Marketing: Common Questions About Measuring Campaign Results

Start with the intended outcome and the decision you need to support. A fictional enquiry campaign may need a measure of suitable enquiries, not just page views. Define the period, source and relevant event. Keep supporting measures available so that you can investigate the route to the outcome without confusing activity with commercial success.

No. Impressions count displays of an advertisement or content in the relevant system; one person may see it more than once. Check the platform’s definitions for the report you use. Do not describe an impression total as the number of unique readers, or assume that a display means the person understood or acted on the message.

For the usual advertising definition, divide clicks by impressions and multiply by 100. In a fictional practice example, 20 clicks from 1,000 impressions gives 2%. Use the same reporting scope and period for both figures. A zero denominator cannot produce a meaningful rate, and a higher rate alone does not establish better quality enquiries.

No. In Google Analytics, a key event marks an event important to the business. Advertising conversion setup has its own role and configuration. Check how your Analytics and advertising accounts are connected before comparing reports. A label on one screen does not establish that every event is counted identically in another system.

No. Attribution assigns credit according to the system and model used. It can help interpret observed interactions, but it does not by itself establish a causal effect. A fictional enquiry may have several influences that the report cannot observe. Explain the model and limitations instead of treating assigned credit as a controlled experiment.

Check tracking, dates, denominators, spend, outcome quality and comparison conditions. Investigate whether a change in setup explains a movement in the figures. Compare results with the agreed objective and describe uncertainty. A small campaign can produce useful learning without proving a lasting trend, so report what the evidence supports rather than selecting the most flattering metric.