How to choose an ERP system: A 10-step guide

Short answer: to choose an ERP system, first agree why you need one and what success looks like, then document your requirements by process, set a realistic budget, build a long list of five to eight products that fit your size and industry, cut it to two or three with an RFP, run scripted demos using your own scenarios, evaluate the implementation partner as carefully as the software, score everything in a weighted matrix and check references before you negotiate. The whole selection usually takes a few months for a mid-sized company, and it is time well spent: the wrong fit is the most expensive ERP mistake to fix.

This guide lays out that selection process step by step, with the questions to ask, a scoring matrix you can copy, and the mistakes that trip up first-time buyers. It is written for finance, operations and IT leaders choosing their first ERP or replacing an old one.

What is an ERP system, and do you need one?

An ERP (enterprise resource planning) system runs finance, purchasing, inventory, sales orders and often manufacturing and projects on one shared database. If you are not sure whether you have outgrown accounting software, our guide to what ERP is covers the modules, the types of ERP and the warning signs. Common triggers are multi-entity consolidation in spreadsheets, unreliable stock figures, slow month-end close and a growing tangle of disconnected apps.

Step 1: Define goals and build the selection team

Start with the business problems, not the software. Write down three to six measurable goals, for example:

• Close the books in five working days instead of fifteen
• Know stock by location in real time, with count accuracy above an agreed target
• Consolidate three entities without spreadsheets
• Remove manual re-keying of ecommerce orders

Then form a small selection team: an executive sponsor (often the CFO or COO), a project lead, one process owner from each major area (finance, purchasing, sales, warehouse, production) and someone from IT. The people who will live with the system must be in the room when it is chosen.

Step 2: Map processes and write requirements

Document how work flows today, where it breaks and what you need instead. Organise requirements by end-to-end process so they translate directly into demo scenarios:

Process	Example requirements to capture
Record to report	Number of entities and currencies, consolidation, intercompany, dimensions for reporting, close checklist, audit trail
Procure to pay	Requisitions and approvals, PO matching, supplier portal, AP automation, landed costs
Order to cash	Quotes, pricing rules, credit control, multichannel orders, returns, invoicing, collections
Inventory and warehouse	Locations and bins, lot or serial tracking, cycle counts, barcode scanning, costing method
Plan to produce (if you manufacture)	BOMs and routings, MRP, capacity, shop floor data capture, quality, job or process costing
Projects and services	Project budgets, time and expense, milestone billing, revenue recognition
Technical and security	Single sign-on (SAML or OIDC), user provisioning (SCIM), role-based access, API access, data residency, audit logs
Integrations	Ecommerce, CRM, payroll, banks, tax engines, EDI, shipping and WMS
Reporting	Standard financial statements, operational dashboards, self-service reporting, export to BI software

Prioritise each requirement as must have, should have or nice to have. Keep the must-have list short and honest; if everything is a must-have, nothing is.

Step 3: Set a budget based on total cost of ownership

Decide what you can afford over three to five years, including implementation, data migration, integrations, training, internal time and support, not just the subscription. Our ERP cost guide explains the pricing models vendors use and includes a TCO worksheet. A realistic budget also keeps the long list honest: there is little point demoing products you cannot afford to implement.

Step 4: Decide on deployment and architecture

• Cloud or on-premises? Cloud (SaaS) ERP is the default for most new small and mid-market projects: no servers, vendor-managed upgrades, access from anywhere. On-premises or private hosting still suits heavy customisation, plant-floor systems with poor connectivity, or strict data location rules.
• Suite or best of breed? Decide which functions must live inside the ERP and which can stay in specialist software (CRM, HR and payroll, advanced warehouse, planning) connected by integrations.
• One system or two tiers? Groups with a large head office system sometimes run a lighter ERP in subsidiaries and consolidate results centrally.
• Generic or industry ERP? Manufacturers, distributors, nonprofits, construction and food businesses often get a better fit from an industry edition or a partner’s industry add-on.

Step 5: Build a long list

Aim for five to eight products that fit your size, industry and deployment choice. Sources include peer recommendations, industry associations, your accountant or auditor, and comparison sites. Our guide to the best ERP software groups the main products by company size and industry, and our ERP software category lists the wider market. If you are UK-based, see our guide to ERP software for UK businesses.

