feat(linkedin-studio): N16 — out-of-network-andel + patterns-oppdatering + boundary-map [skip-docs]

Reach-splitten (in/out-of-network) er native i LinkedIns post-analytics siden juni
2026, men vises som PROSENT og finnes ikke i CSV-eksporten. Planen antok to
manuelle antall; verifiseringen viste prosent, så modellen er ett felt —
outOfNetworkPct — og in-network er komplementet.

- parseOptionalPercent: egen parser, ikke parseOptionalCount. Komma er desimal
  (36,5 -> 36.5, aldri 365), og verdi >100 avvises: i én kolonne kan ikke et
  absolutt antall skilles fra en andel, så svaret er unknown, ikke en gjetning.
  Blank/ikke-numerisk/negativ -> unknown; ekte 0 beholdes.
- Ett lagret halvpart, kryssjekket: In-network godtas og lagres som komplement;
  et transkribert par som ikke summerer til ~100 (±1 avrunding) forkastes som
  unknown i stedet for å bli halvveis trodd.
- weightedOutOfNetworkPct: impressions-vektet roll-up (avgOutOfNetworkPct, uke +
  måned). Flatt snitt lar en 50-visnings-post slå en på 10 000; poster uten
  avlesning ekskluderes, og null vekt gir undefined — aldri 0, aldri NaN.
- Reach inngår ALDRI i engagementRate (distribusjon != engasjement). Rapporten
  leser den som akvisisjon (ut) vs resonans (inn), og sier «ikke ført for denne
  perioden» framfor å estimere. En reach-innsikt går inn i N15s do-next-kanal.
- Step 7c (A2-F11): rapporten tilbyr diff mot brukerens engagement-patterns.md
  med eksplisitt go — aldri stille skriving, aldri inn i den shippede malen.
- Boundary-map (E#9): dwell eksplisitt umålbar, saves partner-gated, reach
  native men CSV-eksport uverifisert.
- Reach-frie importer er byte-identiske med før, på skjerm og på disk.

TDD: rødt bevist først (10 feilende), analytics 119 -> 144 tester, tsc ren.
test-runner 232 -> 247 (Section 16w, gulv 213 -> 228). Alle suiter grønne.
CHANGELOG: N15-oppføringen manglet og er backfilt sammen med N16.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QxvWAjte7vPcF79QeSRvRJ

assets/analytics/README.md

@@ -28,6 +28,43 @@ comparable to older imports). Saves is the strongest organic engagement signal,
 so the reports surface it as its own line. **Dwell time stays unmeasurable** —
 it is internal to LinkedIn for organic posts, with no count to transcribe.
 
+### Optional: add the out-of-network reach share (manual)
+
+LinkedIn splits a post's impressions into **in-network** (people who already
+follow or are connected to you) and **out-of-network** (people who found it
+through recommendations, reshares or search). It sits in your post analytics
+under **Discovery**, beneath the impressions count — a global rollout that began
+in June 2026 — and it is shown as **percentages**, not as two counts. It is
+**not** in the CSV export, and whether it will ever be exported is unverified.
+
+To track it, add an `Out-of-network` column and type the percentage you read off
+that panel. The `%` sign is optional:
+
+```
+"Content","Date","Impressions","Reactions","Comments","Shares","Clicks","Out-of-network"
+"My post...",2026-03-10,5000,100,30,15,200,37%
+```
+
+Details worth knowing:
+
+- **Either half works.** An `In-network` column is accepted instead and stored as
+  its complement (`In-network 63%` → out-of-network 37). Only the out-of-network
+  share is kept, because the two halves describe one split — storing both would
+  let a record contradict itself.
+- **Both halves are cross-checked.** If you transcribe both and they do not sum
+  to ~100 (one point of rounding slack allowed), the reading is discarded as
+  **unknown** rather than guessing which cell was misread.
+- **Percent, not a count.** A value above 100 is refused (`unknown`): in one
+  column an absolute impression count and a share are indistinguishable, so the
+  importer will not guess. A comma is read as a decimal mark (`36,5` → 36.5%).
+- **Unknown is never 0.** A missing column, a blank cell, or a non-numeric cell
+  leaves the share unknown. A genuine `0` is kept — nothing left your network.
+- **Not part of the engagement rate.** Reach is a distribution signal, not
+  engagement: a high out-of-network share means the post **acquired new
+  audience**, while high in-network engagement means it **deepened the audience
+  you already have**. Roll-ups are impressions-weighted, so a small post with a
+  high share cannot outvote a large one.
+
 ## Directory Structure
 
 ```