Skip to content

Commit 7532553

Browse files
ivanivankaivanivanka
committed
feat(funnel): add funnel playbook, four funnel skills, and CLAUDE.md for v1.2.0
1 parent 0a2ec95 commit 7532553

15 files changed

Lines changed: 1948 additions & 1 deletion

File tree

‎CHANGELOG.md‎

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -6,6 +6,16 @@ This project follows [Semantic Versioning](https://semver.org/).
66

77
## [Unreleased]
88

9+
## [1.2.0] - 2026-04-04
10+
11+
### Added
12+
- funnel playbook with 8 reference sections: funnel types and decision matrix, offer stack architecture, VSL architecture, input specification, affiliate mechanics, high-ticket backend, market landscape
13+
- funnel-sprint skill: quick single-ICP mini-funnel generator (landing page, lead magnet, 3-email nurture)
14+
- funnel-plan skill: full funnel architecture session with funnel type selection and offer stack design
15+
- funnel-build skill: asset writing workflow covering VSL script, landing page, OTO pages, and email sequence
16+
- funnel-review skill: scored funnel audit across 5 domains with prioritized fix list
17+
- CLAUDE.md: repo-level release process and CI rules for agent sessions
18+
919
## [1.1.5] - 2026-03-28
1020

1121
### Fixed

‎CLAUDE.md‎

Lines changed: 121 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,121 @@
1+
# CLAUDE.md - markster-os
2+
3+
Governance for the `markster-exec/markster-os` private repo and its public mirror.
4+
5+
---
6+
7+
## Two-Repo Architecture
8+
9+
| Repo | Role |
10+
|------|------|
11+
| `markster-exec/markster-os` | Private source - where all work happens |
12+
| `markster-public/markster-os` | Public install source - what users install from |
13+
14+
`install.sh` hardcodes `markster-public/markster-os/master` as the install source.
15+
Never commit to the public repo directly. Always work in `markster-exec`, then mirror.
16+
17+
---
18+
19+
## Release Checklist (run every time before pushing)
20+
21+
### Step 1: Make your changes
22+
23+
Work on a branch if the change is non-trivial. Use conventional commits:
24+
25+
```
26+
feat(<scope>): <summary>
27+
fix(<scope>): <summary>
28+
docs(<scope>): <summary>
29+
chore(<scope>): <summary>
30+
refactor(<scope>): <summary>
31+
test(<scope>): <summary>
32+
ci(<scope>): <summary>
33+
perf(<scope>): <summary>
34+
```
35+
36+
Max 72 chars per commit subject. No other prefixes - CI will reject them.
37+
38+
### Step 2: Update CHANGELOG.md
39+
40+
- Add changes under `## [Unreleased]`
41+
- When releasing: move items into a new `## [x.y.z] - YYYY-MM-DD` section
42+
- Always keep an empty `## [Unreleased]` section at the top - CI requires it
43+
44+
### Step 3: Sync README.md version badge (most common failure point)
45+
46+
Find this line in README.md:
47+
48+
```
49+
[![Version](https://img.shields.io/badge/version-v1.1.5-blue.svg)](CHANGELOG.md)
50+
```
51+
52+
