Learning your own diabetes

Telling whether something is food, timing or basal

The three have different fixes and different shapes on a graph. How to tell them apart, and why the background has to be sorted first.

Written by Updated 11 May 2026 6 min read
The 60-second answer

Almost every pattern worth solving comes down to one question: is this about the food, about when the insulin went in, or about the background underneath it all? Those three have different fixes, and telling them apart is the most useful diagnostic skill there is. It is also mostly a matter of looking at when things happen rather than how bad they were.

  • Background shows up when you are not eating: overnight, and in the gaps between meals. That is the only place it can be seen cleanly.
  • Timing shows up as a spike that comes back down on its own. The amount was right, it just arrived late.
  • Food shows up as ending somewhere different from where you started, several hours later.
  • Sort the background first, because a wrong background makes every meal result lie to you.
  • Watch the direction as well as the height: a background running strong, or a meal dose too large, shows up as a hypo (a low below 4 mmol/L where you feel shaky, sweaty or confused and need fast sugar). A hypo can arrive at the time or shortly after, and again hours later including overnight, and lows read on a graph as clearly as highs do.
  • Extra carbohydrate is yours to manage; ratios, timing and background changes are worked out with your diabetes team.
The three shapes

What each one actually looks like on a graph

The background is the one you can only see in the quiet stretches. Overnight, and in a gap of several hours between meals with no insulin recently given, nothing much is acting except your background. If glucose drifts steadily up or down across one of those windows, that drift is your background talking, and nothing about your meals can explain it.

Timing produces a very particular shape: a rise after eating that peaks higher than you would like and then comes back down to roughly where it should be, without you doing anything. That shape says the total amount of insulin was about right and it simply started working after the food did. It is a timing problem wearing the costume of a dose problem.

Food, meaning the amount rather than the moment, shows up in where you end up. If you are still meaningfully above where you started four or five hours later, once the rapid insulin has done its work, that points at the amount not covering what you ate. And the mirror image, ending well below where you started, points the same way in the other direction.

The order

Why the background has to be sorted first

This is the part that saves people the most wasted effort. Every meal result sits on top of the background, so if the background is wrong, every meal reading is being nudged by something that has nothing to do with the meal. A background running a little strong makes every dose look too big. One running light makes every dose look too small. Adjust the meal settings to compensate and you now have two things wrong instead of one, and they are hiding each other.

So the order is not arbitrary. Look at the quiet windows first, get the background approximately right with your team, and only then judge what meals are doing. It feels slower and it is dramatically faster, because from that point on the meal results are telling you the truth.

It also explains a frustration people describe often: changing meal ratios repeatedly and never quite landing. If the ground keeps moving underneath, no amount of adjusting the thing on top of it will settle, and the answer is not to try harder at the meals.

What it means for you

Using this without turning it into a project

You do not need to investigate everything. Pick the thing that bothers you most, decide which of the three it looks like using the shapes above, and gather a few days of evidence for that one question. A fortnight aimed at one thing beats months of general watching, and it gives your team something specific to work with.

What makes this genuinely easier is the quiet windows. You do not have to create them artificially: most people already have several hours overnight and usually a gap somewhere in the day. Those are free experiments that happen whether you look at them or not, and they are where the background is legible.

The one thing to hold onto is that all three fixes belong with your team. What you own is the diagnosis and the evidence for it, and that is the harder half. “My overnight drifts up by about the same amount every night” is a sentence that gets a specific answer; “my numbers are all over the place” gets a general one, and it is usually not even accurate.

What varies

What muddies the picture

Any of these can make a good setting look wrong for a few days, which is why several days beats one and why the boring background stuff, a fresh site, insulin that has not been cooked, is worth ruling out before concluding a setting needs changing. A slow meal from the night before can still be landing during what looks like a clean overnight window, and a hard session in the afternoon can pull all evening.

a site that is not absorbing wellillness startingactivity earlier in the dayfat and protein in the last mealwhere you are in a monthly cyclestress and sleepgrowth
Real-life examples

Same complaint, three different answers

Examples, not instructions or doses.

The drift with nothing acting

You have not eaten since six, took nothing since, and between nine and midnight you climb steadily. No meal can explain that window, so it is the background talking.

The spike that fixed itself

Lunch takes you high for two hours and then settles back to where you started. The amount was fine, the insulin just arrived after the food did, which is timing rather than dose.

The one that never came back

Five hours after dinner you are still well above where you began, with the rapid insulin finished. That points at the meal amount rather than at when it went in.

What to notice

Worth paying attention to

The quiet windows, overnight and between meals, since those are the only places the background can be seen without meals on top of it
Whether a post-meal spike comes back down on its own, which separates a timing problem from an amount problem
Where you end up several hours later rather than how high you peaked, which is what points at the meal amount
Sorting the background before judging any meal, because a wrong background makes every meal result misleading
Ruling out the boring causes first, a site not absorbing or insulin that has been too warm, before deciding a setting is wrong
That the diagnosis is yours to make and the adjustment is your diabetes team to make
What to ask your team

Questions that make an appointment useful

"Can we look at my quiet windows and check whether my background is right before we touch anything else?"
"This spike comes back down on its own, does that mean timing rather than the amount?"
"Which of my meals would you most like a fortnight of evidence on?"
"How would I tell the difference between a background problem and a site that has stopped working?"
"When I bring you data, what would actually be most useful for you to see?"
Sources