ERP and MRP for Small Manufacturers

What is MRP?

Material Requirements Planning is a calculation system that answers one question: given a production schedule and a bill of materials (BOM), what materials do you need to order, and when? A basic MRP loop runs like this:

1. Master Production Schedule (MPS) — what finished goods you plan to build and when.
2. Bill of Materials (BOM) — the recipe: every component, sub-assembly and raw material that goes into each finished product.
3. Inventory records — what you have on hand and on order.
4. MRP calculation — the system explodes the BOM against the MPS, offsets by lead time, and produces planned purchase orders and work orders.

MRP was developed in the 1960s, standardised by APICS (now ASCM), and is still the operational core of almost every manufacturing software product sold today — whether it is called MRP, ERP, or something else entirely.

What is ERP, and how does it differ from MRP?

Enterprise Resource Planning is MRP plus everything else the business runs on: general ledger, accounts payable and receivable, purchasing, sales order management, HR and payroll, and often CRM, quality and maintenance modules too. The term "ERP" was coined by Gartner in 1990 to describe the extension of MRP II into a single integrated system across the whole enterprise.

Types of manufacturing ERP by segment

Not all ERP is the same. The category splits into several tiers, each with different capabilities, price points and implementation demands:

Key modules to evaluate

When evaluating any MRP or ERP product, assess how it handles each of these functional areas — not just whether it has the module, but how deep the functionality goes:

• Bill of Materials management — multi-level BOM, revision control, phantom assemblies, where-used queries.
• Master Production Scheduling — forward/backward scheduling, capacity constraints, schedule locking.
• MRP calculation — how often it runs (live, hourly, nightly), net vs gross requirements, exception messages.
• Inventory management — lot/serial tracking, bin/location, cycle counting, physical inventory.
• Purchasing — purchase requisitions, PO management, vendor management, receipts matched to POs.
• Shop floor / work order management — routing, work centres, labour reporting, job costing.
• Sales order management — quote-to-order, available-to-promise, delivery scheduling.
• Financials (ERP only) — GL, AR, AP, bank reconciliation, job cost P&L, standard costing vs actual costing.

What manufacturing ERP actually costs

The listed software subscription is only one part of total cost. For a small manufacturer (10–30 users), expect these cost categories:

• Software licence / subscription — the monthly or annual fee. Cloud products range from ~$50/month (lightweight MRP, single user) to $3,000+/month (full SME ERP, 20+ users).
• Implementation and configuration — either internal time or consultant fees. A $500/month cloud ERP can easily carry $15,000–$60,000 in implementation costs for a 10–25 user business. Larger systems run $100,000–$500,000+.
• Data migration — converting existing item masters, BOMs, open orders and customer/vendor records. Often underestimated; budget 10–20% of implementation cost.
• Training — formal training for key users, end-user sessions, and ongoing learning as the system evolves.
• Ongoing support and maintenance — cloud systems include software maintenance; on-premise systems charge 15–22% of licence annually.
• Customisation — avoid where possible. Each customisation is a future upgrade liability.

A realistic rule of thumb for SME ERP: expect to spend 2–4× the first-year software cost in implementation, data migration and training. For a $600/month product (~$7,200/year), that means $15,000–$30,000 in implementation — before any consultant hourly fees.

Realistic implementation timelines

Every ERP vendor's sales deck says "go live in 90 days." The reality:

• Lightweight MRP, simple operation: 4–10 weeks is achievable if BOM and item data is clean.
• SME cloud ERP, 10–30 users: 3–6 months for a straightforward implementation; 6–12 months if data is messy or processes need redesigning.
• Mid-market ERP, 30–100 users: 9–18 months. Multi-phase go-lives (financials first, production second) are common.
• Any ERP with poor data: add 3–6 months. Data quality is the single biggest driver of implementation overruns.

The biggest variable is the state of your bill of materials. If your BOMs are in spreadsheets with inconsistent part numbers, conflicting revision levels and missing lead times, clean that up before you sign an ERP contract — not after.

Signs you have outgrown QuickBooks for manufacturing

QuickBooks is an accounting tool, not a manufacturing system. It has an inventory module and some light manufacturing add-ons, but it was not designed for production planning. These are signs you need to move on:

• You maintain your bill of materials in a spreadsheet alongside QuickBooks.
• You cannot tell, in real time, whether you have enough components to fill a sales order.
• Job costing requires manual calculation or a second spreadsheet.
• You cannot run MRP — you plan production manually, on gut feel or a whiteboard.
• Inventory counts differ from QuickBooks because transactions are entered late or not at all.
• You spend more than a few hours per week reconciling production data to accounting data.

Cloud ERP vs on-premise for small manufacturers

Almost all new ERP installations for small manufacturers are cloud (SaaS) deployments today. On-premise still exists, but for SMEs the reasons to choose it are narrow:

How to run an ERP selection process

A structured selection process reduces the risk of choosing the wrong system. Here is a practical approach for a small manufacturer:

1. Define requirements before talking to vendors. List your must-haves (lot tracking, multi-currency, specific integrations) and nice-to-haves. Include current pain points and the processes that are broken.
2. Set a realistic budget — software plus implementation, not just software.
3. Shortlist 3–4 vendors that serve your industry segment and user count. Reject any that cannot give you published list pricing.
4. Run scripted demos. Give each vendor the same 5–8 scenarios to demonstrate (create a BOM, run MRP, receive a PO, enter a work order, close a job and see the cost). Do not let them show only their strengths.
5. Check references — specifically manufacturers of similar size and industry. Ask about implementation time, data migration experience, and what they would do differently.
6. Negotiate on implementation scope, not licence price. The licence is often non-negotiable; the implementation package is where there is room.
7. Plan data migration as a project. Get a written scope of what the vendor will migrate and what you are responsible for. This is where hidden costs live.

AI features in manufacturing ERP: what is actually useful

Every ERP vendor marketed AI features heavily in 2024–2026. Some of these are genuinely useful for small manufacturers; many are marketing noise. The useful cases so far:

• Demand forecasting — AI-assisted forecasting in products like NetSuite, Epicor and others can reduce forecast error for manufacturers with sufficient historical data (typically 2+ years of transaction history).
• Anomaly detection — flagging unusual purchase order quantities, unusual lead time exceptions, or inventory discrepancies. Useful as a QA layer.
• Natural language querying — asking the system questions in plain English. Early implementations are limited but improving.

Less useful currently: AI-generated production schedules for complex job shops (constraint-based scheduling tools do this better), AI-written supplier emails, and AI chatbots as ERP help desks (they hallucinate system-specific details).