$1.2M nearly spent. Saved in twelve weeks.

This case study shows how a mid-sized manufacturer chose an ERP in twelve weeks and avoided $1.2 million of unnecessary spend. It is for CFOs, COOs and CIOs about to start a selection, or stuck in one. Five things made the difference. Process maps came before any demo. Every vendor ran the same scripted scenarios on the client's own data. A five-year total cost of ownership model exposed costs the price tags hid. Executives scored options separately to expose bias. And the board saw a one-page decision pack. The saving came from cutting scope the business did not need, not from vendor discounts. If you are about to start, map your core processes first.

After 25 years in ERP advisory, I thought I had seen most things. Then a mid-sized manufacturing company in Poland nearly signed off on the wrong system, a $1.2 million mistake in the making. Leadership was overwhelmed by vendor sales tactics, internal politics and a requirements list that had grown into a wish list.

That is when I was brought in. I simplified the selection process, introduced structured evaluation criteria and asked leadership about their real pain points. I also reminded them that ERP selection is a governance and risk decision, not a feature contest.

The result: a confident choice in twelve weeks, $1.2 million of unnecessary spend avoided, and a board that went into implementation with clarity rather than hesitation.

1. 01Requirements vendors skipIntegration, lock-in, readiness, site differences
2. 02Scripted demo daySame workflows, the client's own data
3. 03Five-year cost modelExposed a $500K rise after year three
4. 04Separate scoringCIO, CFO and COO scored apart
5. 05One-page board packBoard aligned in 45 minutes

ERP chosen in twelve weeks, $1.2M of unnecessary spend avoided

The company had revenue just under $400 million. Finance, warehouse and production each ran on separate platforms with very little integration. Reporting was slow and unreliable, and departments had built manual workarounds to keep operations running. The board had made its position clear: it could not wait another year.

The pressures that surfaced when I spent time with leadership:

1. CIO: pushing for cloud capability and long-term scalability
2. CFO: pushing hard on cost control, worried about another budget overrun
3. COO: insisting operational stability came first, worried about production disruption
4. Vendors: using quarter-end discounts and "limited time" offers to force a decision

Each priority made sense on its own. Together they created a standstill. Nobody had hard numbers, and decisions were being driven by vendor slide decks with no cost model behind them.

Staff were exhausted. Some weeks included three or four demos with no evaluation framework. One manager admitted quietly that he no longer remembered which product offered which feature. Decision fatigue had set in, and the temptation to just pick something was real.

The table shows common selection problems and the fix for each.

Common problem	Impact	How to address it
Jumping straight into vendor demos	Teams get distracted by features; process fit is ignored	Document process flows first; run demos with your own data
Unclear requirements from leadership	Conflicting priorities stall decisions and widen scope	Lock must-have and should-have lists with executive sign-off
No five-year cost model	Hidden costs surface after go-live	Build a TCO view covering licences, escalators, integration and support
Integration blind spots	Late-discovered interfaces cause delays and overruns	Map every system early; prove risky interfaces with a proof of concept
Politics and bias	Decisions rest on opinion, not facts	Use weighted scorecards and role-specific criteria

Phase 1: requirements vendors never ask about

Most selections focus on modules and functions. I looked at where integration would matter most, because legacy systems never disappear overnight. I raised vendor lock-in: penalties and future migration costs that no sales team puts on the first slide. And I checked organisational readiness.

In the workshops, I found that inventory practices differed across production sites. Some tracked raw materials in spreadsheets; others used local databases. That gap alone could have become a cutover crisis. Flagging it early meant leadership understood that system selection had to go hand in hand with process alignment. It is the same principle I apply everywhere: processes first, then systems.

Phase 2: a scripted demo day with real data

Vendors prefer to show what they do best, which hides weaknesses. I built a scripted demo day where every vendor walked through the same core workflows, order-to-cash and production planning, using the client's own data in the same sequence. No cherry-picked scenarios.

Some vendors pushed back and said it was unusual. I held firm. The client saw side-by-side comparisons that revealed differences the brochures had hidden.

Phase 3: surfacing hidden costs

Price tags mislead. I built a five-year TCO model. It included the annual escalation clauses that quietly raise fees and the cost of backfilling staff pulled off their day jobs. It also covered hypercare support, which often costs two to three times what vendors first quote.

One "mid-tier" option showed a hidden $500,000 increase after year three. Without the model, the client would have chosen it on headline price. That single exercise justified the engagement.

Phase 4: neutralising internal bias

Even with data, bias creeps in. The CIO leaned toward the cloud-first vendor, the CFO toward the lowest upfront price, the COO toward stability. I ran structured scoring sessions separately with each, then compared the results to show where personal preference had outweighed the facts. With structured scoring, it became harder to say "this one feels better" without evidence.

Phase 5: board alignment

I summarised everything on a one-page executive dashboard: a scorecard, cost ranges and a risk heatmap. No jargon and no long reports. The board discussion took less than an hour, and leadership aligned within 45 minutes. They left confident, not because they had more data, but because the data was structured and transparent.

The table shows the selection steps in order and what each one produced.

