Each MVP's Lifecycle — When to Grow and When to Kill

Not all MVPs are equal. Grow some, kill others fast.

4 projects, equal time, all ruined

Last year a developer ran 4 MVPs at once.

• Product 1: users growing ~100/month, almost no revenue
• Product 2: initial interest high but users dropped to zero in the last 3 months
• Product 3: deep niche, small but loyal fanbase
• Product 4: not even covering tool costs

By year-end, splitting time equally killed all four.

The next year he changed. Focus only on Product 1. Result: revenue 10x'd. The other three? Product 2 a graceful shutdown, Product 3 part-time, Product 4 killed immediately.

This post is simple. When running multiple MVPs, "the eye for how far to grow each one" is everything.

MVP lifecycle, 5 stages

1. Birth

From the first user. Through the 14-day build, ProductHunt launch or beta. Initial signals are extreme. Good = "smash hit," bad = "dead." But this stage's signals are noisy.

2. ValidationConfirm whether the initial signal is real. Out of the first 100, "who keeps coming back?" Core is not a raw number but repeated signal. 5 paying users renewing each month means more than 100 one-timers.

3. Growth

MAU growing 20%+. "How to grow faster" becomes the main question. Very energetic, but also very dangerous. Easy to be deceived. SNS posts cause growth, mistaken as product pull.

4. Maturity

Growth below 10%/month. Retention of existing users matters more than new ones. If you run multiple products, automate and minimize maintenance.

5. Sunset

Killing the product. Many makers avoid it, but sunset is also a new beginning.

Stage-by-stage decision criteria

Birth: signal clarity

If most of the first 100 used it once and left, it's curiosity. 10 active 3 weeks in a row signals real need. Simplest test: "if I hadn't built this, how would these people have solved their problem?"

Validation: retention & NPS

Retention: % returning a week after signup. <20% bad, 40%+ good, 60%+ strong. NPS = % promoters (9–10) minus % detractors (0–6). Both rising = validation success.

Growth: cause of growth

MAU +30%/month is good news. But know why it grew. Reliance on external channels (ProductHunt, SNS) may be ephemeral buzz. Real growth continues even when you stop pushing.

Maturity: profitability

Time-vs-revenue. 10h/month for 500,000 KRW/month = ~50,000 KRW/hour. Is that your desired hourly rate? If not, automate the revenue or sunset.

Portfolio management: BCG matrix variant

The classic BCG matrix is useful for solo makers. Two axes: market share (relative competitiveness) × market growth.

Star: high growth + high share

Like Lee's AI prompt tool — 50% monthly growth, dominant in its niche. Invest heavily in time and money. This is your future.

Cash Cow: low growth + high share

Park's marketing tool matured, 5–10% growth/month but stable 2M KRW/month. Keep maintenance minimal, automate, use the revenue to build the next Star.

Question Mark: high growth + low share

New market, low share, fast-growing market. Decide clearly: "make it a Star" vs "give up."

Dog: low growth + low share

Small market, small share. Sunset fast. Your time is too precious.

Grow or kill: data thresholds

1. Repeat-user ratio ≥ 20%

20+ of first 100 returning after a week = some need met. Below 20% → redesign or Kill.

2. Monthly active ≥ 50

50 is the minimum scale to say "this product helps people." Above → invest in growth; below → Pivot or Kill.

3. Monthly revenue ≥ 500,000 KRW

For your time investment to be worth it, you need some revenue floor. Below → automate or sunset.

4. Churn ≤ 5%/month

For mature products, 5%+ churn is a warning. Aging or a better competitor. Choose between feature updates, pricing revisit, or sunset prep.

These thresholds aren't absolute. Adjust to your situation. But without these, decisions become emotional, and emotional decisions create zombie projects.

5-stage diagnostic matrix

Update this matrix monthly and the portfolio is visible at a glance. What to grow, maintain, or kill becomes clear.

Wrapping up

Understanding each MVP's lifecycle is essential series-builder skill. Next post: how all this experience and data becomes an asset. Code goes legacy in years; your judgment, relationships, learned patterns last a lifetime.

Previous: Series-Builder Ops
Next: A Maker's Real Asset — What Remains When Code Is Gone

About the characters (Seoyeon Park, Junho Lee)
Characters are fictional personas. BCG matrix, 5-stage product lifecycle are real management theory.

Minchul Kim, CEO of Freeive, Fail School