Before You Sign the ERP Contract and SOW - An Independent Decision Check

Updated: Oct 22

Updated in October 2026 with the eight areas we review before signature and what our second opinions found.
What should you settle before you sign the ERP contract and SOW?
Before signing, I want clear answers about scope, integrations, testing, data migration, staffing, and support after go-live. I also want the commercial assumptions to agree with the delivery plan. A user count in an order form, for example, should have a basis in how people will actually use the system.
Choosing an ERP and an implementation partner leaves several important decisions to settle before the work begins. The proposal needs to describe the implementation the company expects to buy, and the company needs to understand what it must contribute to make that implementation possible.
We call this a pre-implementation diagnostic. It is part of the decision-and-contract phase when we lead an ERP selection, and it can run separately when a company has already chosen its platform. The outcome may be a recommendation to proceed, to proceed after specific conditions are resolved, or to reconsider an important part of the approach. Still choosing? See how our ERP selection program works, or talk to John.
Why the gap matters
Implementation partners price and plan against assumptions. Some are reasonable and explicit. Others remain buried in a proposal or are understood differently by the client and the delivery team.
A statement of work that assigns data cleansing to the client has not established whether anyone has time to do it. A reference to "integration support" has not established who will build the interface or resolve a failure across systems. Those gaps affect cost and timing whether or not they appear as separate lines in the quote.
The purpose of the review is to identify material gaps while the parties can still agree on responsibilities, scope, and commercial treatment. Our operational review supports the client's legal review by making the intended work and business expectations clear.
What a diagnostic settles
First, define the initial phase and the work being deferred. The business should understand what it can do at the first go-live, what remains in existing systems, and how later additions will be approved.
Retained systems need explicit treatment. If CRM, engineering, quoting, HR, or expense tools will remain, establish how information will move between them and the ERP. A temporary manual process may be acceptable, but it needs an owner and a plan for checking the results.
The review should also test the assumptions that matter to the operation. A manufacturer may need to confirm engineering data, bills of material, revisions, and shop-floor transactions. A distributor may need to examine pricing, rebates, branch structures, handheld workflows, and labels. Regulated businesses need an agreed approach to validation and the records their processes require.
Reporting deserves the same attention. Separate the reports needed to operate and control the business at go-live from enhancements that can follow. A report required to reconcile inventory or complete a financial close cannot be deferred simply because it appears under "reporting" rather than a functional workstream.
Finally, put the partner's plan beside the client's real availability. Identify the process owners, decision makers, and people responsible for preparing data, testing, and training. Agree what must be true before go-live and how the steering committee will assess that evidence.
Eight areas I review before signature
1. Integration ownership
List the interfaces and identify the responsibilities for design, development, testing, monitoring, and ongoing support. Where several vendors or internal teams contribute, name the person accountable for coordinating the full process.
The parties do not all need to report to one supplier, but they do need a shared definition of success. Confirm how a transaction will be tested across systems, who resolves defects, and how an issue will be managed when its source is uncertain. General promises to assist with integration leave too much open.
2. Testing ownership and client effort
Establish who prepares the test scenarios, who provides data, who executes the tests, and who approves the result. The plan should cover the business process across applications as well as individual features.
Client process owners need time to validate that the solution works for the business. Ask the partner to estimate that participation by phase, then compare it with the team's availability. Agree how defects are prioritized, corrected, and retested, and what evidence supports completion.
3. Data migration responsibility
Extraction, cleansing, mapping, loading, and reconciliation are different tasks. Assign each one, identify the data sets involved, and agree the planned migration cycles.
A proposal should explain what the partner will do and what depends on the client or another provider. The client needs an owner who can judge whether the converted information is complete and usable. Agree how opening balances, inventory, and open transactions will be reconciled before the final conversion becomes a cutover dependency.
4. Hypercare coverage and exit conditions
Define the support model for the period immediately after go-live. Specify coverage hours, key roles, onsite needs, issue triage, escalation, and the transition to ongoing support.
The appropriate duration depends on the rollout, operating hours, locations, and processes being introduced. Our RFPs ask vendors for at least four weeks as a starting point. A business may need support through its first financial close or another important operating cycle. Establish what happens if critical issues remain at the scheduled end, including how any extension is staffed and priced. A number of weeks without those details does not describe the service.
5. Environments and sandboxes
Confirm which environments the project needs and which are included in the proposed subscription or services. Consider configuration, testing, migration rehearsals, training, and support after go-live.
Ask who provisions and refreshes them, when they will be available, and what additional environments cost. The training plan should not depend on an environment that will be overwritten for migration work during the same week. Check the relevant order form and service terms alongside the implementation schedule.
6. Key delivery resources
Ask for the proposed key people or, where assignments cannot yet be made, role profiles with minimum experience and an agreed approval process. Clarify availability, responsibilities, replacement arrangements, and how knowledge will be transferred if someone changes.
One selection in our records required additional meetings with the proposed delivery team before the client was comfortable signing. The software evaluation alone had not answered the staffing question. The commitment should describe who is expected to deliver the work and how the client will evaluate a proposed replacement.
