raress96/fundamentals.md

Last active March 11, 2026 19:47

Star (1) You must be signed in to star a gist
Fork (2) You must be signed in to fork a gist

Select an option

Learn more about clone URLs
Clone this repository at <script src="https://gist.github.com/raress96/58ce6de7c3cf6809293a5fc19c72aeff.js"></script>
Save raress96/58ce6de7c3cf6809293a5fc19c72aeff to your computer and use it in GitHub Desktop.

Download ZIP

Claude Code Distribution Framework Idea Validator agent (part of https://beyondfolder.com/distribution)

Raw

fundamentals.md

Growth Fundamentals

For developers/engineers who can build products but have no sales or marketing experience.

Product-Market Fit (PMF)

The #1 reason startups fail: building something nobody wants.

Signs you DON'T have PMF:

You have to convince people they have the problem
Users try once and never return
Growth only happens when you push (ads, outreach)

Signs you HAVE PMF:

Users tell others without being asked
People get upset when it breaks
Demand exceeds your capacity

Goal: Find PMF by talking to real leads before scaling.

The Mom Test (How to Ask Questions)

People lie to be nice. Ask about facts, not opinions.

Bad (lies)	Good (truth)
"Would you use this?"	"How do you handle X today?"
"Is this a good idea?"	"What happened last time X occurred?"
"Would you pay for this?"	"How much time/money did that cost?"

Rule: Talk about their life, not your idea. If they haven't tried to solve it, it's not a real problem.

The Sales/Marketing Funnel

The "funnel" visualizes the customer journey. Many enter the top, few reach the bottom.

TOFU (Top)     │  AWARENESS      │  Cold Lead   │  See your message
               ▼
MOFU (Middle)  │  INTEREST       │  Warm Lead   │  Reply, engage
               ▼
BOFU (Bottom)  │  DECISION       │  Hot Lead    │  Want to buy
               ▼
               │  CUSTOMER       │  Paying      │  They pay you

TOFU, MOFU, BOFU

Term	Focus	Activities
TOFU	Reach many people	Cold outreach, content, ads
MOFU	Build trust	Demos, case studies, webinars
BOFU	Close deals	Trials, sales calls, pricing

The Funnel Math

1,000 contacts (TOFU)  →  100 reply (10%)  →  20 watch demo  →  5 trial  →  2 customers

0.2% conversion is normal for cold outreach. To get 10 customers, need ~5,000 contacts.

Leaky Funnel Problems

Problem	Symptom	Fix
Leaky TOFU	<5% reply rate	Wrong audience or messaging
Leaky MOFU	Replies but no demos	Too much friction
Leaky BOFU	Demos but no trials	Pricing or pain mismatch

Fix leaks top to bottom. No point optimizing demos if nobody replies.

Sales vs Marketing

At the most basic level:

Sales = just helping people
Marketing = sharing your story

	Marketing	Sales
Scale	1-to-many (content, ads, SEO)	1-to-1 (calls, DMs, emails)
Goal	Generate awareness	Close deals, get revenue
Examples	Blog posts, social media, ads	Cold emails, demos, negotiations

For early founders: Start with sales—it doesn't scale, but you learn fast. Talking to 50 people teaches more than 5,000 blog visitors. Marketing comes later when you have a proven message.

Other Fields Technical Founders Should Know

Field	What It Is	Key Insight
Customer Discovery	Validating people want your product before building	Talk to customers first, build second.
Lead Generation	Finding potential customers + contact info	See `../strategies/lead-sourcing.md`
Copywriting	Writing that persuades action	Talk about their problems, not your features. Shorter is better.
Positioning	How you describe your product vs alternatives	See Positioning section below
Pricing	What to charge	Price on value delivered, not your costs. If you save them $500/mo, $50/mo is easy.
CRO	Improving conversion rates	Doubling conversion = doubling customers without more leads
Customer Success	Helping customers get value after purchase	Happy customers refer others; unhappy customers churn

Positioning (Your Headline Is 90% of Your Marketing)

Most technical founders describe what their product does. That's wrong. Describe the one outcome your buyer obsesses over.

The Rule: One Metric, Not Features

Your audience cares about one thing. Find it and make it your entire positioning.

Approach	Example	Result
Bad: Feature-descriptive	"Analytics tool with real-time dashboards"	Sounds like 50 other tools
Good: Outcome-obsessed	"See where your revenue actually comes from"	Gut punch of recognition

DataFast didn't grow as "another analytics tool." It grew when positioned as revenue attribution — the one metric founders actually cared about.

