How to Choose an ERP System: A 12-Step Framework

Choosing an ERP is the single most consequential software decision most growing companies will make. Get it right and you compound efficiency for a decade. Get it wrong and you spend $200K and 18 months learning what you should have asked before signing.

This guide is the framework we use when advising teams. It is independent - no vendor pays us, none of the rankings are pay-to-play, and we'll happily tell you when an ERP is wrong for you.

Step 1: Be honest about why you need an ERP

Before any demo, write down the real reason you're shopping. The honest answer is usually one of:

• "QuickBooks / Tally has hit a wall and we can't reconcile inventory anymore"
• "We're acquiring entities and need consolidated financials"
• "We just had an audit nightmare and our manual processes are the cause"
• "Our investors are asking for monthly KPIs we can't produce"
• "We're growing fast and our data lives in 7 spreadsheets"

If your answer is something vague like "we need to digitize" or "the team wants a system" - stop. ERPs solve specific problems. Without a specific problem, you'll buy software that doesn't fit anything.

Step 2: Map your top 5-7 critical processes

Walk the floor. Not metaphorically - actually walk it. Sit with your AP clerk, your warehouse manager, your CFO. Ask them: "What's the most painful 30 minutes of your week?" The answers map directly to the modules you need.

Common critical processes:

• Manufacturing: BOM management, work order tracking, capacity planning, WIP costing
• Distribution: Multi-warehouse stock, lot/serial tracking, drop-ship, EDI
• Retail: POS-to-back-office sync, returns, multi-store consolidation
• Services: Time tracking, project profitability, milestone billing, resource utilization
• Multi-entity: Inter-company eliminations, FX, statutory consolidation

Write down 5-7 processes, ranked by pain. This is your evaluation rubric. Every demo gets scored against these specific processes - not against shiny features the vendor wants to show you.

Step 3: Define non-negotiables

Hard constraints that disqualify vendors immediately:

• Deployment: Cloud-only? On-premise required? Hybrid OK?
• Data residency: Must data stay in India / EU / US?
• Compliance: SOX, GDPR, HIPAA, GST, e-Way Bill, FDA 21 CFR Part 11?
• Integration: Existing Salesforce / Shopify / Tally that must keep working?
• Headcount: How many concurrent users at launch? In 3 years?
• Budget: Hard ceiling on first-year all-in cost?

A vendor that fails any non-negotiable is out, regardless of how impressive the demo is.

Step 4: Build your shortlist (5-7 vendors max)

Use a directory like ours, talk to peers in your industry, read independent reviews. Build a list of 10-15 candidates, then cut to 5-7 by applying Steps 2 and 3.

Avoid the temptation to invite 12 vendors to demo. You'll exhaust your team and the data will be too noisy to compare. 5-7 is the sweet spot.

Step 5: Send a structured RFP - but make it short

A 90-page RFP is bureaucratic theater. A good RFP fits on 4-6 pages and asks 30-40 questions max:

• 10 questions about the 5-7 critical processes (Step 2)
• 5 questions about pricing structure (per user, modules, implementation, ongoing)
• 5 questions about implementation methodology and timeline
• 5 questions about reference customers in your industry and size
• 5 questions about integration capabilities
• 5 questions about platform / customization (low-code, APIs, sandbox environments)

If a vendor cannot answer in 5 business days, they will not implement on time either. Treat this as a signal.

Step 6: Demos that actually tell you something

Standard vendor demos are sales theater. To get useful data, do this instead:

• Send a use-case packet ahead of time: 3-5 specific scenarios from your business ("Customer A places a sales order, half ships from Mumbai warehouse, half from Delhi, payment terms are 50% advance / 50% net-30"). The vendor demos your scenarios in their system.
• Insist on live keystrokes, not slides. If they can't show actual clicks, the feature probably doesn't work the way the slide implies.
• Have two end-users in every demo, not just leadership. Your AP clerk will spot UX issues that the CFO will miss.
• Run the same demo script across all vendors so you can compare apples to apples.
• Score each demo on a rubric - your 5-7 critical processes (Step 2) plus UX, performance, and customizability.

Step 7: Total cost of ownership, not sticker price

ERP pricing is intentionally confusing. The number on the proposal is rarely the number you pay. Here's the real cost framework:

For a typical mid-market deployment, year-1 all-in cost runs 1.5-2.0x the licensing fee. NetSuite at $50K/year licensing is more like $90-130K all-in for year one. Plan accordingly.

For a 5-year TCO view, multiply year-1 by ~3.5x (steady-state years 2-5 are mostly licensing + ongoing customization + version upgrades).

Step 8: Reference calls - but ask the right questions

Vendors will give you references who love them. That data is worthless. To get useful signal:

• Ask for 3 references, then ask each of them for one more reference each ("who else do you know running this system?"). The second-degree references are the honest ones.
• Ask: "What surprised you in the first 6 months?" (gets you to the real friction)
• Ask: "What would you do differently if you were starting again?"
• Ask: "What does your implementation partner do well, and where do they fall short?"
• Ask: "If you had a hard renewal negotiation, what was the lever that worked?"

