The company has the badge. The proposal has the references. Now ask who will actually arrive on Monday.
Assess a DPG implementation partner through evidence from the proposed team: a working deployment, decisions they can explain, comparable references and a credible plan for local operations. Accreditation is useful evidence within its stated scope. It doesn't answer every question about your programme, staffing or delivery conditions.
A disclosure before the example: I am Commercial Partnerships Lead at OpenCRVS, as shown on its Implementation Partner Programme page. I use that programme here because it makes the assessment areas concrete. My involvement should be clear when you read this. Other DPGs have their own arrangements, or none.
Find out what the accreditation covers
OpenCRVS describes business analysis, technical installation and country deployment, project management and operational support in its partner programme. That is more informative than an unexplained logo. It gives a buyer specific areas to investigate.
Ask the relevant programme owner how to verify the bidder's current status, what was assessed, when, and which people participated. Check whether those people are in the proposed team and what happens if they leave. A qualification held by a person who is no longer available deserves a different discussion from one held by the engineer leading your deployment.
Keep accreditation evidence separate from contract commitments. A programme listing is not your staffing plan, service-level agreement or acceptance certificate. Nor should an unfamiliar programme automatically disqualify a capable local team unless your procurement rules require a particular status.
This is a buyer's evidence method, not a ranking of suppliers. Set the criteria before the interviews and have your procurement team check that the process is permitted and fair.
Interview the people in the delivery plan
Ask for a named roster covering business analysis, deployment, integration, testing, delivery management and operations. One person may cover several roles in a small team. That is worth discussing openly rather than hiding behind a large organisational chart.
For each critical role, record allocation, start date, working location, language capability and the person who can cover an absence. Check which roles are employees, subcontractors or still to be recruited. Ask the bidder to explain how competing commitments affect the schedule.
Evaluate the team being offered, not the firm's best-ever team. A principal engineer's excellent interview is limited evidence if they are only allocated two days at mobilisation.
Give the business analyst an ambiguous requirement from your brief. Ask what they would clarify, who can approve it and how they would test their interpretation. Give the delivery lead a dependency arriving late. Listen for a concrete decision and a revised plan, rather than a promise to escalate everything.

Watch a deployment and one failure
Use a proportionate, agreed demonstration with synthetic data. Do not ask bidders to expose another client's records or credentials. For a substantial exercise, decide how preparation time will be recognised and give all bidders equivalent conditions.
- Start from the release. Ask the proposed engineer to identify the software version, local configuration, changed code and deployment instructions.
- Follow a transaction. Watch a representative service task and an exception. Ask the analyst to explain why the workflow matches the brief.
- Inspect a boundary. Introduce a permitted test failure, such as an unavailable integration. Check how the team detects it and avoids losing or duplicating work.
- Recover. Ask for a restore demonstration or recent, explainable restore-test evidence appropriate to the exercise.
- Hand it over. Have another team member locate the runbook and explain the next operational step.
Agree the scope first. A limited sales demonstration cannot prove national-scale resilience. Record what you observed, what the supplier merely described and what remains untested. Those are three different evidence states.
A useful follow-up is “show us a change you chose not to make”. The explanation can reveal whether the team understands product boundaries, upgrade costs and the service owner's authority.
Check local capability and references
For an international programme, “we have a local partner” needs names and responsibilities. Who can work in the required languages? Who can visit an office? Who will train supervisors? Who has authority to resolve a service-process question? Distinguish local delivery capability from the existence of a local registered company.
Request references whose constraints resemble yours: service domain, connectivity, migration burden, government approvals or rollout model. A large user count alone doesn't establish comparability.
With permission, speak to the service owner and someone who operated the system after handover. Ask what the supplier actually delivered, which proposed people did it, what changed from the plan and how unresolved defects were handled. Ask what the buyer itself had to provide. A reference that admits a difficult dependency can be more useful than an unqualified endorsement.
For UK- or UAE-based commissioners, the country of the buying organisation does not establish the delivery team's suitability elsewhere. Verify the receiving institution's requirements and any applicable donor or procurement conditions separately.
Score evidence without hiding gaps
The downloadable scorecard uses a suggested 0–3 scale: 0 for no usable evidence; 1 for an explanation or incomplete artefact; 2 for relevant evidence the team can explain; 3 for directly observed or independently checked evidence that meets the agreed criterion. These are decision aids, not an industry certification scheme.
| Criterion | Useful evidence | Unresolved risk |
|---|---|---|
| Team availability | Named roles and confirmed allocation | Key person still to be recruited |
| Deployment | Proposed engineer explains and runs the deployment | Only the sales specialist can demonstrate it |
| Local delivery | Named local responsibilities and workable support coverage | Local partner's scope is undefined |
| Continuity | Second person uses the maintained runbook | Recovery depends on one individual |
Let assessors score independently, then reconcile differences against the recorded evidence. Agree any weights and mandatory criteria before seeing the bids. A high total cannot cure a failed essential requirement. Keep that failure, its consequence and the permitted remedy visible.
Consider two illustrative bidders. One has polished references but cannot confirm its deployment lead. The other has a smaller portfolio and demonstrates recovery with the team it proposes. That doesn't make the second bidder an automatic winner. It tells you which evidence is stronger and which conditions would need resolving before appointment.
Record the decision and the conditions
Download the DPG partner evidence and interview scorecard (CSV). Make a copy per bidder. Each row has an interview prompt, requested evidence, observed evidence, score, assessor and unresolved-risk field. Scores are entered manually; there are no formulas or macros.
Finish with a short decision record: the evidence supporting appointment, remaining gaps, who must resolve them and the deadline. Recheck material staffing substitutions rather than assuming the original assessment transfers unchanged.
If you need help assessing delivery capability or programme risk, see my open-source and DPI work, or bring the proposed delivery model for a discussion. Disclose relevant relationships in the assessment team as well as in the bids.
Frequently asked questions
Can a newly formed team demonstrate enough capability?
Potentially. Separate the individuals' relevant experience from the firm's delivery history, and test how the proposed team works together. Record the uncertainty and any permitted conditions rather than treating either a new company or a familiar brand as conclusive.
What if a bidder cannot share client material?
Accept appropriate alternatives such as redacted artefacts, a synthetic demonstration or a reference conversation arranged with permission. Don't ask for confidential records. Record which claims remain unverified so the comparison stays honest.
Should price be included in this scorecard?
Keep the capability evidence clear, then use the commercial evaluation method agreed for your procurement. The worksheet does not calculate a combined price-quality ranking. Adding an improvised price weight after seeing the bids would change the assessment.
What if the winning bidder changes the team?
Review the affected evidence and replacement's suitability using the agreed process. A minor substitution may be manageable; losing the only person who demonstrated deployment or recovery changes a material assumption. Record the decision before relying on the replacement.

