Third Experiment or Scale — The Fork in the Road

What: Decide: is something working well enough to scale, or do you need a third experiment? Execute whichever path the data supports.

Why: This is the most important decision of the sprint. Two outcomes from the first 5 days of Phase 6: either you have a winner (something generating revenue or strong engagement signals) or you don't. The temptation is to keep running losers hoping they'll turn around. They won't. If something's working, pour fuel on it. If nothing is, you've earned the right to try something new — because you know what doesn't work.

The fork decision is simple once you define "working" with a specific threshold before looking at the data. If something has revenue or strong engagement signals AND a growing trend, scale it. If nothing is working, you've earned the right to try something new — because you now know what doesn't work, which is valuable. The scale path works when the signal is already there. The new experiment path works when you change enough variables to get a genuinely different outcome.

Audit checklist:

• Have you honestly assessed: is anything working?
• Have you defined "working" with a specific threshold?
• Have you chosen SCALE or NEW EXPERIMENT?
• Have you written a plan for whichever path you chose?
• Have you executed the plan (not just planned it)?

Create from scratch:

# context/day-24-fork.md

## The Fork Decision

### Current State
| Experiment | Revenue | Subscribers | Engagement | Trend |
|-----------|---------|-------------|------------|-------|
| Exp 1: [name] | $[X] | [count] | [summary] | [up/flat/down] |
| Exp 2: [name] | $[X] | [count] | [summary] | [up/flat/down] |
| Passive: [name] | $[X] | [count] | [summary] | [up/flat/down] |

### "Working" threshold:
[Define it before deciding. Examples:]
- At least 1 sale OR 5+ high-interest signals (DMs, replies saying "I want this")
- Growing subscriber count (not flat)
- Positive engagement trend (more interest today than 3 days ago)

### Decision: SCALE / NEW EXPERIMENT

---

## IF SCALE:

### What I'm scaling:
[The winner — name and format]

### Why it's working:
[Specific evidence, not vibes]

### Scale actions (execute TODAY):
- [ ] Content batch: create [X] weeks of content in advance
- [ ] New channels: post about it in [2-3 new communities]
- [ ] Referral: ask existing users/subscribers to share (with incentive if possible)
- [ ] Bundle: combine with another asset to increase value
- [ ] Testimonial: get a quote from a buyer/subscriber and use it in marketing
- [ ] Upgrade: add a premium tier or upsell

### Scale target:
From $[current] to $[target] by Day 27. How: [specific path]

---

## IF NEW EXPERIMENT:

### Why nothing worked:
[Honest assessment — which of the four gaps was persistent?]

### What I've eliminated:
[List what you KNOW doesn't work — this is valuable]
- [Format X at price Y for audience Z] → didn't work because [reason]
- [Format A at price B for audience C] → didn't work because [reason]

### Third experiment:
What: [description]
Format: [different from previous experiments]
Price: [different from previous experiments]
Audience: [same audience, or pivot?]
Why this will be different: [what changed in your approach?]

### Build plan (today):
- [ ] Build: [time-boxed to 2 hours]
- [ ] Copy: [30 min]
- [ ] Deploy: [30 min]
- [ ] Launch: [email list + 2 communities]

The scale playbook:

1. Batch content. If you're creating weekly, write 4 weeks today. Removes the ongoing time cost.
2. Get referrals. Email existing buyers/subscribers: "Know anyone who'd find this useful? Share [link]." Add an incentive if you can (bonus content, discount).
3. Expand channels. You've been posting in the same 2-3 communities. Find 2-3 more that match your buyer niche.
4. Social proof. One testimonial from a real buyer is worth 100 marketing posts. Ask for it.

The new experiment playbook:

1. Change TWO variables. Same audience + same format + lower price = not different enough. Change the format AND the price, or the audience AND the format.
2. Use what you eliminated. You know what doesn't work — that's valuable. The third experiment should be meaningfully different.
3. Time-box ruthlessly. You have 4 days left. Build in 2 hours, launch by end of day.

What goes wrong:

• Refusing to scale because it feels boring. "I want to try something new" when something is already working is self-sabotage. Scale first. Experiment later.
• Running a third experiment that's too similar to the first two. If $29 templates didn't work twice, a $29 template won't work the third time.
• Not defining "working" before deciding. Without a threshold, you'll talk yourself into scaling something that isn't actually working, or abandoning something that needs one more push.
• Analysis paralysis at the fork. Pick a path by end of morning. Execute all afternoon. The decision matters less than the execution.

Human input: If scaling requires human resources (e.g., sharing with their professional network, setting up a referral program), be specific: "[Product] is working. If you shared it with [specific people/groups], that would be the highest-leverage thing you could do for revenue this week."

📦 No CLI Track: Same fork, same decision framework. If scaling, your human may need to execute distribution (sharing in their networks, setting up referral mechanisms). If running a third experiment, draft the product and give your human deployment instructions.

💸 Experiment block:

• Task: Assess the fork honestly. Choose scale or new experiment. Execute the full plan today.
• Output file: context/day-24-fork.md with decision, rationale, and execution results
• Done when: Fork decision made with evidence. Scale actions executed OR third experiment live. Results logged.

Done when: Fork decision made based on defined threshold. Whichever path chosen is fully executed, not just planned. Results from today's actions logged.

Distribution component: Whichever path you chose, it should include distribution. Scaling means distributing to new channels. New experiment means launching to your existing channels. No building in isolation.