Magendoo Catalog Quality — Magento 2 Module

Catalog data quality observability for Magento 2.
Know exactly which products are hurting your conversion rate before your customers find out.

What It Does

Magendoo_CatalogQuality scores every product in your catalog against a two-axis quality model — Enrichment (are the important attributes filled in?) and Consistency (do they meet your content rules?) — and combines them into a single Product Health Score (PHS) between 0 and 100. Products are graded A–E.

The module answers three questions that every Magento merchant needs:

1. Which products are incomplete? — The Gap List shows every product's grade, PHS, priority gaps, and the specific missing attribute codes.
2. Which products are invisible in layered navigation? — The Filter-Blind detector flags products that have no value for a filterable attribute, making them unfindable when shoppers use your category filters.
3. Which filter attributes have the worst coverage across categories? — The Filter Coverage Matrix shows fill rates as a colour-coded heatmap, scoped by attribute set.

No auto-fix. Just clear, actionable diagnostic data.

Screenshots

Dashboard

[image: Catalog Quality Dashboard]

The dashboard shows the Catalog Health Index (CHI) — an aggregate PHS across all scored products of the default store view, weighted by 90-day product revenue when revenue weighting is enabled and sales data exists — alongside grade distribution, filter-blind product count, and a 30-day CHI trend chart.

Product Gap List

[image: Product Gap List]

Every scored product in one grid: grade badge, PHS, P1/P2/P3 gap counts, a filter-blind chip, and colour-coded missing attribute chips. Red chips are filterable attributes directly causing filter-blindness; grey chips are other priority gaps.

Filter Coverage Matrix

[image: Filter Coverage Matrix]

Rows are user-defined filterable attributes; columns are active categories (filterable by attribute set). Each cell shows the fill-rate percentage, colour-coded from green (≥90 %) to red (<35 %). The callout above the matrix links to the Gap List for drilling into individual filter-blind products.

Attribute Priorities

[image: Attribute Priorities]

The priority grid controls how much each attribute matters to the score. Priority levels are inline-editable; every change is audited and marks the score index for rebuild.

Features

• Two-axis PHS scoring: Enrichment completeness × Consistency conformance multipliers
• A–E grade thresholds: ≥90 → A, ≥80 → B, ≥70 → C, ≥50 → D, <50 → E
• Revenue-weighted CHI: weights products by 90-day revenue (order totals minus refunds, processing/complete orders); falls back to the equal-weighted average when no revenue exists
• Filter-blind detection: flags products invisible in layered navigation per store view, considering only filterable attributes that belong to the product's own attribute set
• Attribute Priority System: P1/P2/P3/P4/Ignore per attribute, with category and attribute-set overrides; every change is audited and marks the score index for rebuild
• Placeholder blocklist: regex and literal patterns to catch values like "N/A", "TBD", "Lorem ipsum"
• 9 built-in quality rules: Meta Title, Meta Description, URL Key, Name, Description, Image, Price, Filterable Attribute, Category Assignment — each rule's enabled flag and severity come from the magendoo_cq_rule_config table
• Conditional attribute groups: score module-specific attributes only when their control flag is on (e.g. base-price fields only when baseprice_is_enabled = 1)
• Magento indexer integration: incremental updates via MView subscriptions on the product entity, EAV value, and category-assignment tables; full reindex via CLI
• Nightly CHI snapshots: 30-day trend data stored automatically via cron (one row per store view per UTC day)
• CSV export of the priority table via CLI (catalog-quality:priorities:export)
• REST API: 6 endpoints for scores and priorities
• Inline-editable Priority Grid: change priority levels without leaving the listing
• Filter Coverage Matrix: attribute set filter, product counts per category column
• UTC everywhere: all timestamps the module writes (score computation times, priority history, snapshots, run logs) are UTC

Requirements

Installation

Via Composer (recommended)

Until the package is listed on Packagist, add the GitHub repository first:

composer config repositories.magendoo-catalog-quality vcs https://github.com/magendooro/magento2-catalog-quality

Then install:

composer require magendoo/module-catalog-quality
bin/magento module:enable Magendoo_CatalogQuality
bin/magento setup:upgrade
bin/magento setup:di:compile
bin/magento setup:static-content:deploy -f
bin/magento cache:flush

Manual

1. Copy the module directory into app/code/Magendoo/CatalogQuality/
2. Run the same CLI commands above

Upgrading from 1.x to 2.0

Version 2.0.0 tightens the database schema: setup:upgrade adds foreign keys to the
score and priority tables plus unique constraints on
magendoo_cq_priority (attribute_id, attribute_set_id, category_id) and
magendoo_cq_catalog_snapshot (store_view_id, snapshot_date). MySQL refuses to add
a unique constraint while duplicate rows exist, so on a 1.x install you must run
Setup/orphan-cleanup.sql against your database before
setup:upgrade — it removes orphaned score rows and deduplicates both tables
(1.x could accumulate duplicates via repeated bulk priority upserts and same-day
snapshot cron re-runs). Skipping this step can abort setup:upgrade mid-run with a
"Duplicate entry" error. The script is idempotent; on installs with a DB table
prefix, prefix the table names before running.

2.0.0 also contains breaking API changes for code extending the module:

• Magendoo\CatalogQuality\Api\ScorerInterface gained the method
  scoreBatch(array $productIds, int $storeViewId): int. Custom scorer
  implementations must implement it.
• ScorerInterface::score() now throws NoSuchEntityException for a product ID
  that does not exist instead of persisting and returning an empty E-grade score row.

