How to Select an ERP System: A Practical Evaluation Framework

ERP selection affects every department, every process, every person. It’s a 5-10 year commitment. Total costs run from hundreds of thousands to millions depending on company size and complexity. Getting it wrong is expensive. Getting it right changes how the business operates.

I’ve watched companies agonise over this decision for 18 months and still get it wrong. I’ve also watched companies nail it in three months because they followed a structured process. The difference is almost always methodology, not luck.

TL;DR: Fifty per cent of ERP implementations fail on their first attempt, but companies that performed an ROI analysis beforehand met their expectations 83% of the time (Panorama Consulting, 2025). ERP selection comes down to five decisions: what processes to cover, which deployment model (cloud vs. on-prem), which vendor, which implementation partner, and how to measure success. Define requirements before looking at products. Most selection mistakes start with choosing a vendor before understanding your needs.

How Do You Define ERP Requirements?

The ERP market is valued at approximately USD 73 billion, with Oracle, SAP, and Microsoft holding over 70% market share (Verified Market Reports, 2025). Choosing between them starts with understanding what you actually need.

Process scope. Which business processes will the ERP manage? Finance, procurement, manufacturing, sales, HR, supply chain? List them and mark which are critical (must have in phase 1) vs. desirable (can add later). Be crisp about this distinction. I’ve seen companies throw 15 modules into phase 1 and wonder why the project takes three years.

Integration needs. What existing systems must the ERP connect to? CRM, e-commerce, warehouse management, third-party logistics? Identify each integration, the data that flows, and the frequency (real-time vs. batch). This is where people underestimate consistently.

Industry requirements. Some industries have specific needs: pharma companies need batch tracking and compliance reporting. Manufacturers need BOM management and shop floor control. Service companies need project accounting and time tracking. If your vendor doesn’t have these built in, you’ll be building custom from day one.

User population. How many users? What roles? Which modules does each role need? ERP licensing typically varies by user type (full user, self-service user, API-only user), so the user profile directly affects cost. Get this wrong and your budget is fiction.

Non-negotiables. Identify the 15-20 requirements that will eliminate vendors. These are typically specific industry functionality, geographic/language support, regulatory compliance capabilities, and integration with existing systems. Every vendor can tick 200 boxes on a feature list. These 15-20 items are the ones that actually determine your shortlist. Everything else is noise.

How Do You Evaluate ERP Vendors?

Only 48% of digital initiatives meet their business outcome targets (Gartner, 2024). The vendor evaluation process directly affects whether your ERP initiative joins the successful half.

Step 1: Longlist (5-7 vendors). Based on your non-negotiables, create a longlist. Use analyst reports (Gartner Magic Quadrant, Forrester Wave), peer recommendations, and industry associations to identify candidates. Don’t skip the peer recommendations. Analysts tell you what’s possible. Peers tell you what’s real.

Step 2: RFI (Request for Information). Send a structured questionnaire to longlist vendors. Focus on fit: do they serve your industry, your company size, your geographic markets? This narrows to 3-4 vendors.

Step 3: RFP and demos. Send detailed scenarios (not feature checklists) to shortlisted vendors. Ask them to demonstrate how their system handles your specific processes. Nota bene: script the demo. Don’t let vendors show their standard presentation. Vendors are excellent at showing their strengths. Scripted scenario demos reveal the gaps that polished presentations hide. I’ve sat through vendor demos where they glossed over three dealbreakers in 90 seconds. Scripted scenarios make that impossible.

Step 4: Reference checks. Talk to 2-3 customers of each vendor, ideally in your industry and of similar size. Ask about implementation duration, hidden costs, ongoing support quality, and whether they’d choose the same vendor again. That last question tells you everything.

Step 5: Total cost analysis. Compare total cost of ownership (TCO) over 5 years, including licences/subscriptions, implementation consulting, data migration, customisation, training, and ongoing support.

Cloud ERP vs. On-Premise: How Do You Choose?

Worldwide public cloud spending is forecast at $723.4 billion in 2025, up 21.5% from 2024 (Gartner, 2024). Cloud ERP is the default choice for most new implementations, but it’s not the only option.

