This op/ed was contributed by Jonathan Marek, senior vice president of APT, the world's largest purely cloud-based predictive analytics software company. This article does not necessarily reflect the opinions of the editors and management of Nation’s Restaurant News.
Getting pricing strategy right may be the most challenging task for large restaurant organizations — wrong decisions can lead to costly trade-downs or lost traffic. Similarly, failing to take price where it’s available can leave millions of dollars of lost profit on the table.
The prevailing approach to pricing optimization in the industry is to use elasticity models, which recommend item-level prices by analyzing correlations between historical prices and sales. While this approach can generate interesting hypotheses, implementing recommendations from typical elasticity models without first validating them in the real world bears substantial risk. Based on our experience working with over 25 restaurant brands across four continents, we have identified the following five often overlooked facts about restaurant pricing.
Fact 1: You can’t understand the impact of changing prices without a test versus control analysis. If you haven’t analyzed price changes on a test versus control basis with a well-matched control group, it is impossible to understand the impact attributable to that price change. As you may have found, elasticity models are often inaccurate. Imagine that you dropped the price on gingerbread lattes in December: Sales of those items would skyrocket, and elasticity analysis would suggest that restaurants across the network should drop prices. However, in December, many of those units might sell anyway, so a price decrease would largely subsidize existing guest behavior. Seasonal biases aren’t the only issues at play, so year-over-year comps analysis won’t suffice here, either. Restaurants are impacted by competitors’ actions, economic trends, weather events and many other uncontrollable factors that don’t happen every year. In short, historical elasticity curves simply fail to reveal causal relationships that can inform future pricing decisions.
Fact 2: Cross-elasticity rules restaurant economics. Introducing a new value item might seem like a great idea if that item sells like crazy in an initial trial. If you test in enough restaurants, you might even be able to read a significant impact versus control, observing that the unit lift far exceeds margin rate compression, making the new item appear profitable overall. However, pricing analysis needs to account for inevitable mix-shift of other menu items. A new value item could drive incremental profits, but it could also cannibalize higher-margin items.
Fact 3: Price elasticity varies by store — a lot. Based on APT’s pricing work with dozens of restaurant concepts, we have observed that restaurant-by-restaurant price elasticity varies greatly based on restaurant characteristics (e.g. size), competitive density and demographic factors (e.g. median income and population density). Restaurants need to do the right analysis to identify which factors impact pricing for their concepts. Such analysis can often yield tens of millions of dollars in annual profits by correctly setting price tiers.
Fact 4: If you have a franchised concept, your franchisees are likely testing prices for you. Franchisees often make independent pricing decisions; as such, there is an opportunity to mine historical item-level data to create a test and control environment without actually designing a price test. However, the biased nature of these unplanned tests, or “natural experiments,” is an even greater case for scientifically constructed control groups, as opposed to balance-of-chain or balance-of-region approaches. For example, franchisees often increase price in their highest-traffic locations. Comparing these high-traffic locations to low-traffic locations will lead to inaccurate analysis.
Fact 5: Pricing analysis should not be outsourced. Why would you turn one of the most impactful levers of your business over to consultants with their own black box? Black box solutions may lead to the mistakes outlined above. Restaurants need to own their pricing analysis by doing it themselves with the right tools.
There are two ways restaurants can make more money: increase check size or increase the number of transactions. Changing prices is one of the most important levers restaurants can pull to impact both of these metrics. However, to correctly set menu prices, restaurants need to move beyond traditional elasticity analysis and determine how guests react to realized price changes through scientific test versus control analysis.
Jonathan Marek, senior vice president with APT, leads engagements with casual-dining and quick-service restaurants, as well as specialty retail, big box retail and banking clients. He has helped clients improve performance through better capital strategy, new concept development, emerging media strategy, media optimization, store labor planning and site selection.
Do you agree with Marek? Sound off in the comment section below.