Product 360

Product 360 in CluedIn — Step-by-Step Implementation Guide

Product 360 delivers a single, trusted view of each product across ERP, PIM, e‑commerce, supplier feeds, and catalog systems. This guide walks you through the end-to-end build in CluedIn: ingest → standardize → match/link → golden record → enrich → govern → publish.


Outcomes

  • One golden product record per product (and per variant) with complete, standardized attributes.
  • Linked entities: Product ↔ Variant ↔ Category/Taxonomy ↔ Supplier ↔ Price ↔ Inventory ↔ Assets (images/specs).
  • Quality, lineage, and auditability for compliance and omnichannel consistency.
  • Publish/sync to PIM, e‑commerce, ERP, search, and analytics.

Prerequisites

  • Source access: ERP (item master), PIM, e‑commerce platform, supplier feeds (CSV/Excel/API), DAM (images/specs), pricing/inventory.
  • RACI: Ensure Accountable access to objects you will import/export.
  • Use a dedicated server/VM for heavy runs (ingestion, matching, toolkit exports/imports).

Reference Data Model

Core Entities

  • Product (base/SPU)
  • Variant (SKU level)
  • Category / Taxonomy (canonical categories, possibly multiple hierarchies)
  • Supplier (links to Supplier 360)
  • Price (list/promo, channel, currency, effective dates)
  • Inventory (location, on‑hand, safety stock)
  • Asset (images, datasheets, manuals, specs PDF)

Key Attributes (examples)

  • Product: Brand, Manufacturer, MPN, Title, Description, Category, Attributes (JSON or normalized columns), LifecycleStatus.
  • Variant: SKU, GTIN/EAN/UPC, Size, Color, Material, Dimensions, Weight, Pack/Case/Unit relationships.
  • Asset: URL/Path, Type, Format, Resolution, Primary/Secondary, Hash.

Relationships

  • Product has many Variants, Assets, Prices, Inventory records.
  • Product belongs to one or more Categories/Taxonomies.
  • Product supplied by Supplier(s); kit/bundle links to components (BOM).

Step 1 — Connect & Ingest Sources

Portal: Data Sources → Add Source

  1. Add ERP item master and attributes.
  2. Add PIM/e‑commerce catalogs (including channel overrides).
  3. Add supplier feeds (cost, specs) and DAM (assets).
  4. Add pricing and inventory sources (if modeled in CluedIn).

Tips

  • Start with samples to inspect field coverage and formats.
  • Capture source metadata: SourceSystem, LoadTimestamp, primary keys.

Step 2 — Define Entities & Mappings

Portal: Entity Explorer → + Create Entity Type

  1. Create Product, Variant, Category, Asset, Price, Inventory.
  2. Map source fields to target attributes (e.g., ShortDescTitle).
  3. Persist provenance (SourceSystem, SourceRecordID, SourceField).

Tip: Keep variant‑level attributes (SKU, GTIN, Color/Size) separate from product‑level attributes (Brand, Title).


Step 3 — Standardization & Normalization

Portal: Governance → Rules (Data Part Rules)
Create rules to normalize inputs before matching:

  • Text: trim, case, punctuation; remove HTML; normalize diacritics.
  • Units: convert dimensions/weight to canonical units (e.g., cm/kg).
  • Identifiers: validate/checksum for GTIN/EAN/UPC; normalize MPN (strip spaces).
  • Taxonomy: map source categories → canonical taxonomy (use RDM if available).
  • Assets: compute file hash; flag duplicates; classify by type.

Tags (recommended): Standardized, InvalidGTIN, MissingKeyAttrs, LegacySource.


Step 4 — Product Identity Resolution (Matching)

Portal: Entity Matching → Product / Variant

  1. Blocking keys to reduce comparisons:
    • (GTIN) when present (Brand + MPN) (NormalizedTitle + Category)
  2. Match rules (examples):
    • Variant (SKU) auto‑merge: exact GTIN OR exact SKU within same Brand.
    • Product (SPU) auto‑merge: exact Brand + MPN.
    • Candidate (steward): FuzzyTitle ≥ 0.92 AND Brand equal AND similar key attributes (e.g., size/pack).
  3. Set thresholds (e.g., Auto‑merge ≥ 0.97; Steward review 0.90–0.97).
  4. Run matching; review in Data Stewardship (approve/reject/Unmerge).

Pitfall: Don’t merge variants across different packs (e.g., single vs 12‑pack). Include PackSize / UoM in rules.


Step 5 — Stewardship: Review & Merge

Portal: Data Stewardship → Match Review

  • Validate candidates using GTIN, Brand+MPN, and key attributes.
  • Approve merges, reject false positives, and tag AmbiguousMatch when needed.
  • Use Unmerge if a mistake is found (audit preserved in History).

Step 6 — Golden Record Survivorship

Portal: Governance → Rules (Golden Record Rules)
Define precedence per attribute and tie‑breakers:

Attribute Precedence (high → low) Tie‑breaker
Brand, MPN ERP > PIM > Supplier Most recent verified
Title PIM > ERP > e‑commerce Longest non‑marketing title
Description PIM (tech) > ERP > e‑commerce Richest (char count + spec fields)
Category Canonical Taxonomy (RDM) > Source Steward override
GTIN PIM/ERP only N/A
Attributes PIM > ERP > Supplier Highest completeness
Images/Assets DAM > PIM > e‑commerce Resolution & aspect ratio
LifecycleStatus ERP > PIM Most restrictive status

Actions

  • Tag Sellable when required attributes complete and status active.
  • Compute CompletenessScore and tag LowCompleteness < 0.85.

