Method 3.1.0
Here is exactly what the tool reads, what it calculates, where artificial intelligence is allowed to help, and where it is not allowed to make the decision.
We read public HTML or the website’s crawler fallback copy. If neither is readable, we use public web research and label the source. Ideas use the founder’s description. Findings carry model-assigned high, medium, low, or unknown labels; these are not independent verification.
Market size is a working estimate, not a fact. The report shows a low, base, and high case plus the arithmetic and assumptions. Repeat scans keep the last completed baseline so the number does not jump without a changed assumption.
You confirm budget, client value, founder hours, team skills, location, urgency, capacity, sales motion, and client goals. Missing inputs lower confidence. Client goals are converted to yearly numbers before they are compared with market size.
Start with library assumptions across 18 factors, then adjust for budget, annual client revenue, time, skills, geography, regulation, sales motion, capacity, and business type. Buyer concentration, intent, and several other factors remain library assumptions, not measured demand. Gross margin informs the written guidance, but is not yet an arithmetic scoring input. The score is the sum of each factor times its effective weight, divided by the sum of weights. High urgency adds 4 to speed weight and subtracts 2 from compounding; low urgency adds 3 to compounding and subtracts 2 from speed, with a minimum weight of 1. New reports save the effective weights. Zero cash or hours stays zero.
The Primary channel is the highest total fit score. Fast-Test is 60% total fit, 25% speed, and 15% budget fit. Compounding is 65% total fit and 35% builds-over-time value. These choices are deterministic: the writing model cannot replace them.
Artificial intelligence explains the fixed picks and drafts buyer-location hypotheses, channel tests, and weekly actions from saved inputs. This step does not browse the web. Proposed locations require a buyer-access check before spending. Each new top-five test includes a cash ceiling, hours, a time box of up to 30 days, an early signal, success threshold, and stop rule. Each test is checked against your available cash and time; tests are alternatives, not concurrent commitments. Missing proof requires a collection plan, never invented claims.
Competitive research uses public marketing evidence only: published pages, pricing, reviews, communities, events, and visible campaigns. Unverified facts must be labeled unknown. The tool looks for positioning moves, not material to copy.
The copy audit uses 10 fixed items on a 0–10 scale. Easy to Read covers message clarity, plain language, scanability, and specificity. Wins Clients covers buyer desire, cost of staying stuck, proof, offer clarity, action clarity, and urgency. Overall weights these 40% and 60%.
Review a recommended test, set dates and a measured baseline, then save it to your experiment tracker. Record what happened, what you learned, and whether to continue, change, or stop. Saved results do not automatically retrain channel scores. The next research step is validating reachable buyers and their actual buying signals.
The Client Global Positioning System checks named market signals weekly for active members. It records what was verified, what could not be verified, whether the lead channel is strong or fading, and the exact threshold for changing course.