We tested it
on ourselves
first.
A self-conducted glucose pilot — two people, ten days, Oslo 2026. Every morning reading, every paired breakfast, the method, and the limits, published in full.
The pilot
What we measured
We wore continuous glucose monitors and ate the same five breakfasts twice each — once normally, once after a can of Yogutide. Only one thing changed each morning: the drink.
01 · The signal
The average morning, with and without
Five mornings of each kind, aligned on the moment of the peak and averaged. The black line is breakfast without Yogutide; the beige line is the same breakfasts with it. The shaded band is the ±1 SD spread across mornings.
Figure 1. Time-aligned mean glucose around the morning peak. Hover for values. Drag the gold divider to scrub between the morning without Yogutide (left) and with it (right). The peak with Yogutide is lower and arrives later; the window mean (−60…+90 min) barely moves.
Where the effect shows up
Two ways of looking at the same ten days.
Paired breakfasts. Each row is one breakfast eaten twice. Black dots are the peak without Yogutide, beige dots with it — a line that slopes down means the peak was gentler, and the row glows gold. Hover the labels on the left to see the full breakfast.
02 · Day by day
Every paired breakfast, up close
Same meal, same person, consecutive mornings. Pick a breakfast and a participant — the curves are the real 5-minute readings from 06:00–12:00.
03 · All mornings
Ten mornings, at a glance
Click any morning to jump to its paired comparison. Black mornings are without Yogutide, beige mornings with it. The dot marks the peak. Four of these mornings produced the biggest spikes of the study — they’re hiding here. Can you find them all?
04 · The numbers
Every reading, every day
Computed from the raw CGM files by study_core.py — the same module behind reports v1–v9. Nothing here is typed by hand.
| Day | Breakfast | Condition | Peak (E) | t peak (E) | Post 0–60 (E) | Win −60…90 (E) | Peak (R) | t peak (R) | Post 0–60 (R) | Win −60…90 (R) |
|---|
Glossary. Peak — highest morning reading. t peak — when it happened (minutes since midnight). Post 0–60 — mean in the hour after the peak. Win −60…90 — mean from 60 min before to 90 min after the peak. Units: mmol/L.
| Paired breakfast | Eva Δ peak | Eva Δ t peak | Rahmad Δ peak | Rahmad Δ t peak | Outcome |
|---|
Paired deltas = with Yogutide minus without, per breakfast. Negative peak Δ = gentler peak. Timing (Δ t peak) swings a lot from breakfast to breakfast — the average delay is the robust signal. Click a column header to sort.
05 · CGM 101
How to read the graphs
The plain-language guide we published alongside these numbers — what a continuous glucose monitor measures, the units, and how to read every chart on this page.
What a day of CGM data looks like
The shaded band is the 3.9–7.8 mmol/L reference window. Hover to read values.
Annotated example. After breakfast, glucose rises to a peak, then returns towards baseline.
Reading the graphs — a cheat sheet
Everything on this page uses the same visual language.
- Black line — mornings without Yogutide. Beige line — mornings with Yogutide.
- Higher and earlier peaks mean a sharper rise; lower and later peaks mean a gentler one.
- The shaded band around a line shows how much the mornings varied (one standard deviation).
- The dotted horizontal lines are the 3.9 and 7.8 mmol/L reference marks.
- One day is a snapshot, not a conclusion — the averaged curves and the per-pair tables are the representative view.
How our study used a CGM
Two participants wore the sensors for ten days; five breakfasts were each eaten twice.
On each pair of mornings, one breakfast was eaten normally and the same breakfast was eaten after one 330 ml can of Yogutide (10–20 minutes before). The key quantity is the peak — the highest reading after breakfast. Averaged across the five paired breakfasts and both participants, the peak was about 1.0 mmol/L lower and arrived roughly 19 minutes later with Yogutide.
06 · Method & honesty
How it was done — and what it isn’t
A pilot, not a clinical trial. This is exactly what we did, measured, and the limits we know about.
The honest part
- n = 2. Two people, ten days. A directional pilot — not proof for a population.
- No placebo, no blinding. We knew which mornings had Yogutide. Expectation can matter.
- Not identical days. One morning included a five-hour train trip; two mornings were work days.
- CGM ≠ blood. The sensor reads the fluid under the skin, which tracks blood glucose but isn’t identical to it.
- We have skin in the game. It’s our product. That’s why we publish the raw files and the code — check us.