How we reconcile a category. A single dataset can mislead. Our method cross-checks four independent reads of a category against each other and resolves them into one claim-graded call — with every signal carrying its evidence grade, the reasoning shown, and the limits stated.
A single dataset can mislead. Our method cross-checks four independent reads of a category against each other and resolves them into one claim-graded call — with every signal carrying its evidence grade, the reasoning shown, and the limits stated. This page explains the method, not a guarantee.
We read a category as an inventory system, not a ranking. Amazon best-seller movement is treated as a proxy for the rate at which a category's inventory is depleting or refilling — a flow signal — rather than a static measure of demand.
We read a category as an inventory system, not a ranking. Amazon best-seller movement is treated as a proxy for the rate at which a category's inventory is depleting or refilling — a flow signal — rather than a static measure of demand. A rank that is rising can mean depletion outpacing replenishment, not necessarily a structural rise in demand. Separating the two is the first thing the method does, because a snapshot read is where most single-signal research goes wrong.
Each read draws on a different dataset. The value is not any single read — it is whether they agree . When they contradict, that contradiction is the finding. Whether the demand signal is real movement or a transient spike — read as an inventory flow rate, with promotional and seasonal distortions called out rather than smoothed over.
Each read draws on a different dataset. The value is not any single read — it is whether they agree. When they contradict, that contradiction is the finding.
Whether the demand signal is real movement or a transient spike — read as an inventory flow rate, with promotional and seasonal distortions called out rather than smoothed over.
Whether availability and inventory behaviour agree with the demand read. Depletion that looks like demand, or stock that is being pushed ahead of an anticipated stockout, is a different regime — and the method is built to catch it.
Whether the supply base is concentrated, fragile, or shifting — read from public import records and supplier-overlap evidence, with concentration measured, not asserted, and each claim carrying its source.
Whether replenishment cadence in the import record lines up with the demand story. Import-flow is the anchor that tells a real demand trend apart from an inventory push — the two look alike on a rank chart and diverge in the shipment record.
Promote, watch, hold, or reject — with the reasoning shown, the assumptions named, the evidence grade on every signal, and the contradictions, if any, surfaced rather than resolved silently.
Before we interpret what a number does, we ask which underlying stock is changing and which flow is driving it. A category read that cannot answer that question carries a low-confidence flag rather than a confident interpretation. The method is designed to surface plausible-but-structurally-wrong reads — the failure mode that looks correct and is not — and to flag a.
Before we interpret what a number does, we ask which underlying stock is changing and which flow is driving it. A category read that cannot answer that question carries a low-confidence flag rather than a confident interpretation. The method is designed to surface plausible-but-structurally-wrong reads — the failure mode that looks correct and is not — and to flag a contradiction instead of smoothing it over.
We never present an inference as a fact. Each signal is labelled by how well the evidence supports it, so you can tell a confirmed read from a working hypothesis at a glance — and weight your decision accordingly.
We never present an inference as a fact. Each signal is labelled by how well the evidence supports it, so you can tell a confirmed read from a working hypothesis at a glance — and weight your decision accordingly.
We prove the method where a single component carries most of the risk: motorized appliances, where the category read routes to a motor-readiness and OEM-feasibility path grounded in a working Shenzhen motor manufacturer. It is the first proof vertical — the method is vertical-general by design.
We prove the method where a single component carries most of the risk: motorized appliances, where the category read routes to a motor-readiness and OEM-feasibility path grounded in a working Shenzhen motor manufacturer. It is the first proof vertical — the method is vertical-general by design.
See the motorized-appliance proof caseRequest a Category Intelligence Pilot, or start with a short, no-cost category screen to see whether the story reconciles before you commit. This page describes a method, not a guarantee. We reconcile public datasets into evidence-graded signals and human-reviewed recommendations, and we don't guarantee outcomes, cost savings, supplier replacement, or commercial success — a category read isn't certification, legal advice.
Request a Category Intelligence Pilot, or start with a short, no-cost category screen to see whether the story reconciles before you commit.
This page describes a method, not a guarantee. We reconcile public datasets into evidence-graded signals and human-reviewed recommendations, and we don't guarantee outcomes, cost savings, supplier replacement, or commercial success — a category read isn't certification, legal advice, or a binding quote.