Update the version number to match the latest `## [x.y.z]` in CHANGELOG.md exactly.
53+
These two must always match. CI hard-fails if they do not.
54+
55+
### Step 4: Run the validator locally before pushing
56+
57+
```bash
58+
python3 tools/validate_markster_os.py
59+
```
60+
61+
Fix all errors before pushing. CI is a hard gate with no warn-only mode.
62+
63+
### Step 5: Commit and tag
64+
65+
```bash
66+
git add -p
67+
git commit -m "chore(release): prepare vX.Y.Z"
68+
git tag vX.Y.Z
69+
git push origin master
70+
git push origin vX.Y.Z
71+
```
72+
73+
### Step 6: Mirror to public repo
74+
75+
```bash
76+
git push git@github.com:markster-public/markster-os.git master
77+
git push git@github.com:markster-public/markster-os.git vX.Y.Z
78+
```
79+
80+
If a `public` remote is already configured: `git push public master && git push public vX.Y.Z`
81+
82+
---
83+
84+
## CI Hard Gates
85+
86+
The GitHub Action runs `tools/validate_markster_os.py` on every push and PR.
87+
On PRs it also validates every commit message. All failures block merge.
88+
89+
| Domain | What causes failure |
90+
|--------|-------------------|
91+
| Repo hygiene | `STATE.md` or `MASTER_LOG.md` committed; local absolute paths or home-directory shortcuts in any tracked file - exceptions: validator and validation-spec only (see `validation/validation-spec.md` for exact patterns) |
92+
| `company-context/` | Missing or extra files vs `manifest.json`; any file missing front matter keys (`id, title, type, status, owner, created, updated, tags`); headings missing or out of order |
93+
| `learning-loop/` | Missing required files; markdown without front matter; `canon/` containing `TODO`, `tbd`, `Speaker N:`, `Interviewer:` |
94+
| Release metadata | `## [Unreleased]` missing from CHANGELOG; README badge version does not match latest CHANGELOG version |
95+
| Commit messages (PRs) | Any commit not matching `<type>(<scope>): <summary>` or exceeding 72 chars |
96+
97+
---
98+
99+
## Files That Must Never Be Committed
100+
101+
- `STATE.md` - internal session state
102+
- `MASTER_LOG.md` - internal log
103+
- Any file containing local absolute paths or home-directory shortcuts (see `validation/validation-spec.md` for exact banned patterns) - unless editing the validator or validation-spec itself
104+
105+
---
106+
107+
## Versioning
108+
109+
Follows semantic versioning. Current version: see README badge and latest CHANGELOG entry.
110+
111+
- `patch` (x.y.Z): bug fixes, metadata sync, minor corrections
112+
- `minor` (x.Y.0): new features, new commands, new playbooks
113+
- `major` (X.0.0): breaking changes to workspace model or CLI contract
114+
115+
---
116+
117+
## Open Branches
118+
119+
Check `git branch -r` before starting work. Do not merge unknown branches without understanding what they are.
120+
121+
`feat-paperclip-bootstrap` on origin - not Ivan's work, leave it alone.