How to Find Your One Metric

What does your user complain about at 1am? Not "I need better analytics" but "I have no idea which channel is making me money"
What lie are they telling themselves? "I know my MRR" (but they don't know their profit). "My product is great" (but nobody knows it exists)
What's the gap between what they track and what actually matters? Revenue vs profit. Traffic vs conversions. Features vs retention

The Gut Punch Test

Read your headline out loud. If your target customer would pause and think "shit, that's me" — it works. If they nod politely and keep scrolling — it's too generic.

Positioning Formats

Format 1: Category reframe (headline/tagline) Take your generic category and prefix it with the one outcome that matters. Two or three words that reframe what the product is about.

"Analytics" → "Revenue-first analytics" (DataFast)
"Financial dashboard" → "Profit-first finances"
"Project management" → "Shipping-speed project management"

This works for headlines, taglines, and anywhere you have under 5 words. It tells people immediately how you're different from every other tool in the category.

Format 2: Gap reveal (subheadlines, email openers, social posts) Call out the gap between what they think they know and what they actually don't.

"You know your MRR. You don't know your profit."
"You track traffic. You don't know which visitors pay."
"You ship features. You don't know which ones retain users."

This works for longer copy where you have room to create tension. It's a gut punch that makes people want to close the gap.

Use both together: Category reframe as the headline, gap reveal as the subheadline. Example: "Revenue-first analytics" (headline) + "See where your money actually comes from" (subheadline)

Common Positioning Mistakes

Mistake	Why It Fails	Fix
Listing features	Nobody cares about features, they care about outcomes	Lead with the outcome, features are proof
Being too broad	"For everyone" means for nobody	Pick the narrowest audience that still makes a business
Copying competitors	You'll always be the worse version of them	Find the angle they ignore
Being too calm	"See your data in one place" doesn't create urgency	Call out the pain they're avoiding

Onboarding & the Aha Moment

The aha moment is when the product clicks — the user suddenly gets why this exists. Everything before that moment is friction. Everything after is retention.

Speed to Aha = Conversion

Shorter onboarding → less friction → more dopamine → higher conversions

DataFast's aha: "Google → $1,240 revenue" appearing on the dashboard. The product made sense instantly. Every step before that moment was a leak.

How to Design for Aha

Define the aha moment in one sentence. "The user sees [specific thing] and immediately understands the value." If you can't define it, your product might not have a clear one yet
Count the steps to get there. Sign up, connect data, configure settings, wait for processing, navigate to dashboard... every step is a drop-off point
Remove steps ruthlessly. Can you pre-fill instead of asking? Can you show a preview before requiring setup? Can you skip the tutorial entirely?
Use smart defaults over empty states. Pre-fill common options (popular SaaS tools, industry templates, sensible settings). Let users toggle/adjust instead of typing from scratch

Onboarding Anti-Patterns

Anti-Pattern	Why It Kills Conversion	Do This Instead
Multi-step setup wizard	Users quit before seeing value	Get to the aha first, configure later
Onboarding tour / walkthrough modals	Nobody reads them, they click "skip"	The product should explain itself by doing the thing
Requiring full data before showing anything	All effort, no reward	Show partial results immediately, improve as they add more
Email verification before access	Adds friction before any value	Let them in, verify later

The Free Tier IS the Onboarding

Your free tier should be generous enough that users hit the aha moment naturally. If the free tier is too restrictive, users leave before understanding the product. The paywall should feel like unlocking more value, not hitting a wall right when things start working.

Growth Loops (Build Features Worth Sharing)

Traditional marketing: you push your product to people (outreach, ads, SEO). Growth loops: you build something people pull others toward.

Anti-Marketing Principle

Build features people want to show off. If users screenshot your product and share it unprompted, you never have to "do marketing."

DataFast built a real-time visitor map that looked cool. Users shared screenshots → people asked what tool it was → organic signups. The feature was the marketing.

Types of Growth Loops

Loop	How It Works	Example
Shareworthy feature	Users show off a visual/output from your product	Analytics maps, stats cards, dashboards, "wrapped" summaries
Public by default	User's output is visible to others, each one a landing page	Public dashboards, portfolio pages, shared reports
Creative stunts	Build something fun on top of your own product	DataFast's Severance-themed analytics (6-hour side project, drove hundreds of users)
User-generated distribution	Give users tools to build things with your data	APIs, embeds, integrations that carry your branding

How to Design a Shareworthy Feature

It must look good. Not just functional — visually impressive enough that someone wants to post it. Screenshots are your unpaid ads
One click to share. If generating/sharing the output takes more than one click, most users won't bother
Subtle branding. The output itself carries your product name — not obnoxiously, but enough that people ask "what tool is that?"
Serves the user first. The shareable thing must be genuinely useful to the person sharing it. If it only serves your marketing, nobody will use it

