Notification Respect Index. Swedish publishers (pilot v1)
Eight Swedish news sites observed for 14 days on ordinary-user devices. Methodology, rubric, and aggregate frequency data, no spam labels.
Published: 2026-08-14 · Collection window: 2026-01-06 → 2026-01-19 (14d)
Sample
- Sites observed
- 8
- Active subscriptions
- 8
Desktop Chrome + Android Chrome, ordinary-user conditions, no automation beyond standard browser permission grants.
Aggregate findings
Median notifications/day
2.4
75th percentile/day
4.1
Peak hour (local)
08:00–10:00 CET
Sites with preference control
3/8
Sites with clear unsubscribe path
5/8
Repetitive send days
38%
Days with 2+ sends
62%
Dimension summaries
Notifications per day
Median 2.4/day; 75th percentile 4.1/day across the pilot sample.
Send timing
Morning cluster 08:00–10:00 CET on weekdays; fewer sends after 21:00.
Repetitiveness
38% of site-days had 3+ notifications about the same topic cluster within 6 hours.
Preference control
3 of 8 sites exposed a visible in-product preference or category control before or after opt-in.
Unsubscribe path
5 of 8 had a clear browser or in-page path to revoke permission; 3 required menu digging only.
Tonality (rubric)
Most breaking alerts used neutral newsroom tone; 2 sites used urgency language on non-breaking items (observer rubric, not sentiment ML).
Observed relevance
Same-day news matched site section in 71% of sends where section metadata was visible in payload title.
Site-level data (neutral labels)
| Site | Notifs/day | Timing | Preference control | Unsubscribe path | Repetitiveness |
|---|---|---|---|---|---|
| Publisher 1 | 1.2 | Single morning window | None detected | Clear | Low |
| Publisher 2 | 3.8 | Morning + afternoon bursts | Partial | Clear | High |
| Publisher 3 | 2.1 | Morning cluster | Visible | Clear | Medium |
| Publisher 4 | 4.5 | All-day, peak 09:00 | None detected | Unclear | High |
| Publisher 5 | 1.8 | Morning only weekdays | Visible | Clear | Low |
| Publisher 6 | 2.9 | Morning + evening | Partial | Clear | Medium |
| Publisher 7 | 3.2 | Afternoon-heavy | None detected | Unclear | Medium |
| Publisher 8 | 2 | Two fixed windows | Visible | Clear | Low |
Scoring rubric
- Frequency: count of notifications received per site per calendar day (local timezone).
- Timing: histogram of receive timestamps in CET; no send-time optimization claims.
- Repetitiveness: 3+ sends on overlapping topic keywords within 6 hours = repetitive day.
- Preference control: visible UI to narrow categories or frequency before/after opt-in.
- Unsubscribe path: browser permission revoke or in-product off switch reachable in ≤2 clicks from notification settings.
- Tonality: human rubric (neutral / urgency-on-non-breaking), not automated sentiment.
- Relevance: title/section alignment with same-day homepage section when metadata present.
Limitations
- Pilot v1: N=8 sites, one industry, one country, not a national census.
- Homepage opt-in paths only; paywall or app-only flows may differ.
- Observation used EngageNudge team devices; no subscriber PII collected.
- Sites labeled Publisher 1–8 in public report; internal mapping available to research team only.
- Cannot infer Chrome enforcement outcomes. Browser-behavior claims require a cited public source.
Share this report
Know when to nudge.
Know when a notification is worth sending.
The useful question is who should not get this one.