ERP Selection Criteria Checklist for Manufacturers: 8-Category Scorecard
Score every platform on your shortlist before the demos begin.
Rate each criterion 1 (not met), 2 (partially met) or 3 (fully met). Weight each category by your operational priority. The weighted total determines which platform advances to the demo shortlist.
Request the Fillable VersionAs featured in
Client
Partner
Member
How to Use This Checklist
Use this checklist to score each ERP platform on your shortlist before vendor demos begin. The point is not to collect vendor self-scores. It is to have a documented basis for checking them.
For a printable and Excel-fillable version of this checklist, contact GrayCyan.
Complete the checklist for your operation first
Mark each criterion as required (R), preferred (P), or not applicable (N/A).
Send the checklist to each vendor on your longlist
Ask them to self-score and provide evidence for each criterion.
Verify vendor self-scores during demos
Use the demo tests below to check each criterion claim rather than accepting the self-score.
Weight each category by your operational priority
Compliance carries more weight for regulated manufacturers. AI readiness carries more weight for manufacturers planning AI deployment in the next 18 months.
Compare weighted totals
The platform with the highest weighted score against your specific requirements advances.
ERP Selection Criteria: The Eight Categories to Score
Each category below lists the criteria to score and, where the scoring is easy to get wrong, the test that settles it. Open a card to see the full criterion list.
Manufacturing Module Depth
Priority weight: 25% default
- Supports the manufacturer's production type natively: discrete (BOM and work order), process (formula and batch), make-to-order, configure-to-order
- Multi-level BOM with phantom assembly support
- Work orders with routing and operation-level tracking
- MRP and MPS: demand-driven planning connected to current inventory and open POs
- Shop floor scheduling: work order dispatch by work center and operator
- Scrap and rework tracking with cost capture
- Production variance reporting: standard vs actual material, labor, overhead
- AI readiness sub-check: production order data accessible via ERP API for ML anomaly detection and schedule intelligence
Inventory and Warehouse Management
Priority weight: 20% standard, higher for high pick volume or compliance
- Multi-location inventory with bin-level tracking
- Lot tracking: lot number assigned at receiving, tracked through every inventory movement to shipment
- Serial number tracking: unit-level tracking for aerospace, medical devices, or defense
- FEFO rotation: expiry date-driven pick sequencing enforced at the scan level, not just in the pick list
- Cycle counting: ABC-driven task scheduling, count recording, and discrepancy posting
- Scan-based receiving: PO-driven receipt scan confirmation, or does it require manual entry after the fact?
- WMS adequacy for pick volume: is the native module sufficient at current pick volume, or is a custom WMS layer required?
- AI readiness sub-check: inventory transaction data (stock moves, lot assignments) accessible via API for real-time anomaly detection
Compliance Documentation
Priority weight: 25% regulated, 5 to 10% non-regulated
Score 3 (fully met) only if the ERP generates compliance documentation at the scan event, not from manual form entries after the fact.
- FSMA 204: KDE capture at receiving, transforming and shipping CTEs in FDA-required structured format
- FSMA 204: forward and backward lot trace producible in under 30 seconds from any lot number
- AS9100: serial genealogy linking each unit to raw material COA, NADCAP special processes, and quality inspection results
- 21 CFR Part 11: electronic signatures and immutable audit trail for regulated transactions (pharmaceutical)
- METRC: API integration connecting inventory movements to METRC plant tags, packages and transfer manifests (cannabis)
- USDA and HACCP: batch records and lethality evidence generated at production completion (meat and poultry processors)
- If the compliance module is custom-built: who built it, who owns the source code, and who maintains it through ERP version upgrades?
Financial and Cost Accounting
Priority weight: 15% default
- Costing method supported: standard cost, average cost, FIFO, FEFO lot-actual. Does the ERP support the method your production type requires?
- Job costing: actual material, labor and overhead captured against individual production orders
- Landed cost allocation: freight, duty and insurance allocated to received unit cost at receipt
- Multi-entity: intercompany transactions and consolidated reporting if you operate multiple legal entities
- Period close: automated period close with variance analysis, WIP valuation and inventory reconciliation reports
- Three-way match: PO, goods receipt and supplier invoice matched automatically before AP release
Integration Architecture
Priority weight: 15% standard, higher for a complex external system landscape
- API documentation: published, versioned, stable API documentation available before purchase
- Read access: all operational data (inventory moves, production orders, lot records, quality records) readable via API
- Write access: inventory transactions, purchase orders and production order status writable via API, not read-only
- Webhook support: ERP fires event webhooks on operational transactions for real-time external system integration
- Native connectors: pre-built connectors for carrier APIs, EDI partners and WMS platforms on your shortlist
- Integration cost: per-integration cost for systems without native connectors, and who owns the middleware source code
AI Integration Readiness
Priority weight: 15% when AI is planned within 18 months, 5% otherwise
This category does not appear on most ERP selection checklists. It should. The ERP you select today is the AI infrastructure your manufacturing operation will live with for the next decade. An ERP that constrains AI integration now will constrain it for that whole period.
