30-day X plan, day by day (2026-10-12 → 2026-11-08)¶
Disclosure: product posts are about the maintainer's own products (OneClickTool), from each product page (checked 2026-10-11). Built on the real baseline in flow/analytics-pattern.md and the day in flow/a-day-on-x.md. Agent version of any day:
/social-x today.
Where we start (2026-10-11): 2K impressions / 28 days, 7 profile visits, 0 shares. Last 3 days ran 3–6 posts and 5–9 replies a day, and X put a temporary reach label on the account. So this plan lowers posts to 1–2 a day and moves the effort into replies, then grows slowly.
Targets for 2026-11-08 (goals, not a forecast): 10K impressions / 28 d · 50 profile visits · 15 bookmarks + shares
· utm_source=x link clicks tracked weekly · 0 labels.
How to read a day¶
- P1 = post at 08:00 · P2 = post at 17:30 (at least 4 h after P1).
- Product posts: hook letters from product-rotation.md; link in the first reply,
utm_content=<date>. - Talk posts: IDs from talk-bank.md (A = tip, B = question/poll, C = take, D = build log, E = numbers).
- R = replies by hand, split big accounts / peers / buyer sources (docs/06-reply-playbook.md).
- 🔗 = may drop one product link today in a "what are you building?" post, only if one passes the 6 checks in docs/06-reply-playbook.md.
- F = max new follows (people you replied to and would reply to again).
Day 0 · Sun 2026-10-11 (tonight, 45 min)¶
- Don't touch photo, name or @handle for 2 weeks (the checkmark is under re-review).
- Bio says who it's for + the hero offer; one link with
utm_medium=bio(links.md). - Pin the AgentBell demo (best fit for the dev audience) until a product post beats it.
- Build 3 private X Lists: Big (10 builders in AI coding / Mac / indie), Peers (10 accounts your size), Buyers (10 reviewers, newsletters, Mac/productivity accounts). Notifications on for 5 of them.
- Record the 4 clips for week 1 (list in week-02-mix.md).
- Copy last 28 days into templates/weekly-numbers.md as week 0.
Week 1 · Oct 12–18 · calm down, replies first (≈40 min/day)¶
Posts 1–2 a day · replies 8–10 · F ≤ 3. Content = week-02-mix.md (that file is "week 2 of posting").
| Day | P1 | P2 | R (big/peer/buyer) | 🔗 | Extra |
|---|---|---|---|---|---|
| Mon 12 | AgentBell B + clip | — | 5 / 3 / 0 | — | Answer every reply on P1 within 1 h |
| Tue 13 | Sweepy B + screenshot | B3 first Mac app | 5 / 2 / 2 | — | |
| Wed 14 | Edge Tabs B + clip | — | 4 / 3 / 2 | ✓ | Look for one "what are you building" post |
| Thu 15 | Markdown Everywhere B + map | C3 take | 5 / 2 / 2 | — | |
| Fri 16 | B4 poll | — | 4 / 3 / 2 | ✓ | Reply to every poll answer |
| Sat 17 | Redliner C | D4 tools I use | 5 / 2 / 2 | — | Record week-2 clips |
| Sun 18 | E1 week numbers + Sweepy ask | — | own replies only | — | Check 1 below |
Check 1 (Sun 18, 15 min): impressions, profile visits, link clicks per utm_campaign. Which product post got most
profile visits? → it becomes Monday's post in week 3 and the pinned post. Any label? → stay at week-1 volume one
more week.
Week 2 · Oct 19–25 · build the reply habit (≈50 min/day)¶
Posts 2 a day (Sun 1) · replies 10–12 · F ≤ 5. First week with a full P2 every day.
| Day | P1 | P2 | R (big/peer/buyer) | 🔗 | Extra |
|---|---|---|---|---|---|
| Mon 19 | AgentBell D (small settings) | A13 Esc in Claude Code | 6 / 3 / 2 | — | |
| Tue 20 | Sweepy C (byte for byte) | B1 read every diff? | 5 / 3 / 3 | ✓ | |
| Wed 21 | A1 hidden files tip | A9 git switch - |
5 / 3 / 3 | — | No product today: pure share-bait day |
| Thu 22 | Edge Tabs A (pain) + clip | C1 fewer clicks | 5 / 3 / 3 | — | |
| Fri 23 | B2 poll: where are your to-dos | A6 pbcopy |
5 / 3 / 3 | ✓ | |
| Sat 24 | Markdown Everywhere D (read-only) | D1 bug story (real) | 5 / 3 / 3 | — | Long post optional: A14 checklist as a thread |
| Sun 25 | E1 week numbers + Markdown Everywhere C (trial offer) | — | own replies only | — | Check 2 |
Check 2 (Sun 25): profile visits ÷ impressions ≥ 0.4%? If not, the profile is the leak: rewrite the bio's first line and the pinned post (social-x-profile-funnel). Drop 2 reply targets that never brought a visit, add 2.
