In a conversation with my financial advisor the other night I asked him why he thought brands advertise. His response was “to get people to try the product”. He couldn’t be more correct. The role of advertising is to generate trial, break penetration into households that have not previously purchased the brand.  We should measure our success based on how well we accomplish that, but in most cases we are not.

Over the last 20 years or so marketers have become enamored with marketing mix modeling in an effort to better understand the contribution of every effort against total sales. The modelers will tell you precisely what percent of sales for the year each effort contributed. It sounds like a good concept but in practice it doesn’t really work. Every market mix modeling exercise I have ever seen analyzes advertising impact against total sales. If advertising’s role is to generate trial for new products and penetration for existing ones why are we measuring its success vs. total sales? We end up undervaluing the impact of advertising and overvalue the impact of in-store promotions.

The modelers are counting every transaction as it crosses the scanner at retail. The model cannot distinguish between the first sale, the second, third or fourth by an individual buyer.  If a product is purchased four times in a year by an ad driven new buyer the model is understating the ads impact by 75%. The ad is only responsible for the first sale. The product performance/satisfaction is responsible for sales 2, 3 and 4. Why is this an important issue? Because sales promotion actually hurts the effect of advertising. Huh? Here’s how it works. Ads get people to consider buying without price incentives. If an ad driven potential new buyer is in the store they are going to take advantage of the price promotion that day. They will not ask to pay full price even though they were willing to before they got in the store. If it’s a repeat buyer whose first purchase was driven by an ad they may even pantry load, taking them out of the market for at least two cycles. The worst thing is since they bought on deal the model gives the in-store promotion the credit for making the sale—even the first one.

We now have the ability, with loyalty card data, to separate first time purchases from repeat ones. We can track not only total sales but the impact ads have on generating new customers. We can better understand individual users repeat rates. We can also test pricing strategies to determine if the in-store pricing is harming profitability/equity. If advertising’s role is to drive trial/penetration our single-minded focus should be on that. Ask yourself: “What am I doing today to bring in new customers?”