7. Subscription terms and user definitions
Review user types, committed quantities, additional-user pricing, annual increases, renewal provisions, and the rights to change quantities or license levels. These terms often sit in the subscription agreement or order form rather than the implementation SOW.
Do not assume that additions retain the original discount or that reductions can be made whenever the business requests them. Establish what the specific agreement allows, when changes take effect, and whether notice is required. Reconcile those commitments with the phased rollout and expected use. Our five-year cost comparison shows where those terms land over the full term.
8. Services cost and change control
Understand whether the services proposal is a fixed fee, a time-and-materials estimate, or an agreed not-to-exceed amount. Each needs defined scope, deliverables, assumptions, client dependencies, and acceptance criteria.
Where the scope is sufficiently developed, seek a fee structure that gives the company useful cost protection. For work that remains uncertain, agree how it will be investigated, estimated, and authorized. Require visibility into spending and the forecast to complete, with written approval for scope or budget changes. A low estimate without enough detail to explain the work gives the client little basis for managing delivery.
Two lessons from work after signature
In one implementation, an integrated order process supported substantially less of the expected order flow than intended. The issues included fragmented integration ownership, insufficient review of the integration delivery arrangement, and a lack of complete end-to-end testing.
The lesson is to assess the delivery model with the same care as the product. The selection needs to establish who will make the full transaction work and how the client will see it demonstrated before relying on it.
In another program, onsite hypercare arrangements were still being settled in the weeks before go-live because the support model had been left as an option. That put decisions about availability, coverage, and cost under pressure just as the team needed to focus on readiness.
Clear agreements can reduce these problems. They still need to be carried into the project plan, resourced, and managed through execution.
When the client has already chosen a direction
A second opinion does not automatically mean reopening the selection. Three examples from our work illustrate the range of questions a review can answer.
One company had selected a platform and wanted an assessment of the proposal. We reviewed the approach, saw a viable path, and continued in the project management role.
Another asked its incumbent reseller to price an upgrade against a requirements register we prepared. Further discovery was needed before the reseller could quote the work. The client ultimately chose an on-premise upgrade with a clearer understanding of what it would and would not receive.
A third came to us expecting to replace its ERP. The assessment found that some missing capabilities were already available but unused, while other gaps could be addressed with additional applications. We recommended continuing with the current system subject to conditions and identified the circumstances that would justify revisiting replacement.
Across these three reviews, two recommendations were to proceed and one was to proceed with conditions. None required a new selection. Those outcomes reflect the evidence in those engagements. The value of an independent review is a supported recommendation, including changes to the proposed direction when they are necessary.
How the review runs
Start with the requirements, proposals, draft agreements, project plan, and open decisions. Working sessions should concentrate on the assumptions most likely to change scope, cost, timing, or readiness.
The appropriate duration and level of participation depend on how much is already known. Our full diagnostics have typically run four to seven weeks, depending on the size of the program, while a document review of a defined upgrade can take far less. Agree the questions, participants, deliverables, and effort before starting.
The review should finish with a clear recommendation, an updated scope and responsibility model, and a list of unresolved items with owners and decision dates. Some uncertainties may require a bounded discovery phase before a larger commitment. Make that explicit rather than implying that an estimate is a settled delivery obligation.
When implementation begins, the agreed plan, risks, decisions, and readiness measures should become part of the program's normal management. The review is useful only if its conclusions are carried into the work.
If you are about to sign, or still choosing, talk to John about your ERP program.
Frequently asked questions
Is this part of an ERP selection?
Yes. It connects the selection recommendation with the implementation commitment. It can also be a separate review when a company chose the platform without us or is deciding how to improve a system it already owns.
Who should lead the review?
Someone representing the client who understands how these implementations are delivered. The partner contributes its proposed solution and methodology. The client's business owners remain responsible for priorities, internal commitments, and approval of the operating approach.
Which contract topics deserve particular attention?
Integration ownership and hypercare have exposed important gaps in engagements we have reviewed. Testing, data migration, staffing, environments, licensing, and services assumptions also need attention. The priority depends on the project's actual risks.
How much time will our team need?
As a guide, process owners have typically spent four to six hours a week in the working sessions, IT six to eight, and the sponsor about an hour a week plus the readout. Plan participation around the questions that remain open. Process owners need time to validate scope and responsibilities. IT needs time for data, architecture, and integrations. The sponsor needs to resolve decisions. Agree that effort with the review plan.
What follows the diagnostic?
The agreed scope, plan, responsibilities, open decisions, and readiness measures provide the starting point for client-side implementation management. The company should know who will maintain them and how issues will reach the people authorized to decide.
Related reading - What Actually Moves the Price - ERP Contract and License Negotiation From the Client Side · Five-Year Cost of Ownership - How Finance Should Compare ERP Proposals · When an ERP Implementation Stalls - How Program Recovery Works · Why Outsourcing the “Internal” ERP Project Manager Often Saves the Program


