Decades after supermarkets introduced computerized cash registers and barcode scanners, the “point of sale” or POS system now lies at the heart of information technology used in retail stores, restaurants and hotels. Data recorded by the POS system has become a rich source of information about work carried out by cashiers, salespersons, restaurant servers and other service workers. Overall, work in retail and hospitality is increasingly mediated and managed through digital technology, from shelf stocking in stores to food preparation in kitchens to room cleaning in hotels.
This case study explores, examines and documents how software used by retailers, restaurants and hotels processes personal data on employees in order to track behavior, monitor performance, direct work and automate task allocation. First, it gives an overview of software systems in retail and hospitality, specifically addressing functionality for managing and monitoring workers and the subsequent effects this has on them, based on a brief review of field reports, survey-based studies and media articles, with a focus on Europe. Second, the case study investigates retail and hospitality systems offered by the enterprise software vendor Oracle, based on a detailed analysis of software documentation and other corporate sources. Oracle, a major vendor with a significant customer base in Europe, was selected as an illustrative example of wider practices. The examination documents a wide range of mechanisms that help employers to structure, direct, monitor and control work:
- Performance monitoring and behavioral control. Retail stores can use rich behavioral data on sales transactions recorded by the POS system to rate and rank cashiers and salespersons by their speed of work and sales performance, down to every instance of scanning an item with the barcode scanner. Similarly, restaurant servers can be rated and ranked by the number of table turnovers and guests served, average sales per guest, tips received and by how often they have returned late from breaks. Employers can rank workers “from best to worst”, identify “underperformers”, single out the “least profitable” employees and predict how productive they might be in the future. Assessing workers by their sales performance or even by the tips they receive from guests can be considered a form of quantifying affective work. This kind of monitoring can put workers under pressure. Fraud prevention systems for retail stores use POS data for continuous risk profiling, constantly assessing whether a cashier’s behavior may point to “employee fraud”, “policy violations” or “training issues” and singling out “high risk” cashiers, who are then put on a “watch list”.
- Automated task allocation and algorithmic management. One system for retail stores can automate task allocation by assigning pick lists to shelf stockers, who move store inventory across the backroom and shelves, while monitoring their every step based on handheld devices with barcode scanners. A housekeeping system for hotels can automate task allocation for room attendants, who receive instructions about which room to clean next via a mobile device, including a target time for each room. Workers see a timer that counts the minutes and seconds they have already spent on a room. Cleaning tasks are distributed based on rules, booking data and predefined “credits” that represent the time required to complete different kinds of tasks and other rules. Employers can view reports about the time spent cleaning rooms. Another system subjects kitchen workers to rigid micromanagement. In order to “optimize kitchen workflows” and “speed of service”, restaurants can determine target times for the preparation of each food item or component. The system then distributes guest orders and assigns them to workers across several kitchen stations, who see their preparation tasks on video monitors. Timers and red-blinking alerts notify employees when the specified preparation time for a food item has been exceeded and put workers under immediate pressure. A table management system for restaurants can display alerts that remind servers to collect the bill or stop by a table once again.
Several systems build on earlier practices, but by expanding the share of work activities that are subject to digital recording and automated direction, they significantly increase the potential for performance and behavior control. When fully implemented, these systems can be used to maximize productivity and keep costs low while leaving little room for agency and discretion at work. The recorded data can also serve as evidence for employee misconduct at a later point in time. When these systems fail to create an exhaustive and realistic digital representation of the work process, this may lead to deviations between actual work and digital records, to arbitrary decisions being made about employees, and to discomfort and stress.
According to Oracle, its technology processes $100 billion in retail transactions annually and is installed in 350,000 restaurants and 40,000 hotels, from small businesses to quick service chains, large resorts, theme parks and even cruise ships. Some of the systems examined in this case study are part of, or can be interlinked with, higher-level systems for store operation, inventory management, property management (PMS), enterprise resource planning (ERP), supply chain management (SCM) and business intelligence (BI). In part, Oracle’s performance monitoring systems for retail appear to be based on a standard for data processing defined by a retail industry association.
The findings of this case study will be incorporated in the main report of the ongoing project “Surveillance and Digital Control at Work” (2023-2024) led by Cracked Labs, which aims to explore how companies use personal data on workers in Europe. The main report will draw further conclusions.