Week 3 · Oct 26 – Nov 1 · double down on what worked (≈60 min/day)¶
Posts 2 a day · replies 13–15 · F ≤ 5. Monday = the product that won check 1.
| Day | P1 | P2 | R (big/peer/buyer) | 🔗 | Extra |
|---|---|---|---|---|---|
| Mon 26 | Winner of check 1, new hook (AgentBell E if it's AgentBell) | A12 CLAUDE.md | 7 / 3 / 4 | — | Pin today's post if it beats the old pin by Wed |
| Tue 27 | Redliner B poll | C4 undo = trust | 6 / 4 / 4 | ✓ | |
| Wed 28 | A3 copy path tip | A10 git log -S |
6 / 4 / 4 | — | Share-bait day |
| Thu 29 | Sweepy D (sort by month) | B6 feature nobody used | 6 / 4 / 4 | — | |
| Fri 30 | B5 longest agent wait | A8 Quick Look | 6 / 4 / 4 | ✓ | B5 replies → reply with AgentBell only if someone asks for a fix |
| Sat 31 | Long post: what 3 weeks of replies did (real numbers, E3) | D3 decision | 6 / 4 / 4 | — | |
| Sun Nov 1 | E1 week numbers + Sweepy ask | — | own replies only | — | Check 3 |
Check 3 (Nov 1): link clicks by utm_campaign. A product with zero clicks in 3 weeks → change its hook angle
(pain → number → question) or swap it out of the five.
Week 4 · Nov 2–8 · keep the best, measure money (≈60 min/day)¶
Posts 2 a day · replies 13–15 · F ≤ 5. Only the 2 best post formats from weeks 1–3, plus products.
| Day | P1 | P2 | R (big/peer/buyer) | 🔗 | Extra |
|---|---|---|---|---|---|
| Mon 2 | Best product of weeks 1–3, newest unused hook | Best talk format | 7 / 3 / 4 | — | |
| Tue 3 | Edge Tabs C (never steals focus) | B7 poll | 6 / 4 / 4 | ✓ | |
| Wed 4 | A4 screenshot location tip | C2 take | 6 / 4 / 4 | — | |
| Thu 5 | Markdown Everywhere: new hook from the fact bank | D2 before/after | 6 / 4 / 4 | — | |
| Fri 6 | Redliner D (mixed formats) | A7 mdfind |
6 / 4 / 4 | ✓ | |
| Sat 7 | Long post: "30 days posting on X as a solo Mac dev: the numbers" (real) | C5 take | 6 / 4 / 4 | — | |
| Sun 8 | E1 + one ask for the product with the best click → install rate | — | own replies only | — | Day-30 review |
Day-30 review (Nov 8, 30 min): add the row to the log in analytics-pattern.md;
compare with the targets at the top. Revenue or trials from utm_source=x vs. Premium cost
(docs/04-premium-30-day-plan.md). Then plan the next 30 days: keep the volume,
change only the weakest link (reach → profile → click → install → paid).
Every day, no matter what¶
The one number: profile visits, last 7 days, read each morning (targets: wk1 ≥ 5 · wk2 ≥ 8 · wk3 ≥ 11 · wk4 ≥ 14). How to read it: analytics-pattern.md.
[ ] morning: profile visits (7D) = ___ target ___
[ ] answer overnight replies [ ] P1 [ ] reply session 1 [ ] lunch answers + peers [ ] P2 [ ] reply session 2
[ ] 🔗 only if a post passes the 6 checks [ ] follows ≤ F [ ] one fact-bank line [ ] no hashtags unless a real community uses one
Stop signs (do this, not more posting)¶
| Sign | Do |
|---|---|
| Reach label again | Drop to week-1 volume for 7 days; no links in replies that week |
| Impressions up, profile visits flat | Fix bio + pin, not posts |
| Profile visits up, link clicks flat | Bio link / offer unclear; test the pinned post's first line |
| Clicks up, installs flat | The landing / store page, not X (launch-quick-download-wins) |
| Someone says "spam" under a link reply | Stop link replies for 2 weeks |