line is flat. The job is to spot patterns, not to punish yourself for spikes, and any change to insulin is worked out with your team, never read off the graph.
- Time in range is the share of the day your glucose sits inside a target band, most often 3.9 to 10 mmol/L, and a figure many people aim toward is around 70%.
- Your own target and percentage are set with your team, because they differ for some people, including in pregnancy and for older or frailer people.
- A spiky line is normal: everyone rises after meals and drifts overnight, so the shape and the pattern matter more than any single peak.
- Time spent below range (going low) matters as much as time high, and keeping lows small is part of reading the graph well.
- The graph shows you patterns to take to your team; it does not tell you what to dose, and that stays a shared decision.
What time in range actually measures
Time in range is simply how much of your day your glucose stays inside a band, shown as a percentage. The band most clinics use is 3.9 to 10 mmol/L, and a common goal is spending around 70% of the day inside it, which works out at roughly 17 hours. Alongside it your report shows time above range (too high) and time below range (too low), so the three add up to your whole day.
Two things are worth knowing straight away. First, that 70% is a widely used starting point, not your personal target: teams set a different figure for some people, and pregnancy uses a tighter, lower band. Second, you can look at time in range for a single day, a week or a fortnight, and the longer views are steadier, because one bad night barely moves a two-week figure. Your report may also show a GMI (glucose management indicator), an estimate of your longer-term average from the sensor data.
Why the graph tells you more than the average
For years the main measure was HbA1c, a blood test that gives your average glucose over about the last three months. It is genuinely useful, but an average hides how you got there. Someone who sits fairly steady and someone who bounces between highs and lows can land on exactly the same HbA1c, because a string of highs and a string of lows average out to a middling number.
That is the whole point of the graph. It shows the bouncing that the average erases, including the lows, which is why time below range is such an important line to read. A comfortable-looking average with a lot of time spent low is telling you something an HbA1c alone never would. So a smooth graph and a spiky graph at the same average are not the same picture, and the spiky one is worth understanding rather than feeling bad about.
Spikes, wobble and the patterns worth spotting
Some rise and fall is unavoidable. Every person rises after eating and drifts a little overnight, so a line with peaks in it is normal, not a failure. The word people reach for is wobble, and what it really means is variability: how far and how often your line jumps around. A calmer line is generally easier to live with, but no line is flat.
What turns a graph into something useful is looking for repeats. One spike is noise. The same climb after breakfast most mornings, a regular dip in the middle of the afternoon, or a slow drift up in the small hours are patterns, and a pattern that shows up over several days is the thing worth acting on. A useful pattern tells you when and roughly what, for example a predictable high two hours after the same breakfast, which is far more workable than a vague sense that your numbers are all over the place.
The graph shows you the pattern. It does not tell you what to change, because a change to insulin, its timing or your background dose is a decision made with your team from that pattern, not a figure you calculate off the screen.
What makes a graph spikier or calmer
The same person can have a smooth week and a jagged one, and the reasons are usually in this list rather than in anything you did wrong. Fast carbohydrate and insulin given a bit late both make sharper peaks; illness, stress and poor sleep push the whole line up and make it jumpier. Exercise moves it both ways and for hours afterwards. Many people also see an early-morning climb from hormones (often called the dawn effect), which is why waking numbers can look higher than the middle of the night. Sensor readings also lag your actual blood glucose by a few minutes, so a sharp spike can look slightly later and higher on the graph than it truly was.
How the graph reads day to day
Examples, not instructions or doses.
Your line climbs steeply after cereal, peaks around two hours, then comes back down on its own. That shape, up fast and back to normal without you doing anything, usually points at the timing of the meal dose rather than the amount, which is a pattern to show your team.
You and a friend both have an HbA1c in the same place, but your graph bounces and theirs is smoother. The average hid your lows and highs cancelling out. Your graph is not worse, it is just showing you more to work with.
Most afternoons you drift down around the same time. One day is chance, but the same dip on most days is a pattern worth flagging, not a reading to fix on the spot.
A new sensor reads jumpy and a bit off for the first day, then settles. Early-wear noise is common, so a strange first-day graph is worth confirming with a finger-prick before reading too much into it.