Create the comparison structure before proposals arrive
Build the evaluation table from the brief and publish the important criteria to participating agencies. Include mandatory requirements, weighted preferences and the evidence expected for each. Setting the structure early reduces the temptation to reward whichever proposal introduces the most appealing new idea or presentation format.
Ask suppliers to respond in a common order while leaving room for a different recommendation. A comparable response should cover diagnosis, approach, scope, named team, plan, evidence, costs, assumptions, risks and client responsibilities. Page limits can help focus priorities, but they should not prevent necessary detail from being supplied in appendices.
Put this into practice
- Criteria and weights agreed before review
- Mandatory conditions kept outside the weighted score
- Same requested response structure for every agency
- Evidence expected for each material claim
Check eligibility before evaluating preference
Review pass-or-fail conditions first. An attractive strategic idea should not compensate for an agency being unable to satisfy a genuine security, contracting, capability or delivery requirement. If a condition can be clarified, request the same opportunity for clarification from every supplier affected by it.
Be careful not to turn preferences into hidden eligibility rules. Familiarity, agency size or physical proximity may be useful context without being mandatory. Record the reason behind each condition so evaluators apply it consistently rather than relying on personal assumptions.
Put this into practice
- Required capabilities supported
- Contracting and compliance conditions addressed
- Timing and availability plausibly confirmed
- Conflicts or exclusions disclosed
- No preference treated as an unstated pass-or-fail rule
Compare the diagnosis and proposed approach
Look for a clear connection between the stated problem, available evidence and recommended work. A long activity list is not a strategy if the agency cannot explain priorities, dependencies and what would change its decisions. Reward useful trade-offs and a plan for testing assumptions rather than confident certainty.
Separate genuinely different thinking from different language. One agency may call a phase discovery while another splits similar work across audit, research and planning. Translate both into the decisions, activities and outputs they contain before deciding that one proposal is more complete.
Put this into practice
- Problem and audience understood
- Assumptions and evidence gaps identified
- Priorities and trade-offs explained
- Approach is feasible within stated constraints
- Measurement supports future decisions
Normalise scope, team and timetable
Put every proposed workstream into a common scope table. Record deliverables, frequency or volume, owners, approval rounds, dependencies and acceptance points. This exposes where one agency includes implementation, production or ongoing support that another leaves with your team or treats as optional.
Do the same for staffing and time. Compare responsibilities and expected involvement rather than job titles alone. Check whether the timetable includes client reviews, procurement, content, technology access and other dependencies, or whether it represents only the agency's production time.
Put this into practice
- Deliverables, volumes and review cycles
- Named roles and expected involvement
- Client work and approval turnaround
- External partners and hand-offs
- Milestones, dependencies and acceptance points
Compare total expected cost and delivery risk
Separate the base agency fee from media, production, platforms, data, travel, external specialists and likely optional work. Model the total under the same realistic scenarios. A lower quote may depend on more client work, fewer deliverables or a different service level; a higher quote may include work another proposal has omitted.
Record commercial assumptions beside operational risks. Consider account ownership, continuity, transition, rights, reliance on one specialist and the consequences of a delay. Risk should not become an unexplained penalty score; identify the evidence, likely impact and proposed treatment.
Put this into practice
- Same cost categories and tax treatment
- Expected internal effort considered
- Likely optional scenarios priced
- Change and cancellation rules visible
- Material risks linked to evidence and mitigation
Score independently, clarify gaps and record the decision
Have evaluators score the written response and pitch evidence independently before moderation. Require a brief reason or source for each score and use an explicit unknown option. Giving missing information a neutral score rewards gaps and makes due diligence harder to see.
Use clarification to resolve differences that could change the outcome, not to coach one agency into a better proposal. After moderation, record the evidence behind the recommendation, important residual risks and commitments that must enter the agreement. Keep an audit trail appropriate to the scale and governance of the appointment.
Put this into practice
- Independent first-round scores
- Evidence note for every material score
- Unknowns tracked separately
- Clarifications shared fairly where relevant
- Decision rationale and contract actions recorded
Common questions
Should the lowest-priced agency receive the highest value score?
Not automatically. Compare what the price includes, the client effort required, likely optional costs, service level and delivery risk. Value is the fit between a feasible approach and its complete expected cost, not simply the smallest headline number.
Can agencies use their own proposal format?
They can express their thinking in their own way, but require a common set of answers or a completed comparison schedule. This preserves room for a distinctive recommendation while ensuring scope, team, costs and assumptions can be compared.
How should missing information be scored?
Mark it as unknown and request clarification if the answer could change the decision. Do not quietly assign an average score. After clarification, record the evidence that supports any revised assessment.