ERP Transformation Field Notes

Better questions create better implementations

Practical observations for executive sponsors and program leaders who want earlier visibility, stronger decisions, and fewer surprises.

RequirementsRequirements are where implementation success is designed—or quietly lostGovernanceFive signals your implementation partner is managing perception—not riskDataData migration is not a technical workstream. It is an operating-model decisionReadinessUAT is not a test phase. It is evidence that the business can operate

Requirements are where implementation success is designed—or quietly lost

Requirements are the foundation of an ERP implementation, yet they are often rushed, delegated, or reduced to a list of requested features. Every configuration choice, integration, conversion rule, test scenario, training decision, and scope discussion depends on whether the business need was understood and expressed clearly enough to act on.

Begin with the business outcome—not the requested feature

“Add a field” or “build an approval workflow” may describe a proposed solution, but not the underlying requirement. A strong requirement explains who needs to accomplish what, in which business context, and why the outcome matters. That context lets the team evaluate configuration, process change, controls, and alternatives instead of automatically recreating the current system.

Describe one clear need without room for interpretation

A requirement should be specific, concise, and understandable to the business owner, implementation partner, and tester. Avoid combining several behaviors into one statement or relying on words such as fast, flexible, automatic, or user-friendly without defining what they mean. If two capable people can read it and design different outcomes, it is not finished.

Include the rules, data, exceptions, and controls

The happy path is rarely where implementations fail. Well-formed requirements identify triggering events, required information, business rules, approvals, security, integrations, exceptions, volumes, timing, audit needs, and regulatory constraints when they apply. These details reveal effort and risk before they become late surprises.

Make acceptance observable and testable

A requirement is not complete until the team can explain how the business will prove it was met. Acceptance criteria should describe the expected result, including meaningful scenarios and exceptions. This creates a direct bridge from requirements to design, configuration, testing, and business acceptance.

Give every requirement an owner, priority, and source

The accountable business owner must be identifiable, and priority should reflect business value, risk, and operational necessity—not who spoke most recently. Documenting the source and rationale prevents old assumptions from being treated as permanent truth and gives governance a basis for resolving conflicts.

Maintain traceability as decisions change

Requirements are not a document that disappears after design. Each important requirement should remain connected to the future-state process, solution decision, configuration or build item, test evidence, training impact, and final acceptance. Traceability makes scope changes visible and helps leaders understand the consequence of saying yes, no, or later.

A well-formed requirement is a shared, testable statement of business intent. When requirements are weak, the program pays repeatedly through redesign, change orders, defects, delayed testing, and disappointed users. When they are strong, every downstream workstream has a firmer foundation.
Built to share

Use the title and opening paragraph as a LinkedIn post, then link readers back to this article for the full point of view.

Five signals your implementation partner is managing perception—not risk

A polished status deck can coexist with a program that is losing control. Client leaders should look beneath the color of the status and test whether the implementation has real decision discipline.

Every issue is “being worked”

A controlled program names the owner, the decision needed, the due date, the impact, and what happens if it slips. Activity without consequence is not risk management.

The integrated plan is not truly integrated

If data, integrations, testing, change, and business readiness live in separate schedules, the critical path is partly fictional.

Change orders arrive before root causes

Some changes are legitimate. But scope growth should be traceable to a requirement, decision, assumption, or genuinely new business need—not simply a budget conversation.

Steering updates omit client obligations

The integrator cannot compensate for missing business decisions, unavailable subject-matter experts, or late data ownership. Honest governance makes client-side risk visible too.

UAT is discussed as a date

UAT readiness is evidence: stable configuration, testable end-to-end scenarios, converted data, trained testers, managed defects, and clear acceptance criteria.

The client-side leader’s job is not to distrust the implementation partner. It is to create enough clarity that trust is supported by evidence.
Built to share

Use the title and opening paragraph as a LinkedIn post, then link readers back to this article for the full point of view.

Data migration is not a technical workstream. It is an operating-model decision

Data conversion is often described as mapping fields from an old system to a new one. That framing is too small. The real question is whether the future business can operate with the data that arrives on day one.

Historical data has a cost

Moving everything feels safe, but it adds mapping, cleansing, testing, reconciliation, and support burden. Retention, migration, and archival should be explicit business decisions.

Open transactions are operational commitments

Open orders, receivables, payables, inventory, projects, and service activity do not become clean because the calendar reaches cutover. Their ownership and reconciliation rules must be designed early.

Cleansing cannot be delegated away

A system integrator can provide tools and structure. Only accountable business owners can decide whether customer, product, supplier, pricing, and financial data are fit for use.

Mock conversions are rehearsals

Each mock should prove more than file movement. It should improve timing, exception handling, reconciliation, business validation, and the cutover runbook.

A technically successful load can still produce an operationally failed go-live. Treat data as a business-readiness program with technical execution—not the other way around.
Built to share

Use the title and opening paragraph as a LinkedIn post, then link readers back to this article for the full point of view.

UAT is not a test phase. It is evidence that the business can operate

User acceptance testing is frequently compressed into a late project activity: provide scripts, schedule testers, log defects, and get signatures. That process may test screens. It does not necessarily prove readiness.

Test end-to-end operating scenarios

The business experiences quote-to-cash, procure-to-pay, plan-to-produce, record-to-report, and service delivery—not isolated configuration objects.

Use realistic converted data

Perfectly prepared test records hide the exact mapping, cleansing, security, and exception problems that surface after go-live.

Define acceptance before testing

Severity, retest rules, unresolved-defect thresholds, workaround ownership, and decision authority should be agreed before pressure rises.

Measure tester capacity honestly

Business experts still have jobs. If the plan ignores the time needed to prepare, execute, document, retest, and decide, the schedule is not credible.

UAT is where the business earns the confidence to accept the system. The signature matters. The evidence behind it matters more.
Built to share

Use the title and opening paragraph as a LinkedIn post, then link readers back to this article for the full point of view.

A program worth protecting

Is one of these risks showing up in your implementation?

Share what you are seeing