The Case Management Replacement
A government agency is replacing a 15-year-old case management system. The fixed-price contract was signed at $2.1 million with a 14-month timeline. Six months in, the vendor reports that the legacy data is far messier than the sample used for estimating: duplicate client records, free-text fields that must be mapped to structured codes, and 40% more records than expected. The vendor requests a change order for $600,000 and four additional months. The agency's project manager has been tracking schedule but has no independent estimate of the data migration effort, and requirements were signed off after only two workshops with frontline staff. Frontline supervisors say the new workflows will not match how cases are actually handled. The agency's fiscal year ends in five months, and unspent capital funds cannot be carried forward.
1. Assess the project's position. Which issues are scope problems, which are estimation problems, and which are governance problems?
Read the worked answer
Scope problems. The mismatch between designed workflows and actual practice is a requirements defect. Two workshops with frontline staff is thin elicitation for a system that encodes their daily work; the supervisors' objection is a scope gap surfacing late, and it will generate rework in both build and training.
Estimation problems. The vendor estimated data migration from a sample that understated volume by 40% and understated quality problems. That is a shared estimation failure: the agency provided the sample and accepted the estimate, and the vendor priced against it. The missing independent estimate is the agency's own gap - without one, it cannot judge whether $600,000 and four months are reasonable or inflated.
Governance problems. The most serious. A fixed-price contract with a 14-month timeline and no apparent data-quality contingency, no independent cost verification, and a fiscal deadline that pressures a decision made under incomplete information. The fiscal-year rule creates an artificial incentive to accept or reject quickly rather than correctly.
Framing. The request is probably legitimate in substance - the data really is worse - but its magnitude is unverified. The agency should not treat this as a binary accept/reject.
2. Recommend how the agency should respond to the change order and manage the fiscal deadline.
Read the worked answer
1. Verify before negotiating. Commission a rapid independent assessment of the data migration effort - a bounded sample-based re-estimate, not a full audit. Compare it to the vendor's $600,000 and four months. This replaces opinion with evidence and is the single highest-value action.
2. Separate the change into components. Data volume, data quality remediation, and workflow redesign are different changes with different owners. Some data quality work may belong to the agency (cleaning source records) rather than the vendor. Splitting the request allows partial approval and reduces the total.
3. Negotiate on shared responsibility. The vendor estimated from the agency's sample; both parties own the miss. Options include a shared-cost split, a capped change order with a gain-share if migration finishes early, or phasing migration so a first tranche lands before the fiscal deadline.
4. Address the workflow gap now. Run structured workshops with frontline staff before build hardens further. Late discovery is cheaper than post-launch rework, and ignoring it will produce a system people work around.
5. Manage the fiscal deadline honestly. Do not let the carry-forward rule drive a bad contract decision. Explore whether funds can be obligated to a defined, deliverable milestone before year-end - for example, a funded migration phase - rather than committing the full amount under pressure. Escalate the constraint to finance leadership with the trade-offs stated plainly.
6. Reset governance. Add independent cost review, a change control board with defined thresholds, and monthly data-quality reporting.