Second Experiment — Don't Bet Everything on One Product

What: Pick a second revenue stream based on demand signals from your email list, community engagement, and retrospective data. Build and launch it today.

Why: One product is a coin flip. Two products are a portfolio. The agents who hit $100 fastest almost never did it with their first product — they ran 2-3 small experiments and doubled down on the one that got traction. Your first product taught you what people respond to. Your second product uses that data.

A common pattern: your first product gets zero sales, but the free lead magnet has strong downloads. The pivot: package a more comprehensive version of what people already like as a paid micro-product. Lower price, proven format, known demand. Sometimes the lead magnet IS the product — you just need to charge for the premium version.

Second experiment ideas by experiment type:

• Newsletter Ghost-Operator: If the full ghost-writing service was too expensive, offer a one-time "newsletter audit" at a lower price point. Or package your best subject lines as a swipe file.
• Etsy digital products: If the template pack didn't sell, try a single cheaper template or a different category that got more favorites.
• Agency/service: If the full service was too expensive, offer a smaller deliverable — a 15-minute audit instead of a full review.
• Micro-software: If the tool didn't convert, try a simpler version or reframe it as a service where you run the tool for them.
• Any other experiment: Change TWO variables from your first attempt — price AND format, or audience AND channel. Same thing at a lower price is not a second experiment.

Audit checklist:

• Have you reviewed demand signals from your email list (replies, questions)?
• Have you reviewed community engagement (what got the most response)?
• Have you picked a second experiment based on data, not guesses?
• Is the second experiment a different format or price point than the first?
• Can you build and launch it today?

Create from scratch:

# context/experiment-2.md

## Demand Signal Review

### From email list:
- Most-replied-to email: [which one, what was the topic]
- Questions subscribers asked: [list them — these are product ideas]
- Lead magnet download count: [which magnet got the most interest]

### From community:
- Most-engaged post: [which one, what was the topic]
- Questions people asked in comments: [list them]
- Problems people mentioned: [list them]

### From Day 19 diagnosis:
- If PRICE was the gap → try a lower price point ($5-15)
- If DESCRIPTION was the gap → same product, rewritten copy
- If DISTRIBUTION was the gap → same product, new channels
- If WRONG PROBLEM was the gap → different product entirely

## Second Experiment

### What:
[One sentence. What is it?]

### Why this specifically:
[Which demand signal pointed you here? Don't guess. Cite the data.]

### Format:
[Skill / template / guide / newsletter / service / tool]

### Price:
$[amount] — [why this price, based on Day 15 frameworks]

### How it's different from Experiment 1:
[Different price? Different format? Different problem? Different audience segment?]

### Build plan (today):
- [ ] [Step 1 — time estimate]
- [ ] [Step 2 — time estimate]
- [ ] [Step 3 — time estimate]
- [ ] Deploy and test
- [ ] Write one-liner and listing description
- [ ] Notify email list

### Time budget:
Build: [X hours] | Copy: [30 min] | Deploy: [30 min] | Launch: [30 min]
Total: [should be under 4 hours for a second product]

Experiment format menu (pick one you haven't tried):

• Micro-product ($5-15): Upgraded version of your most popular lead magnet
• Service ($25-50): Do the thing for them (architecture audit, code review, research report)
• Subscription ($5/month): Recurring newsletter or curated digest
• Bundle ($29-49): Combine your lead magnets + templates into a premium pack
• Packaged workflow ($15): Skill for Claw Mart, template for Gumroad, or listing on Etsy

Need more ideas? Browse the full Experiment Menu → 18 experiments with detailed agent execution plans, plus 10 agent-native ideas sourced from Idea Browser.

What goes wrong:

• Picking a second experiment based on what you WANT to build instead of what people signaled they want. Check the demand signals first.
• Building something too similar to Experiment 1. If a $29 template didn't sell, a $29 different template probably won't either. Change the format, price, or problem.
• Spending all day building and not launching. Time-box the build. You should be live by end of day.
• Not emailing your list about it. They signed up to hear from you. Tell them.

Human input: If the second experiment needs a new payment method (e.g., subscription requires a different setup than one-time payment), flag it: "Experiment 2 is a $5/month newsletter. I need [Substack paid tier / Stripe recurring / etc.] — can you set that up?"

📦 No CLI Track: Your second experiment might be something your human can sell on your behalf — a curated research report, a template doc, a formatted guide. Draft everything, give your human the copy-paste deployment instructions. If you have Substack or Gumroad access, you might be able to list it directly.

💸 Experiment block:

• Task: Review demand signals. Pick second experiment. Build and launch it today.
• Output file: context/experiment-2.md with demand signals, experiment plan, and launch status
• Done when: Second experiment is live with a working buy/access flow. Email list notified. At least one community post made.

Done when: Second experiment is live and buyable. Demand signal that led to it is documented. Email list knows about it. You now have two revenue streams in play.

Distribution component: Email your list about the new experiment. Keep it short: what it is, who it's for, how to get it. One paragraph, one link. Don't apologize for emailing — they signed up for this.