Initial Setup

After installation, run the full indexer to score all products:

bin/magento indexer:reindex magendoo_cq_product_score

This may take several minutes on large catalogs. The batch size is configurable under Stores → Configuration → Magendoo Extensions → Catalog Quality → Performance.

Once the reindex finishes, open Catalog Quality → Dashboard in the admin menu: start with the worst grades in the Gap List, fix the missing P1 attributes, and watch the CHI climb on the next reindex.

Admin Screens

Navigate to Catalog Quality in the top-level admin menu.

Configuration

Stores → Configuration → Magendoo Extensions → Catalog Quality

Cron schedule: snapshots are written daily at 02:00 (magendoo_cq_rebuild_snapshot); old snapshots are pruned Sundays at 03:00 (magendoo_cq_prune_snapshots).

Scoring Algorithm

See docs/scoring-algorithm.md for the full specification.

Summary:

PHS(product, store_view) =
    [Σ weight(attr) × completeness_score(attr) × conformance_mult(attr)] / [Σ weight(attr)]
    × product_rule_mult                     ← product-level rules (category assignment)
    × 100

completeness_score:  empty / placeholder → 0.0 | present → 0.5 | meets_bar → 1.0
conformance_mult:    1.0 baseline; per failed rule the multiplier for the rule's
                     configured severity: critical × 0.7, warning × 0.9, info × 1.0

CHI = Σ(revenue_90d × PHS) / Σ(revenue_90d)     (falls back to equal-weight)

Rule severities and enabled flags live in the magendoo_cq_rule_config table — disabled rules are skipped entirely.

Priority System

See docs/priority-system.md for the full hierarchy.

Quick reference:

Priority resolution order (first match wins):

1. Explicit row with matching category_id
2. Explicit row with matching attribute_set_id
3. Explicit global row (no set, no category)
4. Implicit: SEO attrs → P1; user-defined filterable → P2; user-defined → P3; else → Ignore

Priority rules are unique per (attribute, attribute set, category) scope — saving the same scope again updates the existing row instead of creating a duplicate. Every create, change, and delete is recorded in magendoo_cq_priority_history and marks the score indexer invalid, so scores refresh on the next reindex.

Conditional Attribute Groups

Some module attributes should only be scored when they are relevant for a specific product. Configure this in etc/di.xml:

<type name="Magendoo\CatalogQuality\Model\Scoring\ProductScorer">
    <arguments>
        <argument name="conditionalGroups" xsi:type="array">
            <item name="baseprice" xsi:type="array">
                <item name="enabledBy" xsi:type="string">baseprice_is_enabled</item>
                <item name="attributes" xsi:type="array">
                    <item name="0" xsi:type="string">baseprice_product_amount</item>
                    <item name="1" xsi:type="string">baseprice_unit_id</item>
                    <item name="2" xsi:type="string">baseprice_reference_amount</item>
                </item>
            </item>
        </argument>
    </arguments>
</type>

The enabledBy attribute is never scored itself — it acts as a gate. The listed attributes are only evaluated when the gate attribute is truthy (1). The example above wires up Magendoo_BasePrice so that unit-pricing compliance fields only count as gaps on products where base pricing is switched on.

CLI Commands

# Full reindex — score all products, optionally scoped to one store view
bin/magento catalog-quality:rebuild [--store=<store_view_id>] [--batch=<size>]

• --store limits the rebuild to a single store view. The ID is validated; an unknown ID fails with a message listing the valid store view IDs. Without it, all store views are scored.
• --batch overrides the configured batch size for this run (must be a positive integer).
• The command reports the number of product-store pairs scored and the elapsed time.

# Debug a single product (table output by default)
bin/magento catalog-quality:score <sku> [--store=<store_view_id>] [--format=table|json]
# Example JSON output (--format=json):
# {"sku":"WH-X500","store_view_id":1,"phs":42.15,"grade":"D","filter_blind":true,
#  "p1_gaps":1,"p2_gaps":2,"p3_gaps":0,"missing_attrs":["cq_connectivity","cq_color"],
#  "rules_failed":["name_quality","meta_description_quality"]}

# Export priorities as CSV (stdout by default)
bin/magento catalog-quality:priorities:export [--set=<attribute_set_id>] [--output=<file>]

REST API

See docs/rest-api.md for request/response schemas.

All endpoints require an admin bearer token. Authenticate with POST /V1/integration/admin/token. Priority writes validate the priority_level value (400 on invalid input), are audited to magendoo_cq_priority_history, and invalidate the score indexer.

Database Tables

Extending

Adding a Custom Quality Rule

1. Implement Magendoo\CatalogQuality\Api\RuleInterface
2. Register it in etc/di.xml under RulePool

<type name="Magendoo\CatalogQuality\Model\Rule\RulePool">
    <arguments>
        <argument name="rules" xsi:type="array">
            <item name="my_custom_rule" xsi:type="object">
                Vendor\Module\Model\Rule\Enrichment\MyCustomRule
            </item>
        </argument>
    </arguments>
</type>

Optionally seed a magendoo_cq_rule_config row for the new rule code to control its severity; rules without a config row run as enabled with warning severity.

Adding a Conditional Attribute Group

Register a conditionalGroups item as shown in the Conditional Attribute Groups section above.

Running the Tests

Pure unit tests live under Test/Unit and run with Magento's standard unit-test configuration (no database required):

vendor/bin/phpunit -c dev/tests/unit/phpunit.xml.dist app/code/Magendoo/CatalogQuality/Test/Unit

License

MIT — © 2025 Magendoo Interactive