- API completeness: full production, inventory and quality transaction history readable via API for ML model training
- API rate limits: rate limits support real-time anomaly detection at your transaction volume without throttling
- Write-back: AI results (updated reorder points, draft purchase orders, anomaly alerts) writable to ERP records via API
- Webhook event coverage: webhooks available for the specific events agentic automation would need, such as production completion, inventory threshold and quality hold
- On-premises deployment: fully on-premises deployment available with no cloud dependency (required for ITAR-controlled operations)
- Test in demo: ask the vendor to demonstrate how an external Python service would read 30 days of inventory moves and write an updated reorder point back to an item record via API. If they cannot demonstrate this in the demo, score 1.
Implementation Partner Ecosystem
Priority weight: 10% default
- Manufacturing industry depth: the partner has documented implementations for your specific vertical (food, pharma, aerospace, cannabis, discrete)
- Compliance experience: the partner has deployed the required compliance modules on this specific ERP platform. Ask for a live reference.
- US-based partner availability: if US-based support is required, confirm the partner has US-based resources
- Reference quality: the partner can provide a reference from a live manufacturer in the same vertical with the same compliance requirements, post go-live, not mid-implementation
- Source code ownership: custom modules built during implementation are owned by the manufacturer at completion. Confirm this in the contract.
- Change order policy: what constitutes an out-of-scope addition, and how is it priced? Fixed-fee integrations require correct scoping.
Total Three-Year Cost of Ownership
Priority weight: 10% as a scoring category, but use it as an absolute filter
Eliminate any platform whose three-year TCO exceeds your budget, regardless of score in other categories.
- Year 1 license cost: per user or per module fee structure, minimum user count, included modules vs add-on modules
- Implementation services estimate: phase-by-phase breakdown from the partner, not a single total number
- Integration development: per-integration cost for each external system, ownership of middleware source code
- Custom compliance module development: if required, scoped cost and ownership
- Data migration: explicit scope and cost, including data quality remediation if source data quality is below threshold
- Years 2 to 3 ongoing costs: annual license, maintenance (on-premises), support tier, upgrade testing labor
- Hidden costs identified: sandbox environments (cloud), annual price escalation clause, module additions
Scoring Summary Table
Score each platform 1 to 3 per category, multiply by the category weight, and compare weighted totals.
| Category | Weight | Platform A | Platform B | Platform C |
|---|---|---|---|---|
| 1. Manufacturing Module Depth | 25% | __ /3 | __ /3 | __ /3 |
| 2. Inventory & Warehouse Management | 20% | __ /3 | __ /3 | __ /3 |
| 3. Compliance Documentation | 25% | __ /3 | __ /3 | __ /3 |
| 4. Financial & Cost Accounting | 15% | __ /3 | __ /3 | __ /3 |
| 5. Integration Architecture | 15% | __ /3 | __ /3 | __ /3 |
| 6. AI Integration Readiness | 15% | __ /3 | __ /3 | __ /3 |
| 7. Implementation Partner Ecosystem | 10% | __ /3 | __ /3 | __ /3 |
| 8. Total 3-Year Cost of Ownership | 10% | __ /3 | __ /3 | __ /3 |
| WEIGHTED TOTAL | 100% | __ | __ | __ |
Weight adjustment
The figures above are defaults. Set them to your own operational priorities and normalise the set to 100 before scoring begins, because the weighting is what makes the totals comparable across platforms.
Regulated manufacturers should increase Category 3 (Compliance) and reduce Category 4 accordingly. Operations planning AI in the next 18 months should increase Category 6 (AI Readiness) and reduce another category by the same amount. Total weights must always sum to 100%.
Three Criteria That Are Harder to Score Than They Look
These three are where vendor self-scores and verified scores diverge most often. Each one has a specific demo test that settles it.
Compliance documentation
Category 3
Vendors universally claim compliance capability. The real test is whether compliance documentation is generated at the scan event or assembled manually afterward.
In the demo, ask: “Show me how an FSMA 204-compliant receiving record is generated when I scan a receiving barcode.” If the answer involves navigating to a separate screen after the scan, score 2.
AI integration readiness
Category 6
Most ERP vendors cannot demonstrate live API write-back in a demo. They will show screenshots or documentation instead.
Require a live API demonstration: a Python script reading inventory data from the ERP and writing a reorder point update back. If they cannot do this in the demo, they cannot support real-time AI integration in production.
Implementation partner reference quality
Category 7
A reference from a manufacturer still in implementation is not a reference. A reference from a manufacturer in a different vertical is not applicable.
Require a reference that is post go-live, in the same industry, and with the same compliance requirements. Ask the reference: “What went wrong, and what would you scope differently?”
Ready to Score Platforms? GrayCyan Can Help With the Hard Parts.
Filling in this checklist is the easy part. Verifying vendor claims in demos, checking references from your industry, modelling the honest three-year total cost, and identifying compliance gaps before the implementation contract is signed: that is where GrayCyan's ERP consulting adds value.
See the full ERP selection guide or independent ERP selection consulting →