The classic single-server queue problem. See why one person doing everything creates exponentially longer lines than you'd expect.
See It In ActionWatch the line grow even though the barista never stops working
One barista takes orders, makes drinks, and handles payment. Customers arrive every 5 minutes on average. Service takes 5 minutes on average. Should be balanced, right?
The line grows anyway. Why? Because "average" hides variability. Sometimes 3 customers arrive in 5 minutes. Sometimes a complex order takes 8 minutes. These mismatches accumulate.
Separate tasks. One person takes orders, another makes drinks. This parallelization dramatically reduces wait times even without adding total labor.
M/M/1 is queuing theory notation: Markovian arrivals (random), Markovian service times (random), and 1 server. It's the simplest queue model and surprisingly powerful for understanding real-world service operations.
When arrivals and service times are random, temporary mismatches occur constantly. Three customers might arrive in a row, then none for 10 minutes. The queue that builds during busy bursts doesn't clear instantly during lulls.
Task parallelization (separating order-taking from fulfillment), reducing service time variability (standardized processes), and adding a dedicated expediter can all reduce waits without adding total labor hours.
No signup required. Watch variability create congestion in real-time.
Try This Simulation Free