Behind the Scenes: Pivoting Indie10k to MVP v2

Dogfooding hurts. I thought I was building a tool for indie devs, but when I ran Indie10k on myself last week, it felt like homework. Twenty giant missions stacked into a week, strategy check-ins... I wasn’t building momentum, I was down in the rabbit hole - each mission yielded to lots of homework and even more subtasks. I mean, they are not wrong, each obviously leads to growth but it is just too much.

So I tore it down. Two long nights with ChatGPT (arguing from first principles like we were trying to reinvent the wheel) and a little Claude Code magic later, I had something new. A pivot in 48 hours.

Work Backwards From the Invariant

The invariant doesn’t change: help indie devs hit profitable cash flow faster, with limited time and budget.

And honestly, only six levers really matter:

1. Does the idea even sniff product/market fit?

2. Can you get qualified traffic (volume × intent)?

3. Do people activate and convert?

4. Do they pay enough (ARPU / LTV)?

5. Do they stick around (retention)?

6. Can you iterate fast enough (cycle time)?

Everything else? Just noise.

Reduce to the Core Loop

At the core, building looks like this:

Hypothesis → Ship → Distribute → Measure → Decide.

And to run that loop, you don’t need a whole playbook of rituals. You need three things, max:

• Next Best Action (NBA): what’s the one thing I should do right now?

• Evidence Log: how do I know I actually did it, and what happened?

• Weekly Check-in: based on evidence, what’s the next experiment?

That’s it. The rest is bloat.

Question the Pillars

AI Coach

Developers don’t need a chat buddy. We already have ChatGPT. We need fewer wrong turns.

• Keep: a micro-coach that answers just two things:

1. What’s next?

2. How do I know it’s done?

• Cut: long chats, fake personas, emotional pep talks.

AI-Generated Missions

The truth? Nobody reads walls of AI text.

• Keep: a tiny library of battle-tested playbooks (SEO links, waitlist boosts, conversion lifts). Let AI adapt them to context.

• Cut: sprawling trees of tasks. Each “Action Card” should take ≤ 30 minutes.

Momentum

Building is a behavioral game. If rhythm breaks, people quit.

• Keep: daily Action Cards (check them off, feel progress). One Metric Sprint per week (stay focused). Optional community nudge.

• Cut: gamified system, elaborate point systems.

A Simpler, Better Model

Core Mechanics

1. One Metric Sprint (OMS): one metric per week. Traffic, signups, conversion—pick one.

2. Action Cards: 1–3 per day. Each with:

• Why (which lever it moves)

• What (≤ 3 steps)

• DoD (definition of done)

• Metric to watch

• Evidence slot (screenshot/link/number)

3. Evidence Library: your personal archive of what works (and what doesn’t).

And here’s the Rule Score I’ve been testing to rank actions:

Score = Impact (1–3) × Alignment (0–1) × Evidence Weight ÷ Estimated Time (hours)

If it scores high, it’s worth doing. If it doesn’t, kill it.

Before vs After

Before:

• 20+ missions per week

• Endless homework and subtasks that cannot be done in a short term

• Felt like PM software disguised as progress

After:

• 3 Action Cards per day

• 1 weekly metric to anchor focus

• Evidence log + Rule Score → a system that’s light enough to actually use

Why It Works

It cuts decision fatigue. It trims away the fluff. It turns your week into a rhythm instead of a burden. And maybe the most important thing—it forces you to build your own evidence-driven playbook over time. No guru required.

Call for Testers

The MVP v2 prototype is live. If you want to try the indie dev's growth gym system and free use all the future pro features during beta phase with me, hop on the waitlist:

👉 [indie10k.com](https://indie10k.com)