← Back to feed

McDonald’s is running a machine-learning pricing engine across its nearly 14,000 company-operated and franchised US restaurants that goes beyond traditional cost-plus margin math. The system analyzes millions of point-of-sale transactions, evaluates publicly listed prices at nearby competitors including Burger King and Wendy’s, and estimates what customers in each specific location are willing to pay before they walk. It outputs what the company calls an “optimal price” recommendation for individual menu items at individual restaurants.

The distinction matters. Traditional location-based pricing starts with costs — rent, wages, local taxes — and adds a margin. The McDonald’s system also asks the harder question: what is the ceiling this particular market will tolerate? Internal dashboards reviewed by Reuters label each restaurant with a sensitivity tier. One screen designated a location as having “MEDIUM SENSITIVITY to Price” and cited “customer willingness to pay in your area” as a contributing factor.

The Fresno Data Point

Reuters identified two company-operated McDonald’s in Fresno, California, just two miles apart, selling the same Big Mac for $5.69 and $6.89 respectively — a 21% spread within a single city. The reporting couldn’t confirm the AI engine produced that specific gap, but franchisees told Reuters that the pricing system has widened some existing price differences between nearby locations.

The widening matters for franchisee economics. Operators in low-sensitivity (high willingness-to-pay) zones can capture more margin per transaction; those in high-sensitivity, price-conscious neighborhoods are constrained. The AI system shifts the decision from franchisee judgment — which has always varied — to a centrally optimized output that McDonald’s can calibrate and update systematically.

The Connecticut Ceiling

The clearest illustration of where algorithmic pricing can go came in 2023, before the current ML system was fully deployed, when pricing tools recommended a Connecticut franchisee charge roughly $18 for a Big Mac meal at one location. That figure circulated on social media and became a symbol of fast-food inflation. McDonald’s subsequently acknowledged it had pushed prices too far and committed to more affordable value options.

The current system is the institutional response: data-driven pricing that tries to find the optimum rather than simply maximizing at the ceiling. Whether optimizing against willingness to pay constitutes fairness or price discrimination is a question Seattle and at least two states — Maryland and New Jersey — have already answered with legislation restricting AI-driven pricing for grocery items. Fast food does not currently fall under those laws.

Scale and Transparency

The challenge McDonald’s faces is that smarter pricing and perceived fairness sit in tension. Customers who pay $6.89 two miles from someone paying $5.69 for the same sandwich are unlikely to view that as optimization. The company’s public position is that prices reflect local market conditions — a defense that has historically worked for franchised chains. The addition of a machine-learning layer that explicitly models willingness to pay shifts the optics, even if the underlying price differentiation has always existed.

The system covers individual menu items, not just meal bundles, and runs across all US store formats. No comparable ML pricing deployment at this scale has been publicly documented for a US quick-service restaurant chain.