The next report should answer what happened to the previous report’s recommendation.
Find the agreement, not just the slide
A recommendation may finish a presentation enthusiastically and disappear from subsequent work. Recover the actual decision: which problem to explore, what action was approved and who would follow up. A discussed idea is not a confirmed assignment. This avoids evaluating something as a failed experiment when it never happened.
Recover a brief commitment description. If it is only a broad phrase, acknowledge an original limitation at the next review. Do not retrospectively rewrite the objective to fit available data. An incomplete definition can itself be organisational learning even if it prevents concluding whether the proposal worked.
Check what was actually executed
Before opening the results chart, inspect materials and dates. Content may have appeared later, in another format or with changed messaging. It may remain pending for commercial reasons. Expose these changes: they determine which activity can be analysed and how much of the original agreement it retains.
Do not build a success or failure story from a campaign name. If the team tested something else, describe that. Review is not about punishing deviations but understanding learning from actual work. A clear explanation supports deciding whether the unanswered original question is still worth pursuing.
Recover the question before selecting indicators
A test of feature clarity may require reviewing people’s questions. A format exploration may require looking at content and observed responses. Data should follow the question, not the reverse. AMEC’s framework connects communication planning and evaluation, helping prevent a close from becoming an opportunistic search for favourable figures.
Select pertinent evidence and explain why it matters. Including everything available does not prove work quality. If an expected indicator cannot be obtained, record that and discuss the remaining possible interpretation. Quietly substituting another changes the experiment’s meaning and may confuse the assignment’s approver.
Review available Wiwo.MetriQ material
Wiwo.MetriQ supports period comparisons, publication and result review, and dashboards or exports in configured areas. It can gather work evidence with brand, project and period references. Confirm enabled sources and historical scope, particularly when different owners supply material.
Clearly distinguish agreement, observed execution and findings. Contextualise captures so another person understands their role. The platform organises available analysis; do not assume automatic verification of everything happening outside its sources or proof of causal relationships.
Explain what else may have changed
Product availability, offer, audience or distribution may change between reports. Improvement coinciding with an experiment does not eliminate those possibilities. Describe known differences instead of treating a simple comparison as definitive proof. Explained uncertainty is more useful than certainty nobody can reconstruct.
Evidence may justify repeating, adjusting or abandoning an idea without establishing a universal market rule. An experiment may also be insufficient. Explain the observation supporting the recommendation and which new information could change it. Reviewability keeps learning open.
Leave a decision that will not disappear again
End with a concrete next action, an owner and a review condition. The report need not become an endless operations list; its main commitment simply needs to be findable in the next conversation. If the decision is to stop, preserve the reason and which question is resolved or remains open.
Useful history is not accumulated presentations but a traceable relationship between a question, action and learning. MetriQ can supply evidence from enabled areas to accompany that memory. People retain responsibility for choosing experiments and explaining why the next decision is justified.
Frequently asked questions
Can an unexecuted experiment be called a marketing failure?
Not as an evaluation of its result. Distinguish an unapproved idea, pending action and changed execution. Each may need coordination, but none establishes how the audience would have responded.
Does comparing two periods prove the experiment caused change?
Not alone. Other conditions may change. Review context and use comparison to discuss decisions, without automatically turning it into a causal claim.
