Your week,
explained.
When your routine holds together. What changes it. One useful thing to try.
Fictional example · Three weeks of invented observations illustrate the design. These findings do not describe you.
Make room around 09:00–11:00
If it fits your routine, reserve part of this window for a document or spreadsheet task that benefits from continuity. Keep necessary communication available.
Documents & spreadsheets activity was more sustained in this window.
58% versus 24% average daily share in 20+ minute category-consistent app runs, on 15 matched weekdays across 3 weeks.
Explore this evidence ↓This looks like a useful scheduling option for this type of activity. Task differences may still explain it.
This is application continuity, not measured output, attention or a biological peak.
Choose this experiment to compare the same category and two-hour window on matched weekdays over the following seven days.
Your best windows for sustained work
09:00–11:00
Documents & spreadsheets · 15 matched weekdays across 3 weeks.
Compared within the same category on the same days. This describes continuity in application use, not your efficiency, concentration or natural energy.
19 h 30 min in the window; 21 h 45 min elsewhere. Higher continuity on 12 of 15 matched days.
Which observations support this window?
History considered: 2026-08-24 to 2026-09-13.
Matched dates: 2026-08-24, 2026-08-25, 2026-08-26, 2026-08-27, 2026-08-28, 2026-08-31, 2026-09-01, 2026-09-02, 2026-09-03, 2026-09-04, 2026-09-07, 2026-09-08, 2026-09-09, 2026-09-10, 2026-09-11.
Window comparisons use coding or documents separately, with at least 30 recorded category minutes inside and outside a two-hour window on the same weekday. Both window hours need 10 category minutes. Recurring means eight matched days across two weeks, with higher continuity on at least 70% of days and an average gap of 15 percentage points.
Continuity here means recorded category time belonging to an uninterrupted application run of at least 20 minutes that stayed in that category. Each day has equal weight. Task differences within a category can still explain the result.
Your week by day and hour
Cells show observed minutes. Deeper green means a larger share in 20+ minute application runs. It is not a performance score. — means no observations; 0 means that hour was observed but this category was absent.
| Day | 00 | 01 | 02 | 03 | 04 | 05 | 06 | 07 | 08 | 09 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mon, 7 Sept | — | — | — | — | — | — | — | — | — | 50 | 50 | 40 | — | 40 | 38 | 40 | — | — | — | — | — | — | — | — |
| Tue, 8 Sept | — | — | — | — | — | — | — | — | — | 45 | 45 | 40 | — | 40 | 38 | 40 | — | — | — | — | — | — | — | — |
| Wed, 9 Sept | — | — | — | — | — | — | — | — | — | 45 | 45 | 40 | — | 40 | 38 | 40 | — | — | — | — | — | — | — | — |
| Thu, 10 Sept | — | — | — | — | — | — | — | — | — | 33 | 33 | 40 | — | 40 | 40 | 40 | — | — | — | 20 | — | — | — | — |
| Fri, 11 Sept | — | — | — | — | — | — | — | — | — | 45 | 45 | 40 | — | 40 | 38 | 40 | — | — | — | — | — | — | — | — |
| Sat, 12 Sept | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| Sun, 13 Sept | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
5 days contain hourly detail this week. Hours use Europe/Dublin; repeated daylight-saving hours are combined. A blank is never treated as focused time.
Every day has its own shape
Mon, 7 Sept4 h 18 min observed
6.5 application changes / observed hour
1 h 55 min in 20+ minute application runs
First observed hour: 09:00. Last: 15:00. These identify hour buckets; the span between them is not measured work.
After 18:00: 0 min recorded, with 0 session starts in those hours. This is a clock reference, not a claim that you worked outside intended hours.
Application runs: Under 2 min: 10 · 2–10 min: 9 · 10–20 min: 10 · 20+ min: 5.
- Documents & spreadsheets 2 h 55 min · 68%
- Messaging 10 min · 4%
- Email 18 min · 7%
- AI assistants 15 min · 6%
- Browser research 40 min · 16%
Tue, 8 Sept4 h 8 min observed
6.8 application changes / observed hour
1 h 15 min in 20+ minute application runs
First observed hour: 09:00. Last: 15:00. These identify hour buckets; the span between them is not measured work.
After 18:00: 0 min recorded, with 0 session starts in those hours. This is a clock reference, not a claim that you worked outside intended hours.
Application runs: Under 2 min: 10 · 2–10 min: 9 · 10–20 min: 12 · 20+ min: 3.
- Documents & spreadsheets 2 h 45 min · 67%
- Messaging 10 min · 4%
- Email 18 min · 7%
- AI assistants 15 min · 6%
- Browser research 40 min · 16%
Wed, 9 Sept4 h 8 min observed
6.8 application changes / observed hour
1 h 15 min in 20+ minute application runs
First observed hour: 09:00. Last: 15:00. These identify hour buckets; the span between them is not measured work.
After 18:00: 0 min recorded, with 0 session starts in those hours. This is a clock reference, not a claim that you worked outside intended hours.
Application runs: Under 2 min: 10 · 2–10 min: 9 · 10–20 min: 12 · 20+ min: 3.
- Documents & spreadsheets 2 h 45 min · 67%
- Messaging 10 min · 4%
- Email 18 min · 7%
- AI assistants 15 min · 6%
- Browser research 40 min · 16%
Thu, 10 Sept4 h 6 min observed
9.0 application changes / observed hour
0 min in 20+ minute application runs
First observed hour: 09:00. Last: 19:00. These identify hour buckets; the span between them is not measured work.
