Dealership marketing measurement and incrementality

What to check before you decide

An incrementality test gives a cleaner answer. In a geographic test, you run a campaign in some markets and hold it back in comparable ones, then compare sales.

It needs enough volume, a clean split and a long enough window, and a single store may not have the volume to read it. Tests can also pause a channel for a few weeks, or hold out part of a customer list from an email or ad campaign, and compare results.

Marketing mix modeling uses historical spend and sales to estimate each channel's contribution. It needs a long record, often a year or more of weekly data, and consistent definitions.

It is a guide to budgets, not a verdict on any one ad. Models work best when checked against experiments.

Ask a vendor what data it needs, how it validates the model, how it handles seasonality, manufacturer incentives and inventory shortages, and what it will show you when it is uncertain.

Keep results practical. Use them to find waste and to decide which channel to test next, and revisit each quarter. A measurement program that never changes a budget is not worth the fee.

Be wary of precision, since a model that reports each channel's return to the second decimal place is claiming more certainty than a single dealership's sales history can support, and a vendor who shows confidence ranges, explains what the data cannot tell it and recommends a specific test to settle the biggest uncertainty is more trustworthy than one who promises a definitive ranking of channels.