An under-delivering vendor is a supplier decision, not a project verdict
Keep the project question separate from the supplier question
When an external software vendor keeps missing commitments, the question on a CTO’s desk is often framed too broadly: is the project failing? This guide deals with a narrower decision: whether to rescue the vendor relationship or replace the vendor. Replacing a supplier can preserve the software investment; whether to stop the project itself is a separate decision and outside this guide. A proportional delivery model assessment should be completed before deciding to outsource, insource or re-procure.
Make the decision inspectable
You need an objective standard to evaluate the incumbent. I recommend structuring that standard as the Vendor Recovery Evidence Framework. It uses four fixed criteria: Delivery reliability, Technical control, Working relationship and Transition exposure. This framework is an organizing method, not a validated scoring instrument.
You should rescue an under-delivering vendor when evidence supports a credible recovery plan. You should replace them when persistent delivery or control failures outweigh the documented costs and risks of transition.
Build an evidence pack before judging the incumbent
Separate observable failures from disputed explanations
Assemble a shared evidence pack:
- Gather acceptance records, release history, and defect reports.
- Document access ownership, contractual obligations, and your decision log.
- Record the source, owner, and relevance of each item.
Separate observed facts from interpretations and unresolved questions. Vendor failures often stem from multiple sources. Ask what the client supplied late or changed. Review exactly what the vendor committed to deliver before attributing responsibility.
Use the Hertz and Accenture dispute with evidentiary care
According to Gibson Law Partners’ account of the complaint, Hertz alleged that Accenture failed to deliver the website or app, delayed the go-live date twice, disregarded code requirements, and created code with security vulnerabilities and defects. These were allegations in the filed complaint, not judicial findings.
One ruling shows why records matter: the judge dismissed Hertz’s claim that Accenture misrepresented its RAPID expertise because Hertz had not provided specific documentation of that claimed expertise.
Commenting on the dispute, Mark Bachmann wrote: “This disagreement is less an indictment of the consultant model and more of a wake-up call to slow down and do a better job scoping a project”.
The case does not establish a general rescue-or-replace rule. It does show how much any later dispute depends on what both sides wrote down at the time.
Apply the Vendor Recovery Evidence Framework
Delivery reliability
Use Delivery reliability to examine accepted work against commitments. Inspect production defects and release records. Evaluate the credibility of revised delivery promises. DORA metrics can establish objective baselines across teams for software-development lifecycle velocity and stability.
Use DORA’s guidance to inform your interpretation. Verify current metric definitions and avoid universal performance cutoffs. Assess whether corrective work produces accepted outcomes. Record client-side dependencies that affected delivery. Keep delivery activity distinct from business acceptance.
Technical control
Use Technical control to inspect repository access and intellectual-property rights. Verify deployment reproducibility, security findings, and operational ownership.
Distinguish mere possession of credentials or documents from demonstrated ability. Ask the vendor to demonstrate that it can build, deploy, restore, and operate the software under authorized access. Test these claims in a staging environment and document the results.
Identify which control gaps have accountable remediation owners. Flag gaps that remain unresolved or unsupported by evidence. Distinguish contractual assertions from verified rights.
Working relationship
Use Working relationship to examine transparency and accountable vendor leadership. Review escalation records and the vendor’s willingness to expose delivery evidence.
Assess the client’s contribution to the failure. Look for scope churn, delayed decisions, or unclear acceptance conditions. Note any access delays or inconsistent product ownership on your side.
Specify what each party needs to change. Require observable commitments and track them against agreed timelines. Do not treat interpersonal confidence or blame as sufficient evidence.
Transition exposure
Use Transition exposure to examine service dependencies and documentation quality. Assess concentrated knowledge, contractual constraints, and the successor’s ability to assume responsibility. The Sourcing Playbook guidance calls for early planning for contract end.
Identify dependencies that make replacement difficult. Do not confuse those dependencies with proof that the incumbent can recover. Separate transition facts already demonstrated from assumptions requiring inspection.
Test successor readiness and the client’s ability to retain operational oversight. Document the exposure points in your shared evidence pack. Assess their likely impact before terminating the contract.
Turn the evidence into a rescue or replacement decision
Support rescue when the recovery plan can be tested
I would support a recovery attempt when the vendor accepts scrutiny and demonstrates progress against agreed acceptance conditions. Require a jointly accepted recovery plan. Assign accountable owners on both sides. Specify required client changes and an agreed review point.
For the recovery plan, track the four DORA measures: deployment frequency, lead time for changes, change failure rate and time to restore service. DORA defines change failure rate as the percentage of deployments causing a production failure, and time to restore service as the recovery time after a failure.
Evidence across Delivery reliability, Technical control, Working relationship and Transition exposure supports conditional rescue. Promises or a revised schedule alone do not establish recovery. Test the recovery plan at the agreed review point.
Support replacement when unresolved failures outweigh a credible transition
I would recommend replacement when persistent failures remain unresolved and a transition assessment supports a credible alternative. Three conditions point to replacement:
- Material failures persist despite corrective work.