‎README.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
# Markster OS
22

33
[![Validation](https://img.shields.io/badge/validation-passing-brightgreen.svg)](validation/README.md)
4-
[![Version](https://img.shields.io/badge/version-v1.1.5-blue.svg)](CHANGELOG.md)
4+
[![Version](https://img.shields.io/badge/version-v1.2.0-blue.svg)](CHANGELOG.md)
55
[![MIT License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
66
[![PRs required](https://img.shields.io/badge/changes-PRs%20required-orange.svg)](CONTRIBUTING.md)
77
[![Security policy](https://img.shields.io/badge/security-policy-lightgrey.svg)](SECURITY.md)

‎playbooks/funnel/README.md‎

Lines changed: 67 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,67 @@
1+
# Funnel Playbook
2+
3+
Build conversion systems that match your traffic, offer, and price point.
4+
5+
A funnel is not a page sequence. It is a trust-manufacturing system. The wrong funnel type for your traffic temperature wastes budget and kills conversion before a single word of copy matters. Pick the funnel type first. Build the assets second.
6+
7+
---
8+
9+
## What this playbook covers
10+
11+
This is the reference library for conversion system design. It covers funnel type selection, offer stack architecture, VSL structure, affiliate mechanics, high-ticket backend design, and what has changed in 2025.
12+
13+
The four funnel skills all reference this playbook. Each skill has a defined scope. Run the right one for your situation.
14+
15+
---
16+
17+
## Skill Routing
18+
19+
| Skill | When to use | What it does |
20+
|-------|------------|-------------|
21+
| `funnel-sprint` | Need a functional funnel fast. Single ICP. No major architecture decisions. | Lead magnet + landing page + 3-email nurture. Done in one session. |
22+
| `funnel-plan` | Starting any major funnel. No plan exists yet. | Diagnoses situation, selects funnel type, designs offer stack with price ratios and conversion targets. |
23+
| `funnel-build` | `funnel-plan` is complete. Ready to write assets. | VSL script structure, landing page copy, OTO sequence, email nurture. Requires plan output as input. |
24+
| `funnel-review` | Funnel exists but performance is below target. Or you built one and want to check it before launch. | Grades funnel against research benchmarks: EPC readiness, offer stack ratios, traffic-funnel match, affiliate readiness. |
25+
26+
**Rule:** Run `funnel-plan` before `funnel-build`. Building assets before architecture is decided produces assets that need to be rebuilt.
27+
28+
---
29+
30+
## Prerequisites
31+
32+
All four skills check these before starting. Pass all four or stop and complete the missing item first.
33+
34+
- [ ] F1 complete -- ICP defined with a specific buying trigger (not a job title, a situation)
35+
- [ ] F2 complete -- Offer statement exists with outcome, mechanism, price, and guarantee
36+
- [ ] Proof exists -- At least one documented result with a real number attached
37+
- [ ] Mechanism named -- Your proprietary process has a name (not "our approach" or "our method")
38+
39+
**If mechanism is not named:** Name it before building the funnel. The VSL script, landing page headline, and OTO framing all depend on the mechanism name. Unnamed mechanisms produce generic copy that does not convert cold traffic.
40+
41+
---
42+
43+
## Reading Order
44+
45+
Read these sections in sequence before running any skill. Each one builds on the last.
46+
47+
| Step | File | Read when |
48+
|------|------|-----------|
49+
| 1 | `funnel-types.md` | Always. Pick your funnel type before anything else. |
50+
| 2 | `offer-stack.md` | Always. Price architecture determines funnel structure. |
51+
| 3 | `vsl-architecture.md` | If your funnel includes a VSL at any position. |
52+
| 4 | `input-spec.md` | Before writing any copy or building any page. |
53+
| 5 | `affiliate-mechanics.md` | If you are building for affiliate or JV distribution. |
54+
| 6 | `high-ticket-backend.md` | If your offer is $3,000 or above. |
55+
| 7 | `market-landscape.md` | Always. Platform behavior and proof requirements have shifted. |
56+
57+
---
58+
59+
## Reference files
60+
61+
- `funnel-types.md` -- funnel type taxonomy and decision matrix
62+
- `offer-stack.md` -- price architecture, OTO ratios, and order bumps
63+
- `vsl-architecture.md` -- VSL script structure, length, and pacing
64+
- `input-spec.md` -- data requirements before building any funnel type
65+
- `affiliate-mechanics.md` -- EPC, commission structures, and JV mechanics
66+
- `high-ticket-backend.md` -- application funnel design, call structure, close rates
67+
- `market-landscape.md` -- platform shifts, proof standards, and VSL position changes