Step	Purpose	Outcome
Define business objectives	Agree what the ERP must solve: cost control, compliance, scalability	Leadership and operations aligned on goals
Map current processes	Document workflows, pain points and dependencies	A baseline to judge fit against
Shortlist vendors	Narrow to realistic options	Evaluation focused on relevant candidates
Run fit-gap analysis	Compare each option with business needs	A clear view of customisation, risk and integration gaps
Model total cost of ownership	Licences, implementation, training and five years of support	Financial clarity for budgeting and board approval
Check references	Talk to peers in similar industries on the same ERP	First-hand view of vendor performance

1. ERP chosen in twelve weeks
2. $1.2 million of unnecessary spend avoided by removing modules with no business value and rejecting inflated premium bundles, not by chasing discounts
3. Decision fatigue lifted; leadership aligned and moved into kick-off with agreed expectations
4. Vendor quarter-end deadlines neutralised; the client bought on its own terms

"Noel did not just help us pick an ERP. He gave us a way to cut through politics, see the hidden costs, and make a board-level decision with clarity. In 12 weeks we saved over a million dollars and avoided a decade-long mistake." (CFO, Polish manufacturing group)

The cost model was the most valuable artefact in this engagement. These are the lines it covered and the ones most often missed.

Cost line	What to include	Commonly missed
Software	Licences or subscription for current and projected users and entities	User growth and price escalators after year one
Implementation	Partner fees, internal project team, contingency	Internal backfill for people pulled off day jobs
Integration	Interfaces to retained systems, middleware, testing	Interfaces discovered late in the project
Data migration	Cleansing, mock loads, reconciliation	Legacy data quality work
Training and change	Training design, delivery, floor support	Refresher training during hypercare
Hypercare and support	Post-go-live support and the run model	Hypercare scope beyond the vendor's first quote
Exit and renewal	Renewal terms, data export, migration costs	Lock-in costs when terms change at renewal

For the contract side of the same picture, see my ERP contract negotiation guide for CFOs.

Three things would change for a mid-market manufacturer selecting in 2026.

SAP GROW would be on the shortlist. SAP GROW, built on SAP's public cloud ERP, now sits in the same conversation as NetSuite, Dynamics 365 Business Central and Infor for companies new to SAP. Its fixed-scope GROW Fast offering targets go-live in a few months, and clean core is enforced by design. The TCO model has to compare five-year subscription growth, not perpetual licence plus maintenance.

AI speeds up the paperwork, not the judgement. AI assistants can now draft requirements summaries, scoring frameworks and first-pass fit-gap analyses. Which vendor actually fits the business is still a judgement for experienced people.

Cloud is the default. Some mid-market manufacturers used to weigh cloud against on-premise. In 2026 the practical default is cloud unless data residency or sovereignty rules say otherwise. Compare cloud options head to head in the TCO instead of building an on-premise scenario for each vendor.

The scripted demo day, the hidden-cost TCO, the separate scoring and the one-page board pack all still apply whatever is on the shortlist. My guide to the best ERP for manufacturing covers the vendors in more depth.

Do not rush the decision: quick choices miss hidden costs and integration risks. Get real numbers, not slide decks, with five to seven years of visibility. Balance leadership priorities in structured sessions early, before positions harden. Protect staff from demo fatigue by standardising criteria and rotating participants. And keep the board engaged with short checkpoint updates, so there are no late surprises. If the shortlist includes SAP and Oracle, my SAP vs Oracle comparison sets out the trade-offs.

How should a mid-sized manufacturer start ERP selection without getting lost in demos?

Not with vendor demos. Start by mapping core processes: order-to-cash, production scheduling, inventory management and supplier payments. Once they are documented and agreed, evaluate how each system supports them. I have seen teams spend weeks on polished demos and then realise half of what they watched had no relevance to their operations. Process mapping first usually makes the evaluation shorter and less political.

What belongs in a five-year total cost of ownership model for ERP?

Licences or subscription with user growth, maintenance and escalation clauses, implementation, integration, data migration, testing, hypercare, and backfill for internal staff pulled off their day jobs. In one case, a team approved a system on licence fees alone and found the budget had doubled by year two. The model forces the conversation about escalators and renewals before you are committed to them.

How do we make vendor demos genuinely comparable?

Script identical scenarios on your own data. Every vendor walks through the same workflows, such as order-to-cash, production planning and returns, in the same sequence and time. Weight the scoring by your priorities, not by presentation quality. One manufacturing team asked vendors to process a complex return including credit issuance, and that single test exposed differences the polished demos had hidden.

What risks should a board see before approving an ERP decision?

Data migration failure, integration gaps with retained systems, uneven process adoption across sites, and vendor lock-in that bites when upgrades or renewals come due. A one-page risk heatmap of probability and impact usually changes the conversation, because without it executives assume ERP is only about licences and timelines. Some risks can be mitigated; others should be accepted knowingly.

How do we prevent scope creep during ERP selection?

Separate must-have requirements from optional ones, and link every feature to a measurable business outcome. Add a formal change process: any new requirement must show risk reduction or financial return. I have seen teams approve every request until the timeline doubled. Every extra feature competes with go-live readiness, and leadership has to say so consistently.

How do we handle vendor pressure and quarter-end discounts?

Treat discounts tied to short deadlines as pressure tactics, not savings. The real savings come from removing unnecessary modules and challenging over-licensing; a 10% discount looks less attractive when half the functionality is never used. In this engagement, delaying the decision did not change the available terms, because vendors rarely walk away from a committed buyer. The $1.2 million saving came from scope, not from any discount.