The Early-Stage Startup Checklist

Since returning to Silicon Valley in July, I’ve felt the wave of AI startups building here. The energy reminds me of the early days of mobile internet — that same sense that something fundamental is shifting.

I’ve been meeting four or five startup teams every week.

AI is very different from mobile in terms of underlying business logic. But entrepreneurship itself hasn’t changed. The confusion, the temptation to move too fast, the uncertainty — those are constant.

That’s why I felt the need to write this down.

Over the years — from my own experience and from watching hundreds of companies — I’ve found that early-stage startups don’t usually fail because founders aren’t smart or hardworking. They fail because they cross stages too early. They mistake enthusiasm for validation.

So I put together this checklist. Not as theory — but as a way to stay structurally clear in the early days. To use stages to manage risk. To let evidence, not emotion, decide when to move forward.

---

Three Questions Before You Build Anything

Before refining the idea — before writing product specs — pause and ask:

1. Whose problem is this, exactly? Can you name the customer? Can you describe the situation in detail? Can the cost of the problem be measured?

2. Why now? Is there a real structural window — technological, regulatory, supply-side, distribution — that didn’t exist before?

3. Why you? Do you have asymmetric insight or resources? Data, access, experience, speed?

If you can answer these clearly on a single page, then it’s worth moving to the next stage.

If not, stay here.

---

Idea

From “interesting” to “worth spending years testing.”

The goal: Turn a spark into a hypothesis that can be proven wrong.

Give yourself 8–12 weeks. Do 20–30 real conversations — not polite chats, but deep interviews. Map out how decisions and payments actually happen.

Write down:

• Who discovers the problem?
• Who decides?
• Who pays?
• Who uses?

Look at what people do today. How do they solve it now? What does it cost them — time, money, reputation?

And set a clear failure line. If by week 8 or 12 you haven’t seen certain signals, you change direction.

At this stage, your job isn’t to make the vision bigger. It’s to make the problem smaller and sharper.

Example: Instead of “AI copywriting,” narrow it to “rewriting English landing pages for cross-border DTC sellers.” High frequency. Clear pain. Measurable impact.

---

POC (Proof of Concept)

Prove the core idea can actually work.

The goal here is simple: Does the key assumption hold?

Pick one killer metric. Just one.

Maybe it’s:

• 70% time saved compared to manual work
• 10% higher recall than existing solutions

Build the simplest experiment possible. Keep records. Make it reproducible. Failure notes are more valuable than flashy demos.

Also list your real risks:

• Data access
• Privacy compliance
• Supply constraints
• Technical bottlenecks

And ask: what happens if this assumption breaks?

The biggest mistake at this stage isn’t technical difficulty. It’s trying to prove too many things at once.

Example: A new image compression algorithm should first prove: On 1,000 images at target resolution, bitrate drops 30% without visible quality loss. Nothing more.

---

MVP (Minimum Viable Product)

Will someone actually use this?

Not say they would. Actually use it.

Build small. Narrow. Focused. It’s fine if the backend is partially manual. Speed of learning matters more than elegance.

Define the simplest closed loop: User arrives → experiences value → comes back (or pays).

Stay with one segment. One scenario. One channel. Go deep before you go wide.

Instrument everything:

• Activation
• Conversion
• D1 / D7 / D30 retention

Iterate every two weeks. Fix one core KPI at a time.

What matters isn’t that users are impressed the first time. It’s whether they come back a second time.

---

PMF (Product–Market Fit)

When the market starts pulling.

You’ll feel it.

You no longer have to push so hard. Retention stabilizes. Referrals happen. Growth doesn’t depend entirely on spending.

Some signals:

• 30-day retention holds at a healthy level (varies by category)
• NPS is high
• Organic users increase
• People can clearly describe the one thing you uniquely do

PMF isn’t about raising a bigger round.

It’s when customers arrive without being persuaded.

---

ARR

Making revenue predictable.

Now growth has to translate into recurring income.

Define your pricing clearly: Seats? Usage? Tiered packages?

Track:

• New ARR
• Expansion ARR
• Churn ARR

Watch CAC payback. Watch net revenue retention. Make sure paper growth doesn’t hide cash flow pressure.

ARR means stability. Not occasional large contracts.

---

Scale

Systems over heroics.

At this point, founder intensity is no longer enough.

You need:

• Clear roles
• Operating rhythm
• Playbooks
• Hiring standards
• Forecast accuracy
• Compliance discipline
• Financial buffers

Scaling speed is determined by how granular your standards are.

If top sales performance can’t be written down and taught, you’re not ready.

---

Stage Gates

Don’t jump early.

• Idea → POC: Clear problem statement + validation plan
• POC → MVP: Core metric proven and reproducible
• MVP → PMF: Real retention and organic pull
• PMF → ARR: Pricing clarity + healthy renewals
• ARR → Scale: Sales and operations are repeatable

Rule: Evidence first. Resources follow.

Before you cross to the next stage, thicken the proof in the current one.

---

Operating Rhythm

Every week, ask:

• What is our most critical assumption?
• What is the smallest way to test it?
• If we’re wrong, what changes next?

Every two weeks: Ship → Measure → Reflect → Decide.

Keep a one-page dashboard. North Star + 2–3 supporting metrics.

Clarity reduces anxiety.

---

Common Mistakes

• Starting too broad
• Over-validating too many variables at POC
• Subsidy-driven fake growth
• Expanding segments before PMF
• Ignoring cash flow while celebrating ARR

---

Final Thought

Early-stage startups are a fight against uncertainty.

This checklist isn’t meant to make you conservative. It’s meant to help you think clearly when emotions run high.

Break uncertainty into stages. Prove what needs to be proven. Scale only what has earned the right to scale.

When each step is backed by evidence, confidence compounds — for your team, your investors, and your customers.

From Idea to Scale — move forward, but move with proof.