Skip to main content
Operational Methodology • Under 60-Second Loop

How MeritSKU Validates Products Before Launch.

From raw supplier quote to human-gated pre-launch decision. Walk through the four deterministic phases that turn unverified product proposals into fully audited financial models.

Deterministic Commerce Intelligence4-Phase Loop

The 4-Phase MeritSKU Operating Loop

Turn raw supplier quotes into audited pre-launch decisions in under 60 seconds with strict draft-only safeguards.

Total SLA: < 60 Seconds
Step 1 • IngestionDuration: < 5 Seconds

Multi-Channel Candidate Ingestion with Strict Boundary Guards

Ingest raw candidate URLs from AliExpress, 1688, Taobao, or direct factory quote PDFs. Automatically validate product dimensions, weight, HS customs codes, and quoted factory FOB tiers with zero exposure to private networks.

Simulation: Intake & Boundary GuardrailsLive Interactive Node
https://1688.com/offer/654812398.html?sku=hydro_ergonomic_cushion
Quoted Unit FOB$14.50 USD
Sample MOQ Tier500 Units
SSRF Guard Verified • RFC 1918 Address Blocked
PASS
Architectural Invariants & Proofs
  • SSRF Protection: Strict private IP and DNS rebinding filters on all external URL fetches
  • Integer Minor-Unit Storage: All FOB cost quotes stored in minor units (cents) to prevent float drift
  • Client-Side Sandbox Isolation: Zero persistent database writes during public evaluations
Governing Standard:SSRF Safe Parser • RFC 1918 Filter • Minor-Unit (cents) Ingestion Engine
Supported Sources
AliExpress, 1688, Taobao, PDF Quotes
Ingestion Latency
< 5.0s p95
Database Mutation
Zero (Sandbox)
Test this full 4-phase loop live with real or synthetic candidate SKUs:

The 4 Validation Gates in Detail

Explore how data flows through each verification stage to produce objective, reproducible decisions.

Phase 01 • Ingestion Gate

Candidate Intake & SSRF Guardrails

Accepts live candidate links from AliExpress, 1688, Taobao, or uploaded factory quote PDFs. Our ingestion engine enforces strict boundary isolation to prevent intranet exposure or database poisoning.

SSRF Defense-in-Depth

URL resolution uses DNS re-check and RFC 1918 private IP address filtering. Zero network access to internal Cloud Run services.

Integer Minor-Unit Storage

FOB prices, minimum order quantity (MOQ) brackets, and packaging dimensions are stored in integer cents and grams to prevent float drift.

Phase 02 • Evidence Gate

Multi-Source Evidence Extraction

Aggregates evidence across six orthogonal dimensions. Unlike black-box machine learning ratings, every score is explainable down to the source quote and market data point.

6 Scoring Dimensions

Demand Momentum, Landed Economics, Competition Density, Supplier Reliability, Testability & Creative, and Regulatory/FTC Risk.

Orthogonal Confidence Capping

Reflects evidence completeness and half-life freshness decay. A high score with unobserved supplier metrics is capped at low confidence.

Phase 03 • Stress-Test Gate

12-Factor Landed Economics & Sensitivity

Calculates true Contribution Margin 2 (CM2) across twelve integer cost factors. Stress-tests margins across a 9-cell matrix simulating retail price shifts and paid media ad auction spikes.

9-Cell CM2 Matrix

Evaluates unit profit against retail prices (±10%) and paid CAC spikes (up to +20%). Identifies fragile products before ad launch.

Brier Skill Calibration

Tracks model calibration against realized merchant post-launch actuals. Enforces explicit epistemic modesty notices whenever N < 20.

Phase 04 • Staging Gate

Governed Draft Action Center

Formulates pre-launch ad test plans with Target CPA benchmarks and protective stop-loss ad ceilings. Enforces strict draft-only staging to Shopify with zero live catalog writes.

Draft-Only Shopify Staging

Candidate products are staged strictly as unpublished drafts (status: DRAFT). Never alters active catalog items.

Stop-Loss Ad Ceilings

Test plans include mathematically enforced stop-loss budgets ($200–$1,000 bounds) to protect testing capital from creative burn.

Frequently Asked Questions

Clear answers on technical boundaries, data privacy, and operational safety.

Does MeritSKU connect directly to my live store?

Yes, via official Shopify OAuth with restricted scopes. MeritSKU only reads historical sales actuals to calibrate unit economics and writes candidate products strictly as unpublished drafts (status: DRAFT).

Can MeritSKU publish products live or start paid ads automatically?

Never. Our architectural boundary strictly forbids autonomous live publishing or automated ad execution. MeritSKU produces strategy blueprints and draft staging that require explicit human approval.

How does MeritSKU calculate landed cost without connecting to my 3PL?

MeritSKU maintains deterministic freight and fulfillment rate benchmarks across 10 e-commerce categories, modeled in integer minor units. You can also input your exact contracted carrier and 3PL pick/pack rate cards.

What happens if supplier data is missing or unverified?

The Orthogonal Confidence metric binds the overall score, alerting you to unobserved risk rather than hallucinating missing inputs. If a supplier quote lacks customs or freight terms, the confidence score drops.

Ready to Audit Your First Candidate Product?

Run a complete pre-launch viability audit in our interactive demo sandbox in under 60 seconds.