Local Marketing Incrementality: Did the Campaign Create More Demand?

The business question

A campaign can receive credit for calls or bookings that would have occurred without the campaign. Incrementality asks a harder question: did the activity create additional outcomes, or did it merely capture demand that already existed?

How to measure it

Define the outcome before launch, then compare a treated group with a credible comparison group that did not receive the same intervention. Where possible, randomise eligible locations, audiences, or time blocks. If randomisation is not practical, document why the comparison is credible and what differences remain.

Example in practice

A multi-branch service business adds a local paid campaign in selected areas. Comparing only this month's leads with last month's leads would confuse campaign impact with seasonality, staffing, weather, and changing demand. A matched set of similar areas can provide a more informative comparison, provided the team checks whether their pre-campaign trends were reasonably similar.

Common interpretation errors

The biggest mistake is to call every tracked conversion incremental. UTM parameters, call tracking, and platform attribution can show that an interaction occurred; they do not independently prove that the interaction would not otherwise have happened. Another risk is choosing a comparison group after seeing the result.

Turning the result into a decision

Use the estimated lift together with uncertainty, cost, and operational capacity. A positive difference that is tiny or highly uncertain may justify another test rather than a broad rollout. A negative result can also be informative if tracking was sound and the treatment was delivered as planned.

Records and governance

Keep the launch date, eligible population, assignment rule, exposure status, outcome definition, exclusions, campaign spend, and any operational changes. Preserve the pre-test plan so that later edits do not silently change what success means.

First steps for a small team

Start with one measurable outcome and a small, defensible test. Write down the expected mechanism: for example, whether additional visibility should increase qualified enquiries, completed appointments, or profitable sales. Select the outcome that best matches that mechanism instead of choosing the metric most likely to rise.

What the metric cannot prove

Local businesses often have low volumes and spillover between nearby areas. People may see an advert in one area and travel to another branch, or a campaign may influence branded searches outside the treated group. Note these risks, use sensitivity checks where possible, and describe the result as directional when the design cannot support a strong causal conclusion.

A focused audit for local marketing incrementality

Start by writing the decision that distinguishing campaign-caused outcomes from demand that would have happened anyway is meant to support. Identify the source records that directly represent that decision and separate them from contextual indicators. For example, a click, a status label, a booked amount, and a completed financial outcome can each be useful, but they answer different questions. Record the definition, date range, filters, exclusions, and known gaps before calculating a result. Then trace a few cases from the original system into the report. If the transformation cannot be explained, fix the measurement or label the output as provisional. Keep a dated note of the action taken so the next review can distinguish a new result from a changed definition.

Questions for the review meeting

• Which part of distinguishing campaign-caused outcomes from demand that would have happened anyway is directly observable, and which part depends on an assumption?
• What operational change could produce the same pattern without the marketing intervention being responsible?
• Which missing or delayed records would materially change the interpretation of local marketing incrementality?
• What is the smallest next action that could improve the evidence or the business outcome?

How to make the comparison fair

When reviewing local marketing incrementality, compare records that had a similar opportunity to produce the outcome. A recent lead may still be waiting for a quote, appointment, or payment, while an older lead has completed the process. A branch may also have different opening hours, staffing, or service availability from another branch. Note these differences before interpreting a gap as a marketing effect. If a perfectly matched comparison is unavailable, describe the mismatch and reduce the strength of the conclusion rather than hiding the limitation.

Keep the result useful next month

Save the calculation alongside a short decision note: what was observed, what remains uncertain, and what the team will do next. For distinguishing campaign-caused outcomes from demand that would have happened anyway, also preserve the definition version and the date on which the underlying data were considered complete. That record makes the next review more informative because the team can see whether the result changed, the measurement changed, or both. If the business changes its workflow, update the metric specification deliberately and mark the break in the historical series. This small discipline prevents a dashboard from quietly changing meaning over time.

Conclusion

For local marketing incrementality, a defensible result depends on consistent definitions, traceable records, and a clear boundary between observed facts and interpretation. Use the measurement to answer the specific business question described above; do not extend it into a broader performance claim that the data cannot support. Review the result again when the relevant customer or financial outcomes have matured.