Three different questions behind marketing return

Return on advertising spend can describe very different things. Before using a number to change a budget, ask what question it actually answers.

What did the platform record?

A platform’s reported return links advertising activity with conversions according to its own rules. That is useful for managing delivery within the platform. It is not automatically an estimate of what the advertising added to the business.

Several channels can take credit for the same outcome. Some demand may have existed without the measured advertising. Other activity may have influenced a sale without appearing in the final conversion path.

What did the activity add?

Incremental return asks how the outcome differs from a credible estimate of what would have happened without the activity. Experiments and carefully specified models can help answer this, with different assumptions and limitations.

Interpret the result with its uncertainty, time period and commercial context. It is evidence to evaluate, not a label that removes the need for judgement.

What might the next investment return?

The next dollar may not perform like the average dollar already spent. Reach, saturation, timing and competitive conditions change the opportunity.

Marginal return helps a team compare potential allocations while respecting minimum activity levels, brand commitments and operational capacity.

For the next budget discussion, ask for the definition of the return, the evidence supporting it and the assumptions behind the proposed change. A defensible measurement system makes those distinctions visible.

How should I use attribution and causal measurement together?

Attribution shows which interactions led to a sale. Causal measurement shows which activity created additional demand. Both are useful, but they answer different questions.

Use attribution to improve the details of a campaign: audiences, placements, keywords and creative. Use experiments and marketing mix modelling to decide how much to invest and where.

The numbers will not always agree. A platform can report a strong return even when much of the demand would have happened anyway. Upper-funnel activity can look weak in click reports while creating sales that appear later or through another channel.

• Start with the decision you need to make.
• Use causal measurement to set the investment direction.
• Use attribution to optimise within that direction.
• Investigate disagreements rather than averaging the numbers.

Attribution helps teams act quickly. Causal measurement helps them act on the right things.