The 90-Day Faceless Channel Plan (Week-by-Week)

Ninety days is the shortest honest window in which a faceless channel can produce a verdict. It's long enough for the algorithm to gather audience data (it needs roughly 8–12 videos to understand who to show you to), long enough for one full test-and-fix cycle on hooks and packaging, and short enough that you can't hide from the numbers. This plan assumes one long-form video per week plus 2–3 Shorts cut from it, the cadence that matches 2026 economics, where Shorts earn 3–14% of long-form RPM and exist to feed discovery (the data).

Print it, commit to the video count in advance, and don't evaluate "should I quit" before day 90: quitting inside the valley at month 4–6 is the #1 way these channels die (the five failure patterns).

Phase 1 (days 1–14): foundation, before any video exists#

Week 1: lock the niche. Run the validation sequence from the starter guide: demand check (small channels getting views = good), the 10-title test (ten clickable titles or walk away), and the RPM sanity check against the niche ranking. Deliverable: one niche, written down, with its RPM band.

Week 2: lock the two formats and the identity. Pick one main format and one experimental format from the format catalog. Set the visual identity now (caption style, color grade, intro sting under 2 seconds), because a deliberately consistent look is how a faceless channel becomes recognizable. Set up channel, banner, and description with the niche's primary keyword in the first sentence. Deliverable: two formats, one style guide, one channel.

Phase 2 (days 15–45): publish and learn the pipeline#

Weeks 3–6: four videos, one format. One long-form per week in your main format, each scripted on the 15-beat skeleton. These four videos exist to teach you the pipeline, not to perform: expect small numbers. Cut 2 Shorts from each video and post them between uploads. Read the retention graph 48 hours after each publish. If first-30-second retention is under 70%, rewrite next week's hook and nothing else. If there is a cliff at minute 2–3, strengthen beat 5's re-hook. One fix per week, always the earliest cliff first.

Week 5 checkpoint: compliance review. Watch your own four videos as a monetization reviewer would: is there visible human creative input, original analysis, visuals that comment rather than decorate? Fix the process now, not at a YPP rejection (policy explainer).

Phase 3 (days 46–75): introduce format B and optimize packaging#

Weeks 7–10: alternate formats, attack packaging. Introduce your experimental format on alternate weeks (A, B, A, B). By now you have 6–8 videos of CTR and retention data: if CTR sits under 3–4%, packaging is the bottleneck; rewrite titles and re-sketch thumbnails before scripting each video, and test one variable at a time. If CTR is healthy but retention bleeds mid-video, your beats 7–9 (midpoint reveal, complication) are weak: the script structure post has the fixes.

Day 60 milestone review. Compare formats head-to-head on 30-second retention and average view duration, not views. The format with better retention wins, even if the other got more views, because retention is what the algorithm compounds. Kill the loser, keep the winner, and define its repeatable template.

Phase 4 (days 76–90): consolidate and plan monetization#

Weeks 11–13: schedule lock and monetization math. Publish weekly on a fixed day. Run the monetization countdown: hours remaining to 4,000 ÷ your current watch hours per week = the weeks until your expected YPP date at current pace (the free monetization readiness calculator shows exactly what remains). Map the post-YPP revenue stack for your niche: affiliates first in high-RPM niches, sponsors from ~50k subs (earnings data). Decide your day-90 verdict with pre-set criteria: continue if average view duration is above ~35% and climbing, first-30s retention above 65%, and at least one video beat your channel median by 2× (proof the algorithm can distribute you). Pivot the niche only if all three failed despite packaging and hook fixes, and note that "pivot" means a relaunch, not a tweak.

The whole plan in one table#

Want the review loops automated? Unfaced audits your channel's retention every 6 hours and flags the exact beat where each video bleeds: the whole phase-2/3 diagnostic, without the spreadsheet. Start with the free channel audit.