- The incumbent cannot substantiate recovery.
- An assessed successor can plausibly assume responsibility.
Where evidence is incomplete, define a bounded assessment. Name evidence owners and set a review point. Use decision questions rather than forcing a verdict.
Compare the qualitative conditions across the framework’s criteria. Do not turn the table into an invented weighted score.
| Criterion | Evidence to inspect | Conditions supporting rescue | Conditions supporting replacement |
|---|---|---|---|
| Delivery reliability | Acceptance records, defects and releases | Revised commitments produce accepted work | Delivery failures persist despite corrective work |
| Technical control | Repository access, rights, deployment and security findings | Control gaps have demonstrated remediation | Control failures persist and an assessed transfer path exists |
| Working relationship | Escalations, transparency and client decision records | Both parties demonstrate agreed changes | Accountability failures remain unresolved after corrective work |
| Transition exposure | Dependencies, knowledge concentration, exit obligations and successor readiness | A credible recovery case withstands comparison with transition exposure | An assessed successor and documented continuity approach support transfer |
| Criterion | Evidence to inspect | Conditions supporting rescue | Conditions supporting replacement |
|---|---|---|---|
| Delivery reliability | Acceptance records, defects, releases | Revised commitments produce accepted work | Delivery failures persist despite corrections |
| Technical control | Repository access and security findings | Control gaps show demonstrated remediation | Persistent control failures; viable transfer |
| Working relationship | Escalations, transparency, decision records | Parties demonstrate agreed changes | Accountability failures remain unresolved |
| Transition exposure | Dependencies, knowledge concentration, exit obligations | Credible recovery outweighs transition exposure | Assessed successor supports documented transfer |
Compare recovery spending with the full cost of switching
Use the same planning horizon for both options
Compare remaining recovery expenditure with replacement expenditure over the same planning horizon. Include internal oversight, remediation, and continued operations. Account for the business effect of delayed usable delivery. For complex outsourcing projects, the UK government’s Sourcing Playbook requires a Should Cost Model Estimate to understand whole-life costs and guard against low-cost-bid bias.
Build the comparison as parallel option estimates:
- For recovery, show remaining vendor expenditure, corrective work, client oversight and delayed usable delivery.
- For replacement, show transition expenditure alongside the successor’s remaining delivery work and continued operations.
- For each entry, distinguish documented contractual costs from estimates and unresolved exposure.
Show which assumptions could change the decision. Test whether the preferred option changes when uncertain handover effort or delivery interruption increases. Treat historical spending as context, not a reason to retain the incumbent. Use qualitative comparisons where defensible monetary inputs are unavailable.
Include handover, knowledge transfer and contract exit
Next to the exit terms themselves, the switching estimate needs the cost of overlapping teams, successor onboarding and access changes, plus delivery interruption and continued internal oversight.
The UK government’s Sourcing Playbook provides sourcing and exit-planning principles. Its public-sector context does not make it a private-sector legal requirement.
Check exit obligations, assistance provisions and intellectual-property arrangements against the actual contract with appropriate legal input. Do not presume that a general playbook determines enforceable rights. Use documented contract terms and clearly labeled scenario assumptions instead of invented prices, durations or percentages.
Preserve operational continuity during a vendor transition
Prepare access, knowledge and service ownership
Map service dependencies before responsibility changes. Establish authorized access to repositories, infrastructure, credentials, deployment tooling and operational records.
Plan structured knowledge transfer using documentation, walkthroughs and practical demonstrations. Have the successor show that it can reproduce a build and deployment. Identify owners for incident response, security, unresolved defects and client decisions during overlap. Include responsibilities the client retains.
Transfer responsibility against observable readiness conditions
Sequence preparation, knowledge transfer, and readiness validation around service dependencies. Follow with responsibility transfer and monitored stabilization.
- Define acceptance conditions for the successor.
- Establish escalation arrangements, contingency actions, and rollback where technically feasible.
- Document residual risks rather than implying uninterrupted service.
Use the build and deployment demonstrations as readiness evidence, not merely completed handover activities. If the successor cannot demonstrate readiness, return the unresolved gaps to their owners before transferring the affected responsibility.
Manage access revocation and contract closure in step with verified responsibility transfer. Preserve incident coverage and the evidence needed for unresolved contractual matters.
Close the vendor review with evidence and a practical next step
Apply the same scrutiny to a prospective provider
For every outsourcing project, the Sourcing Playbook calls for a minimum-standard assessment of suppliers’ economic and financial standing. Apply rigorous financial due diligence to any new partner.
I would expect any successor, Peoplemore included, to be assessed against the same evidence requirements as the incumbent.
Bring a shared evidence pack to the vendor review
The Vendor Recovery Evidence Framework uses four criteria: Delivery reliability, Technical control, Working relationship and Transition exposure. Together, they support a decision between credible recovery and a credible transition.
The first step is to assemble a shared evidence pack containing recent delivery records, access ownership and contractual exit obligations.