Cloud ERP (SaaS): the vendor hosts and manages the system. You access it through a browser. Updates are automatic. Customisation is limited to configuration and API-based extensions. Examples: SAP S/4HANA Cloud (public edition), Oracle Fusion Cloud, Microsoft Dynamics 365. Good fit for companies that want to stay standard and don’t have deep IT infrastructure teams.

On-premise ERP: you host and manage the system on your own infrastructure. Full customisation control. You decide when to update. Higher TCO but maximum flexibility. Examples: SAP S/4HANA (on-premise), Oracle E-Business Suite. I’d only recommend this now if you have very specific regulatory or customisation requirements that cloud can’t meet.

Managed private cloud: the vendor hosts a dedicated instance for you. More customisation than public cloud, less operational burden than on-premise. Examples: RISE with SAP (S/4HANA private edition). This is the sweet spot for a lot of mid-market companies right now.

The choice depends on how much you need to customise, whether you want to manage infrastructure, your data residency requirements, and your budget model preference (OpEx vs. CapEx).

What Are the Common ERP Selection Mistakes?

Fifty-one per cent of companies experience operational disruptions at ERP go-live (Panorama Consulting, 2025). Many of those disruptions trace back to selection mistakes made months earlier.

Choosing the vendor before defining requirements. “We’re going with SAP” or “We want Dynamics” before anyone has mapped out what’s needed. The vendor choice should follow from requirements, not precede them. I reckon this is the single most expensive mistake in ERP selection. It’s also the most common.

Ignoring total cost of ownership. Comparing licence costs without factoring in implementation, customisation, training, and 5-year maintenance. The cheapest licence often comes with the most expensive implementation. I’ve seen a client pick a vendor that was 30% cheaper on licences and end up 50% over budget because of customisation costs. The maths don’t work the way procurement expects.

Underestimating change management. ERP changes how people work. If the selection process doesn’t include end-user representation, the chosen system may technically fit but operationally fail because users reject it. People aren’t features. They have habits.

Scope creep during selection. Starting with “we need finance and procurement” and ending with “let’s add HR, CRM, and warehouse management to phase 1.” More scope means longer timelines and higher risk. Start lean. You can always add modules later.

Skipping the implementation partner evaluation. The product is only as good as the team configuring it. Evaluate partners as rigorously as you evaluate vendors. Check references, team composition, and engagement model. A good partner can make a decent product great. A dodgy partner can make a great product terrible.

FAQ

How long does ERP selection typically take?

A thorough selection process takes 3-6 months: 4-6 weeks for requirements, 4-6 weeks for RFI/RFP, 4-6 weeks for demos and references, and 2-4 weeks for final decision and contracting. Rushing shortens the process but increases the risk of a wrong choice.

Should I hire a consultant for ERP selection?

For your first ERP or a major platform change, yes. An independent advisor (not tied to a specific vendor) helps structure the evaluation, ask the right questions, and avoid vendor-influenced bias. Budget CHF 30,000-80,000 for selection advisory services.

How do I compare SAP, Oracle, and Microsoft?

SAP excels at complex, multi-entity, multi-country scenarios, especially manufacturing and distribution. Oracle is strong in finance and cloud-native architecture. Microsoft Dynamics integrates well with the Microsoft ecosystem (Office, Azure, Teams) and suits mid-market companies. The best choice depends on your industry, size, and existing technology stack.

What’s a realistic ERP implementation budget?

For a mid-size company (200-1,000 users): CHF 500,000-2,000,000 including licences, implementation, data migration, training, and first-year support. Enterprise organisations (1,000+ users, multiple countries): CHF 2,000,000-10,000,000+. These ranges assume a phased approach. Implementing everything at once costs more.

When should I replace vs. upgrade my current ERP?

Replace when your current vendor no longer supports the product, your business has fundamentally changed (new industries, new geographies), or customisations have made upgrades impossible. Upgrade when the core system fits your needs and the vendor offers a clear migration path (e.g., ECC to S/4HANA).