Build a scorecard around evidence you can inspect
Compare legacy modernization service providers by the same business workflow, migration boundary and acceptance packet. Define disqualifiers before reading proposals. Examples include unavailable data export, no named technical owner or a cutover plan with no treatment of intervening writes.
Use evidence states such as demonstrated, documented, proposed and unresolved. If you assign weights, agree them with the business owners first and keep the raw evidence beside the score.
A low price must not cancel a failed mandatory requirement. This editorial worksheet contains no measured company ratings.
Ask for discovery outputs that change the decision
| Question | Requested evidence | Decision it supports |
|---|---|---|
| Which business behavior must survive? | Representative inputs, outputs and exception cases | Scope of the replacement |
| Which dependencies cross the boundary? | Inventory with owners and observed execution paths | Feasibility and wave order |
| Which option removes the constraint? | Retain, wrap, replatform, refactor and replace comparison | Choice of intervention |
| What remains unknown? | Assumption register and bounded investigation plan | Estimate confidence |
Ask the provider to walk through an anonymized assessment sample and explain which finding changed the plan. A large slide deck with no traceable decision is weak evidence of technical due diligence.
Download the evaluation worksheet and record the owner, evidence and decision for each row. Its review fields are blank; it is a reusable template, not a completed assessment.
Review the team and delivery sequence together
Meet the proposed technical lead and the people responsible for domain analysis, data migration and operations. Request disclosure of subcontractor responsibilities and a substitution plan. A company's case study may describe a team that is not available for your work.
Compare the first migration slice, its dependencies and continuation gate. A refactor and a replatform proposal can be legitimate alternatives, but their accepted outcomes differ. Require each to show behavior checks, operational ownership and a rehearsed return path. Knowledge transfer should happen through shared work and customer-operated releases rather than only a final document handoff.
Test data and operational assumptions before cutover
Require a data-authority map for coexistence, reconciliation rules and treatment of writes made during the transition. Ask for the failure cases that stop cutover, the decision maker and the forward-repair option after rollback becomes unsafe. Review disaster recovery separately from application rollback. They solve different failures.
Security evidence should match the data, access and deployment scope. Identify who supplies test data and how private source and customer records are handled. Legacy decommissioning needs evidence of removed consumers, approved historical access and an operating owner. Accepting a replacement does not prove that the old system can be shut down.
Normalize estimates before negotiating the headline price
Put discovery, implementation, integration, data work, parallel operation, support and retirement into the same comparison period. Label each mandatory item included, excluded or unresolved. Add licenses and customer staffing to the ownership-cost view. Unresolved work is a risk to investigate, not a zero-cost line.
Review milestone evidence, change requests, warranty scope, support coverage and exit deliverables. For fixed price, identify the scope boundary. For time and materials, define spending visibility and stop decisions. Warning signs include acceptance based only on activity, inaccessible source or operational data, and assumptions with no owner. Resolve these with the procurement and contract reviewers. This checklist does not establish legal sufficiency.
Use the pilot and references to challenge the proposal
Ask reference customers what the assigned team actually delivered, which responsibilities remained internal and what was still unfinished after go-live. For a case study, distinguish stated business outcomes from evidence available for your own assessment. Obtain permission for the reference discussion.
Before the pilot, record input data, expected behavior, allowed differences, failure cases and acceptance owner. Afterward, preserve actual results and unexplained gaps. The selection decision should identify why the accepted evidence fits your project and what must still be resolved. The modernization provider shortlist supplies candidates. This scorecard supplies the comparison method.
Sources
Documentation checked .
Continue the conversation
Comments (0)