Skip the "are you happy with the system" question. Nobody admits they made a $200K mistake to a stranger.

Step 9: The implementation partner is more important than the vendor

This is the single most under-appreciated factor in ERP buying. The same software, implemented by two different partners, produces wildly different outcomes.

When evaluating partners:

• How many implementations have they done in your industry, at your size? ("Several" is not an answer. Names of companies is.)
• Who specifically will be on your project? (Not the SVP at the pitch. The actual senior consultant who shows up day one.)
• What's their methodology? (Agile, sprint-based deployments beat 18-month waterfalls.)
• Can you talk to their last 3 customers - not the highlight reel?
• Are they incentivized to use vendor-provided customizations or custom code? (The latter is more expensive long-term.)
• What does the post-go-live support look like, day 91? (Many partners disappear after launch.)

A great partner with a good-not-great ERP beats a mediocre partner with a great ERP every time.

Step 10: Negotiate price. Always.

ERP pricing has substantial flex - typically 15-40% off the initial proposal. Levers that work:

• Multi-year pre-payment (1-3 years upfront, 5-15% discount)
• End of vendor's fiscal quarter / year (sales reps need the deal)
• Bundling modules (commit to expansion modules in year 2 for a year-1 discount)
• Reference customer commitment (some vendors will discount if you'll be a published case study)
• Locked renewal pricing (this matters more than year-1 discount; cap renewals at CPI or 3-5% annually)
• Implementation cap with milestones (fixed-price + scope changes priced separately)

The number you should negotiate hardest on is renewal pricing. Year-1 discounts feel good but renewals trend up 5-15% annually if you don't lock them. NetSuite is particularly aggressive on this.

Step 11: Plan for the first 18 months realistically

Most companies underestimate the implementation timeline by 30-50%. A realistic mid-market plan:

• Month 0-2: Discovery, scoping, contract finalization
• Month 2-4: Configuration, data migration prep, integration scoping
• Month 4-8: Build phase - configuration, customizations, integration development
• Month 8-10: User acceptance testing, training, parallel runs
• Month 10-12: Go-live, stabilization
• Month 12-18: Optimization, second-wave modules, real ROI starts

If a vendor or partner promises go-live in under 8 weeks for a non-trivial mid-market deployment, they are either lying or about to disappoint you.

Step 12: Decide who owns success after go-live

Day 91 is where most ERP investments quietly fail. The implementation partner has moved to the next project. Your team is exhausted. Bugs surface in edge cases the demo never tested. Support tickets pile up.

Before signing, define:

• Who internally owns the system? (A named person, not "the IT team")
• Who owns vendor escalations? (Same person, with a defined SLA from vendor)
• What's the budget for ongoing optimization? (10-15% of year-1 implementation cost is reasonable)
• When do you do the first full system audit? (90 days post go-live, with the partner)
• What's the upgrade / version migration plan? (Especially relevant for Odoo, ERPNext, on-premise SAP B1)

The companies that get the most ROI from their ERP are the ones that treat it as a continuous program, not a one-time project.

A 12-question gut check before you sign

1. Does this vendor's product actually solve our top 3 process pains?
2. Have we talked to 5+ customers in our industry and size, including 2 second-degree references?
3. Do we have a written implementation plan with milestone-based payments?
4. Have we negotiated renewal pricing, not just year-1 pricing?
5. Do we have a named internal owner for the system?
6. Is our 5-year TCO budget realistic (1.5-2.0x year-1)?
7. Have we tested the data migration with real production data?
8. Have we defined exit criteria - what does "go-live" actually mean?
9. Have we tested mobile / field-worker scenarios that are common for our team?
10. Do we have a rollback plan if go-live fails?
11. Is the implementation partner's senior consultant actually on our project full-time?
12. Have we listed the 5 things that, if missed, will end this engagement in failure?

If you can answer "yes" to all 12, you're ready to sign.

Common mistakes that kill ERP projects

• Buying for features the company doesn't have processes for yet ("aspirational requirements")
• Letting IT lead the buying without operational stakeholders
• Choosing the cheapest implementation partner
• Trying to migrate every legacy customization on day one
• Skipping data cleanup before migration ("we'll do it in the new system")
• Underestimating change-management work
• Not having a project owner with authority to overrule departmental requests

What this guide deliberately does not do

• It doesn't tell you which ERP is best. Best is contextual. We have vendor profiles and industry hubs for that.
• It doesn't push a specific implementation methodology - waterfall vs. agile, both can work with the right partner.
• It doesn't try to replace consulting. If your decision is over $250K total, hire an independent advisor. The fee pays for itself.

Used well, this framework should compress an 18-month chaotic search into a 4-6 month structured one. The discipline of writing things down - critical processes, non-negotiables, scoring rubric - is what separates buyers who get it right from buyers who learn the hard way.