Market Gate

---
name: market-gate
description: The ฿1,000,000,000 gate — bottom-up market sizing that kills ideas whose ceiling is too low, before a single line of code is written. Use before committing to any business, product, or pivot, and whenever someone asks "is this market big enough", "how big can this get", "should we build this", "ตลาดใหญ่พอไหม". Computes buyers × price × attainable share from cited inputs, ranks survivors by time-to-first-baht and time-to-฿1B, and writes explicit kill criteria into the thesis.
---

# THE GATE — if the ceiling is under ฿1B, we don't enter. No exceptions, no sentiment.

Building a ฿10M business and building a ฿1B business take roughly the same years, the same nights, and the same operator. The only difference is which market you walked into on day one. **This skill is where that decision is made, with arithmetic, before anyone falls in love.**

The gate tests the **ceiling**, not the forecast. The question is never "will we make ฿1B?" — it is **"could this market pay us ฿1B/yr if we won it?"** If the honest answer is no, we do not enter, however charming the idea is.

## The rule
```
CEILING = reachable buyers × realistic annual price × attainable share
GATE:  CEILING ≥ ฿1,000,000,000 / year   → may proceed
       CEILING <  ฿1,000,000,000 / year   → killed, logged, next candidate
```
No year limit on *reaching* it — a slow ฿1B market beats a fast ฿50M one. But among candidates that clear the gate, **speed wins every tie**: rank by time-to-first-baht, then by time-to-฿1B.

## BANNED (sizing theatre)
- **Top-down TAM.** "The global market is $47B, we only need 1%" is the single most reliable predictor of failure ever written into a deck. 1% of a market you cannot reach is 0%.
- **Uncited inputs.** Every number in the arithmetic needs a source or a stated, labelled assumption. "About 50,000 shops, probably" is not an input.
- **Counting people you cannot reach or bill.** The buyer must be reachable through a channel that exists and payable in a currency you can accept.
- **Retrofitting the numbers to save a favourite idea.** If you find yourself widening the definition of the buyer to clear the gate, the idea already failed — you're negotiating with a spreadsheet.
- **Confusing revenue with profit ceiling.** Check gross margin too: ฿1B of revenue at 8% margin is a logistics company with a founder's salary.
- **Ignoring the incumbent's grip.** Attainable share is not 30% when one player owns 80% and every buyer is on a 3-year contract.
- **Gating on feelings.** "It could be huge" is not a number. Run the tool.

## Process
1. **Define ONE buyer precisely.** Not "SMEs" — "accounting firms in Thailand with 3–20 staff." Precision is what makes a count possible.
2. **Count them, bottom-up, with a source.** Registry counts, directory sizes, platform seller counts, association membership, job-posting volume, app installs. Cite each.
3. **Price it from evidence.** What the incumbent charges, or what the manual alternative costs them per year. Not what feels affordable.
4. **Set an attainable share, and defend it.** 1–5% for a crowded market with strong incumbents; 10–20% for a fragmented one with weak ones; above 30% only if it's a niche you can genuinely dominate.
5. **Run the arithmetic** — [tools/market-gate.mjs](tools/market-gate.mjs) prints the ceiling, the gate verdict, the margin check, and the customer count required for the first ฿1M, ฿10M, and ฿1B.
6. **Widen or die.** If the ceiling fails, you get exactly one widening pass along the axes in [references/sizing-math.md](references/sizing-math.md) (geography, adjacent buyer, second product, up-market tier). If it still fails, kill it and log the reason to `venture/candidates.md`.
7. **Write the kill criteria** into `venture/THESIS.md` from [references/kill-criteria.md](references/kill-criteria.md) — the numbers that, if reality misses them, end this venture. Written *now*, while you're objective, not later when you're attached.

## Taking over something that already exists
Same arithmetic, two tiers: **฿1B is the entry gate** and the **path to $1B is stated alongside it, always** — not as a gate, as a direction that makes today's choices legible. Existing work moves nothing except *attainable share*, and only when the asset is real (an owned audience, an aged domain, a finished integration). See [../take-over/references/two-tier-gate.md](../take-over/references/two-tier-gate.md).

## The three checks every survivor must pass
1. **Ceiling check** — ≥ ฿1B/yr revenue potential from the arithmetic above.
2. **Margin check** — gross margin ≥ 60% for software/digital, ≥ 30% for service, ≥ 15% for physical. A ฿1B ceiling on 5% margin is not a ฿1B business.
3. **Path check** — a plausible sequence of customer counts, not a miracle: `first 1 → first 10 → first 100`, at a price already validated. If step one requires a miracle, so does step three.

## The sanity table (write this into the thesis, always)
| milestone | customers needed at ฿X/yr | how they're reached |
|---|---|---|
| first ฿1 | 1 | <the exact channel and the exact person> |
| ฿1M/yr | N | <channel> |
| ฿10M/yr | N | <channel — probably a different one> |
| ฿100M/yr | N | <what has to become true> |
| ฿1B/yr | N | <what has to become true — geography? second product? enterprise tier?> |

If the ฿1B row requires more customers than exist on the planet, the gate has failed regardless of what the ceiling arithmetic said.

## When to be honest about a "no"
Killing candidates is the job, not a failure of it. Nine of ten will die here. That's the gate working. The founder who tells the operator "this fails at ฿180M ceiling, here's the next one" on day two has saved them a year — that's worth more than any code written this week.

## Done check
- [ ] One buyer defined precisely enough to count
- [ ] Buyer count sourced, not guessed — with URLs in the thesis
- [ ] Price evidenced by an incumbent's price or a costed manual alternative
- [ ] Attainable share defended in one sentence about the competition
- [ ] `market-gate.mjs` run; ceiling, margin and path all pass
- [ ] Sanity table written into `venture/THESIS.md`
- [ ] Kill criteria written with numbers and dates
- [ ] Rejected candidates logged with the ceiling that killed them