Chaos in the kitchen. It’s a nightmarish experience for restaurant operators, ranging from incorrect orders, running out of ingredients, and understaffed kitchen personnel, which all leads to dissatisfied customers, exhausted staff members, poor decision-making, and shrinking margins. Chaos in data analytics can have similar consequences. Recent trends such as diverse sales channels (on-premises and off-premises) and customer feedback loops (field inputs, online reviews, surveys, etc.) create a multi-layered restaurant operating environment, turning data analytics into a necessity rather than an option. Similar to challenges around kitchen management, data challenges are multi-faceted, from data compilation and interpretation, to conversion and utilization, and cultural integration.
Data compilation revolves around migration and centralization. Restaurants operate in multi-platform IT environments where data typically resides in multiple places. This creates a need to incorporate a system that compiles data in one place — a central repository. A central repository refers to a collection of accumulated data from an organization’s existing databases over time. It’s the harmonization of records that creates a platform for activities, records, and transactions to be easily accessed, shared, and analyzed. It helps to detect current progress and identify room for future improvements and productivity. A well-managed central repository preserves data quality and integrity, as well as references to past events and records. A primary challenge restaurant operators face when utilizing a central repository is maintaining data integrity. At the onset of implementing a central repository, restaurant data analytics must analyze factors that may distort integrity such as uncommon and incomplete data population, and KPI definitions with embedded algorithms. Moreover, mere data accumulation over time may result in integrity degradation without proper management and caretaking efforts such as normalization, testing, and filtering.
Management needs to understand data interpretation and limitations on embedded algorithms when aggregating data. A simple example is guest count. One chain may capture a meal as one guest, while others may require servers to physically count guests. Chains may count non-meal orders (such as appetizers or drinks) as a portion (one-half or one-third) of a guest, while others count them as one guest if the orders were entered at the bar. There is no perfect algorithm. Instead, management should understand its limitations and embrace an open-minded attitude. Over time revisions may be required due to internal changes from operations, or external from customers or competition.
Restaurant management should view data conversion and utilization from a practical implementation standpoint as a feedback loop to operations. Information from data analytics should be simple, relevant, and accommodative to operations. Easily understandable metrics, such as weight-based food waste instead of dollar-based, represents simplicity. Time-lapse between data timestamp and presentation allowing for timely response represents relevancy. For example, daypart sales and labor hours should be monitored on a daily basis instead of weekly, while marketing campaigns should be measured over a longer period in order to allow trends to materialize. Lastly, accommodative pertains to ease of use from access and presentation standpoints. Tablets, cloud-based apps, and “push” systems such as e-mail or text allow GMs to remotely monitor operations. “Callout/exception alert” systems will highlight immediate focus areas. Visual cues (graphs or charts) are better at highlighting trends compared to tables. Line graphs are optimal for illustrating timelines, bar graphs for comparisons across regions or locations, and pie charts for relative percentage (such as daypart sales or channel breakdown).
Fostering a data-driven culture within field operations can be achieved through several methods. One method is basing GM bonuses on store profitability metrics. Such incentives should be straightforward without many exceptions. Too many exceptions create difficulties in evaluating true performance and present distrust in the data. Management should foster culture on “let’s look at the data first” instead of solving issues based on anecdotes and hypothetical terms. For example, one chain experienced high BOH workers’ comp incidents. Field operations stated that prior management mandated menu rotation every two weeks to foster a better guest experience. However, guest pattern analytics indicated that a majority of guests visited once a month, thus such frequent menu rotation made minimal, or no, improvement on guest experience, while creating chaos in the kitchen. By triangulating guest pattern data, online feedback, and operational inputs, new management decided to rotate the menu once monthly, and workers’ comp incidents declined by more than half.
Lastly, smaller or mid-size restaurants are resource-constrained, unable to aggregate and analyze data without an expensive IT system and FP&A team. In this case, potential solutions include automation and outsourcing. Management often misunderstands automation as advancing from Excel to business intelligence tools such as PowerBI and Tableau, which only partially help. Front-end data aggregation and automation efforts are often relegated to second priority. This is flawed as maintaining data integrity and developing AI-based data gathering and analytics can tremendously reduce human reporting hours and eliminate many errors and inefficiencies, allowing seamless FP&A operations with minimal personnel. However, developing such an architecture in-house is typically feasible only for large organizations. Outsourcing with the right partner can provide management with access to the latest technology solutions with the proper IT infrastructure and personnel.
Similar to an organized kitchen with streamlined production, restaurant management should take the same approach with data analytics. Organized data enables organizations to get a finger on the pulse of their business. Management should embrace data analytics with discipline and agility.
Bryan Shupe is the Founder and CEO of OverlayAnalytics, Inc. a software company that makes it easy for companies to connect their financial data to the teams and applications that need it. Prior to founding OverlayAnalytics, Bryan specialized in value creation for private equity owned businesses supporting turnaround and restructuring activities as a Vice President with Deloitte Corporate Restructure Group. Bryan holds a Chartered Professional Accountant (CPA) designation.
Sugi Hadiwijaya is a partner with CR3 Partners LLC, a turnaround management consulting firm. Sugi has consulted with restaurant companies on financial analytics, capital sourcing and operational improvement. He successfully utilized data analytics alongside field inputs to improve restaurant’s financial performance. He previously served as interim CFO for a casual dining restaurant concept. He holds certifications as a Chartered Financial Analyst (CFA), a Certified Insolvency Restructuring Advisor (CIRA), and a Certified Distressed Business Valuation (CDBV).