You can commission a technically impressive piece of measurement and still leave the budget meeting with the same disagreement you had before it started. That tends to happen when the brief asks for better understanding without saying which decision the understanding is supposed to change.
Before choosing a model or asking for a dashboard, write down the choice facing the business. If the team can’t explain who will act, what money is at stake and when the answer is needed, the project probably needs a better brief before it needs a supplier.
Make the choice specific enough to answer
“Understand our marketing effectiveness” leaves an enormous amount open. “Decide whether to move £100,000 from paid social into non-brand search next quarter, while protecting contribution and cash payback” gives the work a commercial purpose. Putting the amount at stake in your brief forces useful questions about materiality and risk.
Include the option of holding the current allocation or making a smaller, reversible change. Otherwise, the analysis can become an argument between two preferred channels when the sensible recommendation is to test a limited shift and learn.
Define the economics before comparing the evidence
Finance and marketing need to agree what success means for this decision. Revenue may be the right operational outcome, while contribution and payback determine whether the investment is affordable. If teams use different treatments for discounts, cancellations, customer identity or variable costs, those differences can move the recommendation before any sophisticated analysis begins.
State how you’ll treat uncertainty too. A business with limited cash may reasonably prefer a smaller change with a faster review, even when another option has a higher expected return. The brief should expose that constraint so the analyst can recommend something the business can actually do.
Use this brief before selecting a method
Decision: What specific allocation or operating choice will the project inform?
Accountable owner: Who can approve the change, and who will implement it?
Investment at stake: What budget, contribution or commercial exposure could change?
Alternatives: Which actions, including maintaining the current approach, are genuinely available?
Deadline: When must the recommendation arrive to influence the decision?
Economics: Which revenue, contribution, acquisition and payback definitions will be used?
Available evidence: What data, experiments, commercial context and known measurement breaks exist?
Critical uncertainty: What could reverse the preferred action?
Deliverable: What recommendation, ranges, assumptions and implementation plan must be handed over?
Review: When will the outcome be assessed, and what would trigger a change of course?
This isn’t a demand for perfect information. Its purpose is to make missing evidence and disagreements visible early enough to handle them.
Let the question determine the measurement
Once the decision is clear, assess whether attribution, an experiment, MMM or a simpler economic analysis can answer it at proportionate cost. The IPA’s Making effectiveness work guidance discusses the roles of attribution, experiments and marketing mix modelling; choosing among them still requires judgement about your evidence and the decision.
If channels have always moved together, a model may struggle to separate their effects. If a test can’t run before the budget deadline, it won’t answer the immediate question, although it might improve the next one. A supplier should explain those limitations before you commit to the project, together with a useful narrower scope if the original ambition is unsupported.
Be equally careful about commissioning a method because you already own the software. Access to a tool doesn’t establish that your data can support the answer you want.
Agree what a useful handover contains
Ask for a recommendation that names the action, the evidence supporting it and the assumptions that could overturn it. Where several actions remain plausible, the deliverable should explain their tradeoffs and propose a proportionate next step rather than manufacture certainty.
For the £100,000 allocation example, a useful result might recommend a staged £20,000 test, with the remaining £80,000 held until a defined review. That recommendation would need its own evidence; the point is that an effectiveness project can earn its fee by making the decision smaller and more learnable, rather than endorsing the full move.
The handover also needs an owner, an implementation date and a record of what success would look like. Without those, even a sound recommendation can disappear into the presentation archive.
Check whether a project is worth buying
If the unresolved issue is an inconsistent margin definition, an internal working session with finance may be the best first step. If no available evidence could change the decision, ask why you’re buying analysis at all. External support becomes more useful when the investment is material, the evidence genuinely conflicts or the team needs independent challenge to reach an actionable conclusion.
At GV Marketing, we start by defining that decision and the evidence needed to support it. If a focused internal review will get you there, we’ll say so. Where a paid project is justified, the scope should make clear what decision the fee is buying and how you’ll judge whether the work helped.
Sources
Read next: marketing budget allocation and choosing measurement for your decision.