The Stunt Playbook

Small creative projects built on your own product can drive outsized attention:

Build something playful or unexpected using your own API/data
The goal is to make people think "that's cool, what powers it?"
Time-box to hours, not weeks. The ROI comes from novelty, not polish
Each stunt is a distribution channel that costs nothing but time

Growth Channels

Channel	Speed	Cost	Scale	Learning
Cold outreach	Fast	Low	Limited	High
Content/SEO	Slow	Low	High	Medium
Paid ads	Fast	High	High	Low
Referrals	Medium	Free	Medium	High

Start with cold outreach for fast learning, add scalable channels later.

Cold, Warm, Hot Leads

Funnel = the process (stages) Leads = the people (categorized by readiness)

Lead Type	Funnel Stage	Definition	Conversion
Cold	TOFU	Don't know you	1-5%
Warm	MOFU	Engaged (replied, watched demo)	10-30%
Hot	BOFU	Want to buy (asked for pricing/call)	30-70%

Status tracking:

Status	Temperature
🔍 Researched	Cold
📤 Outreach Sent	Cold
💬 Responded	Warm
📞 Call Scheduled	Hot
✅ Converted	Customer
❌ Not Interested	Dead

The math: 100 cold → 5-10 replies → 1-3 interested → 1 customer. Volume + follow-ups matter.

Churn (The Silent Growth Killer)

Churn rate = % of customers who cancel per month.

Why Churn Matters

5% monthly churn = 46% annual loss

Year 1: Add 120 customers, lose 55 to churn = net 65
vs.
2% monthly churn = 22% annual loss

Year 1: Add 120 customers, lose 22 to churn = net 98

Same acquisition effort, 50% more customers with lower churn.

Churn Benchmarks

Monthly	Annual	Verdict
<2%	<22%	Excellent
2-3%	22-31%	Good
3-5%	31-46%	Average
5-7%	46-58%	Concerning
>7%	>58%	Fix before scaling

SMB products typically have 5-7% (higher than Enterprise 1-2%).

Types of Churn

Type	Cause	Fix
Voluntary	Customer cancels	Better onboarding, more value
Involuntary	Payment fails	Dunning emails, retry logic (20-40% of churn!)

What Causes Churn & How to Fix

Cause	Signal	Solution
Never got value	Low usage	Better onboarding (Day 1-7 critical)
Stopped using	Usage dropped	Re-engagement, check-ins
Found alternative	Competitor mentioned	Better positioning
Budget cut	"Too expensive"	Show ROI, annual discount

How to Reduce Churn

Onboarding (biggest impact)—get to "aha moment" fast
Track usage—reach out before they leave
Exit surveys—find patterns
Fix involuntary—payment retries, expiring card emails
Annual plans—discount for yearly, locks in customers

Raw

idea-validator.md

name	description	tools	model
idea-validator	Validates product ideas through desk research - competitors, market signals, pain evidence, and scoring. Use when the user wants to validate an idea before building.	Read, Grep, Glob, Bash, WebSearch, WebFetch, Write, Edit	opus

Idea Validator Agent

You are a market validation researcher for bootstrapped solo developers. Your job is to do thorough desk research on a product idea and determine whether it's worth spending time on community validation.

Your Constraints

The user you're helping is:

Solo developer — No team, must be buildable by one person
Fully bootstrapped — No VC, max $200/month infrastructure
MVP timeline: ~2 weeks — Must be simple enough to ship fast
Limited time: 10-20 hrs/week — Every hour spent validating is an hour not building
Revenue goal: $0 → $10k MRR — Focus on clear monetization paths

Process

Step 1: Understand the Idea

Get the idea from the user's prompt or description.

Clarify if needed:

What does it do? (one sentence)
Who is it for?
What's the price hypothesis?

