Choose criteria that can change the decision
Start with five or six material dimensions. A practical set is: problem understanding, proposed approach, relevant evidence, delivery team, commercial clarity and operating fit. Add a specialist criterion only when it is central to the brief.
Remove duplicate questions. “Strategic quality”, “quality of thinking” and “insight” may be the same factor described three times, accidentally giving it extra weight.
Define the scoring scale
A five-point scale can work when every point has a behavioural anchor. For example: 1 means material requirement not addressed; 3 means credible response with adequate evidence; 5 means highly relevant evidence, clear trade-offs and a feasible plan.
Include “unknown” separately. Giving an unanswered question a neutral three rewards missing information and hides due diligence gaps.
- 1 — material gap or unsupported response
- 2 — partial response with significant concerns
- 3 — credible and sufficient
- 4 — strong, relevant and well evidenced
- 5 — exceptional fit with clear trade-offs
- U — unknown; requires resolution
Set weights before the pitch
Agree weights while the team is focused on the brief, not after it has a favourite agency. Keep the total at 100 and make sure the most consequential dimensions carry the most influence.
Treat mandatory eligibility conditions outside the weighted score. An agency that cannot meet a true requirement should not progress because presentation scores compensated for it.
Score evidence, then record the reason
Require a short reason for each score: a proposal page, answer, example or behaviour seen during the pitch. This makes the group discussion faster and reduces the influence of the most confident speaker in the room.
Score the pitch and written response as one evidence set. Presentation craft can matter for some appointments, but it should not silently dominate the criteria.
Moderate differences rather than averaging them away
Have evaluators submit scores independently. Discuss large differences and unknowns, then let people revise their own scores if new evidence changes their view. Keep a record of the rationale supporting the recommendation.
Check whether a small change to the weights changes the preferred agency. If it does, the decision is close and needs a focused follow-up rather than more decimal places.
Copy-ready scorecard fields
Criterion · definition · weight · evidence expected · evaluator score · evidence note · unknown or follow-up · moderated score. Add pass/fail requirements, reference checks, commercial assumptions and final risks in separate sections.