Stackflow — Founder voice · B2B SaaS

Goal: Fix engagement drift · Cycle: Week of 26 May 2026 · 30-day lookback · LinkedIn + X + YouTube

Guardrails: Format diversity · voice · do-not-touch

2 patterns flagged

What's happening right now

Signals detected

Week of 26 May 2026 · 30-day lookback · triggered 12 min ago

2 Critical·3 Warning

LinkedIn engagement rate

1.9%↓38% below your normal

Critical

Talking-head video posts (LinkedIn)

4 postsAll below benchmark

Critical

X follower growth

+0.2%Flat vs prior 30d

Warning

YouTube watch time (post-mortems)

+12%↑12% vs prior 30d

Warning

Comment sentiment (one post)

32% negativeAbove your ceiling

Warning

LinkedIn talking-head video is underperforming across 4 posts in the window — engagement dropped before reach did, which rules out a pure distribution problem. YouTube post-mortems are the only format beating benchmark.

LinkedInXYouTube

Diagnosis

What went wrong · by platform

Fix engagement drift · LinkedIn + X + YouTube

DONE

LinkedIn engagement rate — talking-head vs carousel/text (30d)

Talking-head video

Carousel + text

Benchmark 3.1%

LinkedIn engagement rate fell to 1.9% account-wide — driven by four talking-head posts averaging 1.2% ER. Carousel and text posts in the same window held 3.4% (above your 3.1% floor). Reach per post was flat — engagement dropped before distribution, which rules out a pure reach problem.

Audience shift

Engagement concentrating on technical practitioners

Comments and saves shifted +11% toward segment B (technical practitioners) vs segment A (engineering leaders) over 30 days. LinkedIn talking-head posts targeted segment A — mismatch explains part of the engagement drop.

Plan

Recommended plan

82% diagnosis confidence · 91% plan score

Selected: Shift LinkedIn away from talking-head · boost YouTube post-mortems · hold X cadence · 7-day calendar with one experimental slot capped at 10%

View plan in Canvas →

Approve held drafts or edit the workflow on Canvas. Next cycle re-runs signals on the same brand memory.

If you want to verify

Venti separates what still matters from what should be archived or dropped entirely — so this cycle does not inherit noise from last month’s chat.

Still relevant

· Brand voice + 3 content pillars (persistent memory)
· Flagged signals only — not the full 30-day post dump
· Confirmed pattern: FORMAT FATIGUE on talking-head video (HIGH)
· Audience segment: engineering leaders on LinkedIn

Archived

· Q1 thought-leadership blitz · ended 30 Apr — kept as reference, not active strategy
· Carousel format · strong in March · available if we test again

Dropped

· 847 raw posts from the 30-day pull — dropped after signals stage
· Ruled out: platform-wide algorithm shift on LinkedIn
· Ruled out: combative competitor hook for engineering-leader segment
· Previous strategy doc: “post daily on all platforms” — superseded

Six inputs · how this cycle handles them

30 days of posts	Ingested once · only flagged posts carry forward
Engagement metrics	Compared to your benchmarks · raw tables dropped after signals
Audience data	Shift surfaced only when it explains a flagged signal
Brand guidelines	Always relevant — stored in brand memory
Campaign notes	Active campaigns only · none in this window
Previous strategies	Archived with date · ruled-out hypotheses dropped

Pull performance data
Dropped: 847 raw posts · Carried forward: 5 flagged signals
Analyze content patterns
Dropped: Ruled-out hypotheses · Carried forward: 4 confirmed patterns with confidence
Compare platforms
Dropped: Averaged cross-platform metrics · Carried forward: Per-platform diagnosis tabs
Identify audience shifts
Dropped: Full audience export · Carried forward: Segment shift explaining LinkedIn drop
Generate content calendar
Dropped: Dead formats + archived Q1 blitz · Carried forward: 7-day plan with guardrails
Schedule posts
Dropped: Unapproved calendar rows · Carried forward: 3 queued · 1 needs approval