The pipeline
A run has seven steps. The sanity check runs before the simulation. A nonsense idea stops at step 02. Click a step to read its inputs, outputs, and gaps.
Monte Carlo
The engine tests three channels under four market conditions. It runs 1,000 trials for each pair. Each trial samples the rates and runs 12 weeks. Click a cell to see the condition.
The three channels
Outreach and community use founder time. Paid uses money.
Saturation is one number. It changes three things at once. The engine uses 0.40 by default. Research can adjust it.
The judgment layer
Three parts. The decision engine makes the call. The writing engine drafts the words. The research layer supplies the evidence. No customer number is invented.
Decision engine
Reads the evidence and judges the idea, the market, and each approach against a fixed set of criteria. Returns scores and confidence. The arithmetic is done in code.
Decision layerWriting engine
Turns the judgment and the evidence into the recommendation, the honest read, and the next step. It does not calculate.
Writing layerResearch layer
Runs three parallel searches — competitors, pricing, and reviews/alternatives — and returns full page content for the strongest results.
Web searchDecision questions
How the verdict is made
Invented numbers: zero. The decision engine returns categories and probabilities. Code maps the saturation score to a bounded number between 0.15 and 0.90. The engine recomputes the winning channel. The writing engine writes prose only. Every number in the report comes from the engine.
The data map
This table lists what the system uses and what it leaves out. Filter the table to see the gaps.