Step 7 — Asset Management (Images/Docs)

Portal: Data Sources (DAM) / Rules

  • Deduplicate by hash; pick PrimaryImage per variant (min resolution, white background if policy).
  • Validate asset links; move broken links to stewardship queue.
  • Channel‑specific renditions (size/aspect) can be generated downstream; store metadata.

Step 8 — Pricing & Inventory (Optional Modeling)

Portal: Entity Explorer / Rules

  • Price: store List, Sale, Currency, Channel, EffectiveFrom/To.
  • Inventory: store Location, OnHand, AvailableToSell.
  • Compute InStock flag for channel exports.

Step 9 — Quality Scoring & Dashboards

Portal: Data Quality → Profiling / Dashboards
KPIs: completeness %, invalid identifiers, duplicate rate, image coverage, attribute fill by category, enrichment freshness.


Step 10 — Governance, Security & Lifecycle

Portal: Governance → Permissions / Policies

  • Restrict sensitive attributes (e.g., cost).
  • Enforce RACI for stewards and domain owners.
  • Apply Retention (e.g., archive discontinued after N years; keep specs/lineage).

Step 11 — Enrichment (Optional)

Portal: Data Sources / Rules

  • Pull manufacturer specs, regulatory attributes (RoHS/REACH), and alternate identifiers.
  • Track LastEnrichedAt and provider; re‑check on schedule.

Step 12 — Publish / Integrate

Portal: Exports

  • PIM/e‑commerce: golden product + variant attributes, primary assets, channel flags.
  • ERP: authoritative updates where CluedIn is source of truth.
  • Analytics/DW: wide denormalized table for BI/search.
  • APIs/Eventing: expose read endpoints and change events.

Start read‑only to validate mappings, then switch to authoritative syncs.


Step 13 — Scheduling & Operations

  • Schedule ingestion, matching, golden updates, enrichment, and exports (off‑peak for heavy jobs).
  • Monitor runtimes; alert on spikes or queue backlogs.
  • Use Product Toolkit for packaging to staging/prod; run on a dedicated box with Accountable access.

Step 14 — Validate & UAT

  • Sample 50–100 high‑volume products across categories.
  • Verify no cross‑product conflation; confirm variant separation and image correctness.
  • Business sign‑off from PIM/e‑commerce/ERP owners.

Go‑Live Checklist

  • All in‑scope sources ingested; mappings validated.
  • Matching thresholds tuned; false‑positive rate low; unmerge tested.
  • Golden survivorship approved; completeness thresholds set.
  • Canonical taxonomy mapping verified.
  • Asset primary/alt images validated.
  • Exports tested end‑to‑end in target systems.
  • Permissions, retention, and audit verified.
  • Promotion package built with Product Toolkit on dedicated VM.

Example Rules (Snippets)

Normalize GTIN / MPN

  • Condition: GTIN present
  • Action: strip non‑digits; validate checksum; tag InvalidGTIN on fail.
  • Condition: MPN present → uppercase, remove spaces/hyphens → MPN_Normalized.

Auto‑merge Variant

  • If GTIN exact OR (SKU exact AND Brand equal) → Confidence 1.0 (auto).

Candidate Product Merge

  • If Brand equal AND MPN_Normalized equal → Confidence 0.98 (auto).
  • Else if FuzzyTitle ≥ 0.92 AND Category equal AND KeyAttrSim ≥ 0.9 → Confidence 0.95 (steward).

Golden Survivorship

  • Title: PIM > ERP > e‑commerce; prefer longest non‑marketing title.
  • PrimaryImage: choose highest resolution; white background preferred; tag Primary.

Sellable Flag

  • If required attrs complete (Title, Brand, Category, GTIN, PrimaryImage) AND LifecycleStatus=Active AND InStock=true → set Sellable=true.

Common Pitfalls & How to Avoid Them

  • Merging different pack sizes → include Pack/UoM in match rules.
  • Over‑reliance on fuzzy title → require GTIN or Brand+MPN for auto‑merge.
  • Channel overrides pollute golden → keep Golden canonical; apply channel transforms in exports.
  • Dirty units/attributes → standardize units before matching; centralize UoM with RDM.
  • Running heavy jobs on laptops → use dedicated compute to avoid timeouts.

Success Metrics

  • Duplicate rate ↓; auto‑merge precision ≥ 98%.
  • Golden completeness ≥ 90% per category; image coverage ≥ 95%.
  • Time‑to‑publish new SKU ↓ (hours, not days).
  • Downstream catalog exceptions ↓; price/inventory sync accuracy ↑.

Summary

Product 360 in CluedIn unifies product and variant data, establishes clear survivorship, governs taxonomy and assets, and reliably syndicates to every channel. Start with clean identifiers and taxonomy, validate merges with high‑confidence evidence, and keep golden canonical while handling channel‑specific needs in exports.