After 18:00: 20 min recorded, with 1 session starts in those hours. This is a clock reference, not a claim that you worked outside intended hours.
Application runs: Under 2 min: 14 · 2–10 min: 11 · 10–20 min: 19 · 20+ min: 0.
- Documents & spreadsheets 2 h 35 min · 63%
- Messaging 14 min · 6%
- Email 17 min · 7%
- AI assistants 20 min · 8%
- Browser research 40 min · 16%
Fri, 11 Sept4 h 8 min observed
6.8 application changes / observed hour
1 h 15 min in 20+ minute application runs
First observed hour: 09:00. Last: 15:00. These identify hour buckets; the span between them is not measured work.
After 18:00: 0 min recorded, with 0 session starts in those hours. This is a clock reference, not a claim that you worked outside intended hours.
Application runs: Under 2 min: 10 · 2–10 min: 9 · 10–20 min: 12 · 20+ min: 3.
- Documents & spreadsheets 2 h 45 min · 67%
- Messaging 10 min · 4%
- Email 18 min · 7%
- AI assistants 15 min · 6%
- Browser research 40 min · 16%
Sat, 12 SeptNo observations
No activity was recorded for this date. Nothing about work, rest or focus can be inferred.
Sun, 13 SeptNo observations
No activity was recorded for this date. Nothing about work, rest or focus can be inferred.
Time, with its context
About 5 h 40 min out of 20 h 48 min belonged to application runs lasting at least 20 minutes. That describes continuity, not work quality.
| Category | This week | Share | Previous week | Previous share |
|---|---|---|---|---|
| Building & coding | 0 min | 0% | 0 min | 0% |
| Documents & spreadsheets | 13 h 45 min | 66% | 13 h 45 min | 66% |
| Meetings | 0 min | 0% | 0 min | 0% |
| Messaging | 54 min | 4% | 54 min | 4% |
| 1 h 29 min | 7% | 1 h 29 min | 7% | |
| AI assistants | 1 h 20 min | 6% | 1 h 20 min | 6% |
| Browser research | 3 h 20 min | 16% | 3 h 20 min | 16% |
| Media | 0 min | 0% | 0 min | 0% |
| Uncategorised | 0 min | 0% | 0 min | 0% |
These are whole-week totals, not adjusted for missing activity. A larger share does not necessarily mean more time: another category or overall coverage may have fallen. Browser category coverage was missing for 0 min.
Compare like with like
5 matched weekdays with similar recorded duration and category mix: 7.2 → 7.2 application changes per observed hour. This measure was broadly similar.
How activity was broken up
Across 5 days with routine summaries. Application runs count continuous use of one app; category visits can continue across apps. Neither is a task count.
| Category visits | Under 2 min | 2–10 min | 10–20 min | 20+ min |
|---|---|---|---|---|
| Documents & spreadsheets | 0 | 0 | 45 | 14 |
| Messaging | 54 | 0 | 0 | 0 |
| 0 | 31 | 0 | 0 | |
| AI assistants | 0 | 16 | 0 | 0 |
| Browser research | 0 | 0 | 20 | 0 |
Most recorded category transitions
- Documents & spreadsheets → Messaging: 38
- Messaging → Browser research: 20
- Messaging → Documents & spreadsheets: 18
- Browser research → Messaging: 16
- Messaging → AI assistants: 16
- AI assistants → Email: 16
Counts cover adjacent observations within a continuous recording period. Gaps and restarts are excluded. They do not identify the reason for a visit or whether two categories belonged to one task.
Another option, if it matters to you
Look at your short messaging visits
Consider whether a few non-urgent checks can wait for a natural break. Keep time-sensitive contact available.
54 category visits under two minutes on 09-07, 09-08, 09-09, 09-10, 09-11. At least half of visits were short on each day.
Some checks may be useful to group; they may also be necessary for your work. We cannot see notifications or why you switched. Category visits can cross applications. This does not count notifications or prove interruptions.
See supporting observations ↓How we would follow it up
If you choose this experiment, compare short visits per observed hour on matched weekdays with a similar category mix.
Your morning-window experiment
58% before → 58% afterwards across 5 matched weekdays. Similar observed duration and broad category mix; different tasks may remain. This does not prove that the experiment caused the change.
Definitions, coverage and limits
- Observed activity is a partial view. Blank hours mean no observations, not rest, focus or zero work.
- An application run is uninterrupted foreground use of one application. It ends at a gap, pause, restart or local midnight. The 20-minute threshold describes continuity, not output or attention.
- Window comparisons use coding or documents separately, with at least 30 recorded category minutes inside and outside a two-hour window on the same weekday. Both window hours need 10 category minutes. Recurring means eight matched days across two weeks, with higher continuity on at least 70% of days and an average gap of 15 percentage points.
- Window continuity additionally requires the same category throughout an application run. A browser tab can change without an application change. A category visit is continuous use of one category, even across applications; it is not a task or notification.
- Hour buckets use your report time zone. Repeated daylight-saving hours are combined; skipped hours stay blank. A session begins after a gap of at least 30 minutes or at local midnight. First and last hour do not measure the full span as work. After 18:00 is a clock-based description, not your agreed work boundary.
- Week and experiment comparisons match weekday, observed duration within 0.67–1.5 times and broad category mix within 20 percentage points total variation. Similar categories cannot establish identical tasks or cause and effect.
- AI-tool activity is one category. It does not establish time saved or work quality. Recommendations use deterministic rules, not an AI model.
Rules: personal-advice/v2. No productivity score. No activity sent to an AI model.