Skip navigation
The interaction between inventory and advertising impact can make an analysis of return on investment more complex.

How Car Dealerships Can Precisely Measure Marketing ROI

The cheapest audiences to reach and motivate to visit can be the least likely to convert.

Chad Morber, the used-car manager at a Ford dealership in suburban St. Louis, inspired this column.

Evaluating his advertising budget, he was trying to figure out whether to drop one of his media channels from the mix. Car dealers face a complicated marketing challenge because their product mix changes daily and their prospective customers are primarily product-driven shoppers.

If you sell one product or service, then marketing ROI testing is straightforward. You systematically vary the dollars spent in each channel and correlate those changes with your revenue.

However, if you have a black ’17 Honda Civic on Tuesday and it’s gone on Wednesday, whatever proportion of the audience that is looking for a black ’17 Civic becomes immune to your advertising. This interaction between inventory and advertising return makes the ROI analysis more complex.

It is further complicated by the choice of conversion metric. If you have a clean way of measuring advertising response – such as a beam-break counter on your front door, a phone system that accurately counts and codes inbound calls and a CRM system that accurately measures inbound emails – then you can measure advertising’s ability to generate interest. Yet, this still does not give a clear measure of return on investment.

When I worked for CarMax, we had the infrastructure to measure interest. We could tell which advertising put more people through the door and generated more calls, emails and web visits.

This shaped the media strategy. We were good at converting ad dollars to foot traffic. But we also discovered different ad channels produced different types of customers with different probabilities of converting to sales. 

The cheapest audiences to reach and motivate to visit were the least likely to convert. This created some conflict between marketing and store management. 

Marketing said, “We’re putting more people in the store, converting them is your job.” Store managers said, “These people are just wasting our time, they can’t get financed.” 

This is the pitfall of using traffic rather than revenue as your success metric.

To go a step further, using monthly revenue as the success metric introduces other variables into the analysis.

Car dealerships often manage their inventory with two levers, spiffs and price. These levers can increase revenue but reduce net profit. Since net profit is the ultimate goal, that should be the unit of analysis when measuring return on marketing investment. 

However, depending on how a dealership does its cost accounting, net profit may be influenced by irregular expenses such as an IT charge or a change in utility rates. Changes in personnel and training can directly impact sales. Anything that impacts the net profit should be included in the analysis. (Wards Industry Voices contributor Philip Moore, left)

In mathematical terms, net profit is a function of marketing expense, inventory mix, other operating costs and inputs that differ from one month to the next. 

With a properly constructed data set, this function can be evaluated using linear regression that will reveal the dollar impact of each variable on net profit. 

The challenge is getting enough observations or rows of data for the regression to find the relationships. The statistical rule of thumb is you need at least 30 observations and at least one more observation than input variables. Thirty months is a long time in the car business.

One alternative is to get input from multiple locations. Many dealerships participate in cooperatives, often called Dealer 20 Groups. 

If 20 dealerships (typically from non-competing markets) are willing to measure, record and share this data, then you could get a clear picture of how specific marketing channels contribute to monthly net profit in six months instead of 30. Chad Morber’s store is part of a dealer group that has several locations around St. Louis, so he could also use data from his sister stores to conduct the analysis.

When the modeling is complete, you will know that every dollar spent with, say, Carfax produces X dollars of net profit, every dollar spent on CarGurus produces Y dollars of net profit and every dollar spent on that sales trainer in February produced Z dollars of net profit. 

Imagine having that insight when a rep shows up to inform you of a price increase.

Philip Moore is a senior research consultant at MDC Research.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.