Step 2: Load the Validation Framework

Read validation.md for:

The scoring criteria and weights
Phase 1 desk research requirements
The VALIDATE.md template structure

Step 3: Research Competitors

Use WebSearch to find:

Direct competitors — Tools solving the exact same problem
Indirect competitors — Adjacent tools, manual processes, spreadsheets
Pricing — What do existing solutions charge?
Weaknesses — Read reviews on G2, Capterra, Reddit for complaints

Search queries:

"[problem]" tool OR software OR app
"[competitor name]" review OR alternative OR sucks
site:g2.com "[competitor]" cons OR negative
"[competitor] alternative" reddit
"[competitor] pricing"

Step 4: Research Market Signals

Use WebSearch to find evidence of pain:

site:reddit.com "[problem keywords]" frustrated OR annoying OR hate
site:reddit.com "[problem keywords]" "any tool" OR "looking for"
"[problem]" "I wish" OR "why isn't there"
site:news.ycombinator.com "[problem keywords]"
"[problem keywords]" indiehackers.com

Look for:

Frequency — How often do people complain about this?
Recency — Are these recent posts (last 12 months)?
Intensity — Are people mildly annoyed or genuinely frustrated?
Specific quotes — Save exact words people use (these become marketing copy)

Step 5: Assess Market Size & Accessibility

Research:

How many businesses/people have this problem?
Where do they hang out online? (communities, forums, LinkedIn groups)
Are they easy to find and contact?
Is the market growing or shrinking?

Step 6: Score the Idea

Apply the scoring framework from the validation strategy:

Criterion	Weight	What to assess
Pain evidence	3x	Complaints, frustrations, "I wish" signals found
Willingness to pay	3x	Do people already pay for solutions? At what price?
Market size	2x	Total addressable market for a solo dev
Market accessibility	2x	Can you find and reach these people?
Competitive gap	2x	Is there a clear angle vs existing solutions?
Build feasibility	2x	Can you ship MVP in 2 weeks on Cloudflare/Nuxt?
Revenue potential	1x	Clear pricing model, realistic path to $2-5k MRR
Personal fit	1x	Do you experience the pain? Have domain knowledge?

Score each 1-5, apply weights, total out of 80.

Be honest and critical. The point is to avoid wasting weeks building something nobody wants. A low score now saves months later.

Step 7: Write Recommended Validation Steps

Based on the research, suggest:

3-5 communities to post research questions in (with specific subreddits/forums)
A post angle — What question to ask that surfaces pain naturally
Who to DM — Type of person most likely to respond to discovery calls
Landing page angle — Headline + subheadline for a quick waitlist test

Output

File Output

Use the template from product/templates/VALIDATE.md.

Single idea: Write directly to product/VALIDATE.md
Multiple ideas: Write each to product/validations/[idea-name].md and create/update product/VALIDATE.md as an index linking to each file

Create the product/validations/ directory if it doesn't exist.

Fill in all Phase 1 sections (desk research). Leave Phase 2 sections empty — the user fills those after community validation.

Summary Output

After creating the file, provide a concise summary:

## [Idea Name] — Validation Summary

**Score:** X/80 ([Strong/Promising/Weak/Kill])
**One-liner:** "[pitch]"

**Strongest signals:**
- [Top positive finding]
- [Second positive finding]

**Biggest risks:**
- [Top concern]
- [Second concern]

**Competitors:** [Name] ($X/mo), [Name] ($X/mo), [Name] ($X/mo)

**Recommended next steps:**
1. [Most important action]
2. [Second action]
3. [Third action]

Handling Multiple Ideas

If asked to validate multiple ideas:

Research each idea separately
Create a validation file for each in product/validations/
Provide a comparison summary at the end ranking all ideas by score
Recommend which ideas to take to community validation (Phase 2)

Important Guidelines

Be brutally honest — A killed idea saves weeks of wasted building
Show your sources — Include links or specific references for every finding
Quantify when possible — "3 Reddit threads" not "some people complain"
Don't over-research — 1-2 hours max per idea, then score and move on
Compare to existing user context — If the user has products, note synergies or conflicts
Use the scoring framework — Don't skip criteria or invent new ones

Raw

validation.md

Idea Validation Strategy

Validate product ideas before building. Desk research + community signals tell you whether to build, pivot, or kill — in days, not weeks.

Why Validate

Building an MVP takes 2-4 weeks. Validation takes 3-7 days. Wrong bet = months wasted on something nobody wants.

Signs you skipped validation:

Built for weeks, launched to silence
"I think people want this" (no evidence)
Competitors exist but you don't know why yours is different

See ../concepts/fundamentals.md → "Product-Market Fit" for deeper context.

Two-Phase Process

Phase 1: Desk Research (1-2 hours, agent-assisted)

The idea-validator agent handles this. It researches and fills product/VALIDATE.md (single idea) or creates per-idea files in product/validations/ (multiple ideas).

Research Area	What to Find
Competitors	Who exists, what they charge, their weaknesses
Market signals	Reddit/forum complaints, "I wish" posts, negative reviews
Willingness to pay	Do people pay for existing solutions? At what price?
Market size	How many potential customers? Growing or shrinking?
Build feasibility	Can a solo dev ship an MVP in 2 weeks on <$200/mo infra?

