Lesson 77.1 told you to choose a platform by behavior, not hype. This appendix shows how to study any channel — search, social surface, newsletter niche, marketplace, community — in under an hour, without copying anyone. Observe, then decide.

How to observe without copying

Study five things, in this order. Take notes in plain words; screenshots with dates beat adjectives.

1. Audience. Who is actually here, in what moment? Read comments, reviews, and questions — not bios. What job brought them today?

2. Format. What shapes win here — long guides, 30-second demos, carousels, templates, live threads? Note length, pace, density, and what the first two seconds must do.

3. Norms. What does this place reward and punish? Disclosure habits, self-promotion tolerance, link behavior, reply expectations, moderation strictness. Violating norms burns faster than bad creative.

4. Competitors. Who already serves your overlap well? Name two. What promise, proof, and destination do they use — and what gap did they leave that a newcomer could fill honestly?

5. Destinations. Where do good posts send people — a tool, template, booking page, archive, signup? If every good post points to a pattern you cannot serve yet, the channel is not first.

Copying is mimicking the surface — same hook, same format, same claim. Learning is borrowing the mechanism — "this audience asks X before buying, so our destination must answer X first" — then building your own asset.

The worksheet

# CHANNEL-RESEARCH.md — [Project], date: ___

## Channel 1 — [name]
- Audience + moment: ___
- Winning formats (3 examples + dates): ___
- Norms (do / do not): ___ / ___
- Competitors (2) + gap left: ___
- Destination pattern + can we serve it: ___
- Verdict: pursue / test / defer + why: ___

## Channel 2 — (same fields)
## Channel 3 — (same fields)

## Decision
- Primary (why now): ___ / deferred + recheck date: ___

Three-channel notes: worked example

Channel 1 — Niche research newsletter sponsorships. Audience: solo analysts skimming Tuesday mornings for one usable check. Formats: single-idea issues with a filled example. Norms: disclosure expected, hard sell punished, replies answered. Competitors: two citation tools with generic "AI research" promises — gap is a worked filing example with sources shown. Destinations: template pages, not homepages — servable now. Verdict: test — one co-send next month.

Channel 2 — Short vertical video. Audience: curious scrollers, low intent, high volume. Formats: 20–40 second demos with on-screen proof. Norms: native pacing, no watermarked reposts, comments answered fast. Competitors: loud "AI hacks" accounts with no sources — gap is quiet, cited walkthroughs, but production cost is real. Destinations: follows and profile links, weak for activation. Verdict: defer — recheck after the template gallery compounds; listening only (77.4) for now.

Channel 3 — Workflow-tool template gallery. Audience: users mid-job, searching "brief template." Formats: preview plus filled example plus 60-second setup. Norms: current screenshots, support owner, changelog. Competitors: three stale packs abandoned at v1.0 — gap is maintained, dated updates. Destinations: in-app install plus optional email — servable. Verdict: pursue as the quarter's test route.

Verify fast

For each channel, cover the verdict and ask a peer to guess it from the notes alone. If pursue, test, and defer all sound identical, the observations are too vague — return to comments, reviews, and destinations until the differences are obvious.