‎playbooks/funnel/affiliate-mechanics.md‎

Lines changed: 101 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,101 @@
1+
# Affiliate Mechanics
2+
3+
How affiliate distribution works and what makes an offer affiliate-ready.
4+
5+
---
6+
7+
## What Affiliates Actually Evaluate
8+
9+
EPC (earnings per click) is the master variable. Serious affiliates will not commit list traffic below $2.00 EPC on email.
10+
11+
Secondary filters:
12+
13+
| Metric | Threshold | Notes |
14+
|--------|-----------|-------|
15+
| EPC | $2.00+ | Gate for committed affiliates |
16+
| Opt-in rate | 25%+ | On warm affiliate traffic |
17+
| Show rate | 75-85% (VSL calls), 30-40% (webinars) | |
18+
| Close rate | 35-45% | For VSL-fed funnels |
19+
| Refund rate | Under 10% | High refunds burn the affiliate's reputation, not yours |
20+
21+
Refund rate matters more than most offer owners realize. A 20% refund rate makes your offer un-promotable to serious affiliates regardless of EPC.
22+
23+
---
24+
25+
## Commission Structures
26+
27+
| Model | Example | Best for |
28+
|-------|---------|----------|
29+
| 30% recurring | Brunson / ClickFunnels | Long-term affiliate retention. Affiliates compound income with every new referral. |
30+
| 40-50% one-time | Kennedy Magnetic Marketing, Kern Mass Control | Short-term cash, strong initial recruitment. No retention incentive. |
31+
| 100% front-end | Hormozi-style loss-leader | Affiliate takes all front-end profit. Offer owner monetizes on backend. Strongest recruitment mechanic. |
32+
| Tiered JV | Launch-style with leaderboard | Competitive promotion behavior. Best for event-style campaigns. |
33+
34+
Recurring model wins on affiliate retention. Flat-fee one-time wins on initial recruitment for launch campaigns. 100% front-end commission wins on affiliate volume for backend-heavy offers.
35+
36+
---
37+
38+
## Stickiness: What Keeps Affiliates Promoting Long-Term
39+
40+
Three mechanisms:
41+
42+
1. **Recurring income that compounds.** Each new referral adds to their monthly base. Affiliates with 50+ active referrals do not stop promoting.
43+
44+
2. **Sticky cookie on the backend.** Affiliates get paid when their referral buys backend products they never pitched. Brunson's sticky cookie structure means an affiliate who drove a $7 book sale still gets commission on a $50K mastermind purchase two years later.
45+
46+
3. **Dream 100 / leaderboard dynamics.** Public leaderboard + competitive prizes triggers affiliates to promote harder against each other. Most effective during launch windows.
47+
48+
Kern's version: write the JV email swipes himself. Removed all friction from the affiliate's promotional decision. The affiliate copies, pastes, and sends.
49+
50+
---
51+
52+
## What Makes an Offer Affiliate-Ready
53+
54+
An offer is affiliate-ready when all five conditions are met:
55+
56+
- [ ] Fixed price - affiliates cannot model EPC on custom-priced offers
57+
- [ ] VSL or standalone sales page - measurable per affiliate link, no live scheduling required
58+
- [ ] Known EPC - from your own traffic data or a seeded test with a friendly affiliate
59+
- [ ] Low refund rate - under 10% to protect the affiliate's list reputation
60+
- [ ] Commission paid on the backend - sticky cookie or tiered structure so affiliates benefit from the full value ladder
61+
62+
**VSL is the dominant affiliate format** because it is standalone, measurable per link, and requires no scheduling. Webinars add friction. Application funnels require a phone call. VSL converts or it does not, which makes EPC modeling clean.
63+
64+
---
65+
66+
## Platform Split (2025)
67+
68+
Not all affiliate platforms are the same buyer or the same affiliate.
69+
70+
| Platform | Volume | Audience | Best for |
71+
|----------|--------|----------|----------|
72+
| impact.com | $120B GMV (2025) | B2B, SaaS, serious operators | High-trust B2B offers |
73+
| PartnerStack | $2.7B GMV (2025) | SaaS, B2B | SaaS and productized services |
74+
| JVZoo / WarriorPlus | High volume, low trust | Info-product buyers | Low-ticket info products ($17-$47) |
75+
| Direct JV deals | Varies | Curated lists | High-ticket offers where reputation matters |
76+
77+
JVZoo and WarriorPlus are not the same buyer as impact.com. If your offer needs to be your market reputation, the low-trust platforms have structural tension with that goal.
78+
79+
---
80+
81+
## The JV Launch Sequence (2025 Version)
82+
83+
Still the Product Launch Formula structure. Updates layered on top:
84+
85+
1. Pre-launch content (2-4 pieces over 7-10 days) - builds desire before cart opens
86+
2. $2-per-optin pre-launch incentives - pays affiliates to drive opt-ins before launch
87+
3. Cart open - first 24-48 hours drive 40-60% of total launch revenue
88+
4. Mid-launch content - case study or proof piece
89+
5. Cart close - final 24 hours drive 20-30% of total launch revenue
90+
91+
Close mechanics still drive 40-60% of total revenue. That has not changed since 2010.
92+
93+
AI layers now added: AI sales agents for follow-up, challenge-format pre-launch upsells.
94+
95+
---
96+
97+
## Reference files
98+
99+
- `funnel-types.md` - VSL funnel is the default for affiliate distribution
100+
- `offer-stack.md` - structure the price stack before setting affiliate commission rates
101+
- `market-landscape.md` - what has changed in the affiliate market

0 commit comments

Comments
 (0)