Quick filters that remove poor fits early:

• Does the vendor serve companies of your size and industry today?
• Does it support your countries, currencies and tax requirements?
• Is there a partner with relevant experience in your region?
• Is the pricing model likely to fit your budget and user mix?

Step 6: Send an RFI or RFP and cut to two or three

A request for information (RFI) or request for proposal (RFP) turns your requirements into a structured questionnaire. Keep it focused: vendors answer long generic RFPs with “yes” to almost everything. Useful sections are:

• Company background, scope, users by role, entities and sites
• Your prioritised requirements, asking whether each is standard, configurable, available from a partner add-on or requires customisation
• Integration and security questions
• Implementation approach, timeline, team and method
• Pricing for software and services in a format you specify, over three to five years
• Two or three reference customers similar to you

Use the answers to cut the list to two or three finalists.

Step 7: Run scripted demos

Do not let vendors run their standard demo. Send each finalist the same scripts, based on your real processes and data, a week or two in advance. For example: “Show a customer order for a lot-controlled item shipped from two warehouses, invoiced in euros, with the cost and margin visible in reporting.” Have each process owner score the part of the demo that affects them, using the same scoring scale.

Ask to see:

• How a user performs daily tasks, not just dashboards
• How configuration is done (by the admin or only by the partner)
• How reports are built and changed
• How upgrades affect configuration and extensions
• The mobile and approval experience for occasional users

Step 8: Evaluate the implementation partner

For most mid-market products the partner does the implementation, so you are really choosing a software and partner combination. Meet the consultants who will work on your project. Ask how many similar projects they have delivered in the last few years, how they handle change requests, what their testing and training approach is, and how they support customers after go-live. Our ERP implementation guide explains what a good partner-led project looks like.

Step 9: Score the finalists

A weighted scoring matrix keeps the decision objective and gives the steering group something concrete to discuss. Adjust the weights to your priorities; the ones below are only an example.

Criterion	Example weight	What to score (1 to 5)
Functional fit to must-haves	30%	Share of must-haves met with standard or configured features
Usability	10%	Process owners’ scores from scripted demos
Implementation partner	15%	Relevant experience, team quality, method, references
Total cost of ownership (5 years)	15%	Relative TCO including services and internal costs
Integration and technology	10%	APIs, connectors, SSO and provisioning, security posture
Scalability and roadmap	10%	Support for more entities, users, countries and volume; product direction
Vendor viability and support	5%	Customer base in your segment, support model, partner network
Contract terms	5%	Renewal caps, flexibility, data export and exit terms

Multiply each score by its weight and add up. If two products end up close, the partner and the references usually decide it.

Step 10: Check references and negotiate

Speak to at least two customers of each finalist who are similar in size and industry and who used the same partner. Ask what they would do differently, how long the project really took, how the first month-end close went and how responsive support has been since.

When negotiating, look beyond the headline discount:

• Caps on price increases at renewal
• Ramped user counts that match your rollout
• Included sandbox or test environments
• Clear statement of work with deliverables and assumptions
• Data ownership, export formats and exit assistance

How to choose an ERP for a small business

Small businesses can run a lighter version of the same process. Keep the requirements list to one or two pages, look at three products, and insist on a scripted demo of your two or three most important processes. Favour cloud ERP with a strong partner or vendor onboarding package, standard processes and a pricing model that suits a small number of full users. Check how easily you can add modules, users and entities later, because the right small business ERP is one you will not have to replace after two years of growth.

Common ERP selection mistakes

• Buying the demo, not the fit. Polished dashboards are not the same as handling your order and costing processes.
• Letting one department choose. ERP touches everyone; a finance-only selection often fails in the warehouse, and vice versa.
• Ignoring the partner. A good product with the wrong partner is a bad project.
• Comparing subscriptions, not TCO. Services, integrations and internal time often outweigh licence differences.
• Over-specifying. Hundreds of must-haves lead to customisation and higher cost.
• Skipping references. Customers like you are the most reliable guide to what the project will really be like.

ERP selection checklist

• Goals and success measures agreed by the sponsor
• Selection team with a process owner for each area
• Requirements documented by process and prioritised
• Five-year TCO budget approved
• Deployment and architecture choices made
• Long list of five to eight products, filtered to two or three finalists
• Scripted demos scored by process owners
• Implementation partners evaluated
• Weighted scoring matrix completed
• References checked and contract terms negotiated

Frequently asked questions