elevate techConsulting Group

Independent ERP and system selection

How do you choose an ERP system? A step-by-step process

Choose an ERP in eight steps: decide who owns the project, map how a sale becomes cash, write requirements, build a shortlist, run scripted demos with your own data, check references, review the contract, and then select the implementation partner. Do the first three before you talk to any vendor.

Most failed ERP projects were lost at the start, not at go-live. The company chose a product before it understood its own process. This guide puts the steps in the order that protects you. It is the process behind our independent ERP and system selection work, and it assumes you are an owner-led, founder-led or PE-backed company of real size, with more moving parts than a spreadsheet can hold.

Step 1: Who owns the selection?

Name one executive who owns the outcome. Not a committee, and not the IT manager by default. Most IT teams spend their days keeping the lights on and do not have the time, or the mandate, to redesign how the company operates. Appointing your IT manager to head technology strategy is like appointing your bookkeeper to be your CFO. The role needs authority over process, budget and people.

Around that owner, build a small team: an operational lead who knows how work really gets done, someone from finance, and a neutral voice on your side of the table. A vendor-led process has a conflict of interest, because the vendor's team works for the vendor.

Step 2: How does a sale become cash?

Before you write a single requirement, draw the path from quote to payment in your business. Use one page and plain words.

The places where information is retyped or emailed are where your pain is. This map also reveals a common finding: often the right system is already in place and used wrong. If the map shows the problem is a missing approval step or an unclear hand-off, a new ERP will not fix it. Most technology problems are people and process, not the software.

Step 3: How do you build ERP requirements?

Write requirements in plain business language, each one a sentence a manager could test: "When a customer places an order in either storefront, the order appears once in the system with the right tax and ship-to." Group them by area, and mark each as must have, should have or nice to have.

Length is not the goal, clarity is. On the multi-brand auto parts distributor project, the company ran thirteen storefronts and two systems of record that disagreed with each other, and the team wrote 240 requirements and scored every platform considered against them. A simpler company may need far fewer. What matters is that every requirement is written down, owned and testable, because the list becomes your scorecard for every later step. For a fuller list of what to capture, use our ERP requirements checklist.

Step 4: How do you shortlist ERP vendors?

Start with a long list from your industry peers, your advisor and your own research. Screen it against your must-haves and your size, your industry and the number of locations or entities. Then cut it down.

Scale the list to the project. In the wireless infrastructure services case study, the company ran an eight-vendor selection for its first ERP and CRM. That was workable because the requirements were written and the screening was disciplined. Eight vendors in full demos would have exhausted the team. Before you invite vendors to demo, aim for three or four.

Not sure whether your current systems are the problem or how they are used? Check first.

Take the Systems Check

Step 5: How do you run scripted demos?

Never let a vendor run its standard demo. Give each finalist the same script, built from your sale-to-cash map, and some of your real data: a messy customer, an unusual order, a return, a split shipment. Ask them to run it live.

"Promised" means it is not there. A feature on a roadmap should not earn the same score as a feature you watched run.

Step 6: What should reference checks cover?

Ask for customers of your size and complexity, including at least one that has been live for more than a year. Call them and ask about the project, not the product. What slipped? What cost more than expected? What did the vendor do when something broke? Would they pick the same system again? Reference checks are the cheapest honest information you will get.

Step 7: How do you review the contract?

Read the statement of work as closely as the price. It should name the people, the deliverables, the acceptance tests, what the customer must supply and what is excluded. Check how change orders work, how data migration is divided, and what the exit terms are. We cover this in detail in the questions to ask an ERP vendor before you sign, and the money side in what an implementation costs beyond the quote. Have counsel review it, and have someone who has run these projects read the scope.

Step 8: How do you pick an implementation partner?

The software and the people who install it are two separate decisions. A good product with a weak partner will disappoint you, and so will the reverse. Ask the partner for the named team, their experience with companies like yours, and their record on the specific modules you need. Meet the people who will be on site. Where possible, choose a partner whose incentives do not depend on selling you more modules. Independence at the selection stage matters because the choice of software shapes everything that follows.

The sequence at a glance

StepWhat you produceWho owns it
1. OwnerOne named executive and a small teamCEO
2. Sale to cashA one-page map with pain points markedOperations and finance
3. RequirementsA written, ranked, testable listSelection team
4. ShortlistThree or four finalistsOwner and advisor
5. Scripted demosScored sheets from live demonstrationsSelection team
6. ReferencesNotes from calls with comparable customersOwner and finance
7. ContractA reviewed statement of work with exclusions listedOwner, counsel, advisor
8. PartnerA named delivery team you have metOwner

What goes wrong most often?

Three patterns, in our experience. Companies skip step 2 and buy a system to fix a process problem. They skip step 3 and let the vendor's demo define what "good" looks like. Or they treat the project as an IT project and not a business one, so nobody with authority resolves the disagreements between departments. Each of these is cheap to fix at the start and expensive to fix after signing.

If you would like help running this process on your side of the table, see how we work on ERP selection.

Common questions

How long does an ERP selection take?

It depends on company size and how clear your requirements are. Expect a matter of months, not weeks, if you do it properly. Rushing the early steps is the most common reason a selection has to be redone.

How many vendors should you evaluate?

Start with a long list, cut to a shortlist of three or four, and run scripted demos only with those. One wireless services company we worked with began with eight vendors. A tight, written set of requirements is what makes a number like that manageable.

Should you pick the software or the implementation partner first?

Pick the software on fit, then choose the partner who will do the work. They are different decisions. A great product with a weak partner fails, and the reverse is also true.

Who should lead an ERP selection?

One named executive who owns the outcome, often the CFO or COO, with an operational lead and a neutral advisor. Appointing your IT manager to head technology strategy is like appointing your bookkeeper to be your CFO.

Not sure where to start?

Take a free three-minute score. If a conversation makes sense, it will offer one.