Phase 2: Community Validation (3-7 days, manual)

You do this. Post in communities, talk to real people, collect signals.

Activity	Goal	Time
Research posts	Post genuine questions in 3-5 communities	Day 1
Engage	Reply to every comment, note exact words used	Day 2-3
DM interested people	"Would you have 10 min for a quick chat?"	Day 2-3
Discovery calls	3-5 calls per idea minimum	Day 3-5
Landing page test	Simple page with waitlist CTA	Day 5-7

See customer-discovery.md for call scripts, outreach approach, and the Mom Test methodology from ../concepts/fundamentals.md.

Scoring Framework

Rate each idea 1-5 on these criteria:

Evidence Criteria (Does the pain exist?)

Criterion	1 (Low)	5 (High)	Weight
Pain evidence	No complaints found	Multiple people actively frustrated	3x
Willingness to pay	Free alternatives satisfy everyone	People pay $20+/mo for inferior solutions	3x

Market Criteria (Can you reach them?)

Criterion	1 (Low)	5 (High)	Weight
Market size	<1,000 potential customers	50,000+ potential customers	2x
Market accessibility	Hard to find, scattered	Concentrated in known communities	2x
Competitive gap	Crowded, strong incumbents	Few/weak solutions, clear angle	2x

Feasibility Criteria (Can you build and sell it?)

Criterion	1 (Low)	5 (High)	Weight
Build feasibility	Months of work, complex infra	2-week MVP, simple stack	2x
Revenue potential	Hard to monetize, low ARPU	Clear pricing, $20+/mo ARPU	1x
Personal fit	No domain knowledge, no interest	Experience the pain yourself	1x

Scoring Template

Criterion	Weight	Total
Pain evidence	3x
Willingness to pay	3x
Market size	2x
Market accessibility	2x
Competitive gap	2x
Build feasibility	2x
Revenue potential	1x
Personal fit	1x
TOTAL		/80

Score Interpretation

Score	Meaning	Action
65+	Strong	Validate with community, likely worth building
50-64	Promising	Validate with community before committing
35-49	Weak	Deprioritize unless community signals surprise you
<35	Kill	Move on to next idea

Parallel Validation

When you have multiple ideas, validate them simultaneously.

Process

Desk research all ideas — Run the idea-validator agent on each
Rank by desk research score — Pick top 2-3 ideas
Run community validation in parallel — Different communities, same timeframe
Compare signals after 5-7 days — Use the decision rules below

Decision Rules

Clear winner (>2x the signals): Build that one.

Both strong (similar signals): Pick the one where:

Calls revealed higher emotional pain
People mentioned specific dollar amounts they'd pay
You personally feel more excited to build it

Both weak (<5 waitlist signups each): Neither validated. Go back to idea list.

One strong, one weak: Build the strong one. Save the weak one.

Research Posts (Templates)

Adapt these to your idea. Post genuinely — you're researching, not selling.

Reddit/Forum Post

Title: How do you handle [problem area]?

[Describe the frustration in 2-3 sentences using your own experience or a relatable scenario.]

What's your process? Have you tried any tools for this? What actually works?

LinkedIn Post

[State the problem in a punchy opening line.]

[2-3 specific examples of the frustration.]

For [target audience] — how do you handle this?
Genuinely curious — [tool? spreadsheet? manual process?]

DM After Engagement

Hey [name], thanks for your reply about [topic]. I'm actually
exploring building a tool to solve this. Would you have 10 min
for a quick chat? No pitch, just trying to understand the problem better.

Connecting to the Framework

Validation is the first phase in the product lifecycle:

idea-finder agent → ideas backlog
                         ↓
              idea-validator agent → product/validations/*.md
                         ↓
                    Decision: build / kill / pivot
                         ↓
              new-product-setup agent → product/PRODUCT.md
                         ↓
              ICP evaluation → outreach → growth

After Validation

If building: Run new-product-setup agent to create product/PRODUCT.md, evaluate ICPs, and plan distribution.

If killing: Document why in the validation file. Move idea to bottom of backlog.

If pivoting: Update the hypothesis and re-run desk research.

Common Mistakes

Mistake	Why It's Bad	What to Do Instead
Building before validating	Weeks wasted on unvalidated ideas	Spend 3-7 days validating first
Asking "would you use this?"	People lie to be nice (Mom Test)	Ask about their current process and pain
Only doing desk research	Desk research is educated guessing	Talk to real people
Falling in love with idea before data	Confirmation bias	Let the signals decide
Validating too long	Perfectionism disguised as research	7 days max, then decide