Blog

Analytics ROI vs. the high cost of ignorance

It's best if restaurateurs approach analytics by first asking themselves what their highest value questions are and then consider the many ways those questions might answered.

October 21, 2016

By Mike Lukianoff/Chief Analytics Officer - Fishbowl

I'm often asked what return on investment analytics offer. The question always gives me pause and invariably the person asking thinks I'm hesitating out of uncertainty, but that's far from the truth. After almost 20 years helping restaurants make better decisions with data, the question is a bit like asking what the ROI is on learning to read or what the payback period is on a thermometer.  How does one begin to answer such a question? 

For people who value making decisions based on all knowable facts, I think analytics can be equivalent to literacy. It's a requirement for obtaining knowledge, but does not guarantee if or how that knowledge will be used. 
 
It's best then, if restaurateurs approach analytics by first asking themselves what their highest value questions are and then consider the many ways those questions might answered. Better still, restaurateurs should first consider what would be the most costly mistakes they might make in managing their businesses. Then consider what answers and information is needed to best avoid those pitfalls.

Many clients start their forays into analytics by instituting price optimization programs. They do this for three primary reasons:

It's expensive to incorrectly price products: An error here can be among the most expensive ones you can make, while repairing such missteps can be among the biggest challenges you can face. 
The impact is predictable:Decades of empirical evidence has repeatedly proven to me that econometric-based pricing can produce measurable returns. Repeated controlled tests show this often comes in the form of either increased profits or retained traffic counts, resulting from informed pricing. 
It's a great starting point: Establishing comprehensive econometric pricing analytics requires collection of all the data needed to fully understand which forces best contribute to actual consumer purchase behavior. So price optimization is a great way to self-fund development of a total multi-purpose analytics program. 

I have repeatedly found that econometric-based pricing most always reigns supreme over other methods brands traditionally use for pricing. A comprehensive econometric pricing analysis typically has an initial benefit of one to two points in sales flowing through to profit in the first year. So a restaurant chain doing $100 million in annual sales could safely estimate $1-2 million in profit benefit after switching from a do-it-yourself pricing approach to one based on econometrics. 

Benefits beyond price optimization

A good analytics program can also have numerous other potentially high impact initiatives. For instance, a solid ongoing analytics program often provides some of its greatest rewards as a result of what it guides you away from implementing. 

As an example, I remember recent talking to a regional restaurant chain CFO who sought my opinion on a beverage company's recommendation that his chain reduce its three beverage size offerings to two, with more ounces per cup.  The beverage company's research showed this would produce huge numbers of happy customers who would more than make up for the lower margin per cup.  

The initiative was teed up to go system-wide, but the CFO remained skeptical.  After some analysis, I quickly gave him compelling information to put the cup-change initiative on hold until a six-week market test using the proposed cup set could be performed and analyzed. 

The results of that test proved customers failed to respond as predicted, but instead total beverage purchases dropped which compounded the problem of margin reduction. Indeed, if they had fully implemented the program system-wide, they would have taken a multi-million dollar hit on profit equal to 22 percent of the chain's annual net profit. In this way, this executive's simple, well-thought-out question turned out to be the chain's most profitable analytics deployment of the year. 

A well-designed analytics program must be continuous and progressive because insights build upon each other, meaning the highest value questions next year will likely be far different than those asked last year. In essence, this means that the true key to success is a business culture that puts facts ahead of instincts and feeds that culture with factual information that drives the best decision-making every day.