Search

Starbucks Leadership: ‘Everything Is Fine.’ Starbucks Customers: ‘Where’s My Coffee?’

Your algorithms will never be better, or more important, than customer perception.

Starbucks has faced criticism for long lines and long wait times, despite the company’s staffing algorithm being upgraded. This is a classic case of relying too much on technology, as it can lead to problems. While technology can be cheaper than employees, it can also lead to issues if not properly managed.

Algorithms and predictive models can be helpful in planning, but it is important to remember the programming phrase “garbage in; garbage out.” The past isn’t as important as the future, as computer models use past data to predict the future. However, humans can add flexibility by adding in flexibility where algorithms see numbers.

Staffordable algorithms can handle new things like promotions and specials, but Starbucks employees argue that their algorithms don’t account for these. While it’s doubtful that the Starbucks algorithm completely leaves out information about promotions, it doesn’t feel like it takes it into account.

Feelings matter, and how employees perceive staffing levels and wait times matters more than factual reality. For example, lines at Disney parks are often surrounded by interesting activities to change customers’ perceptions of the time they stand in line for a two-minute ride. If employees believe they are understaffed, then they are likely understaffed.

Customer perceptions are crucial, and companies should consider their employees’ thoughts and feelings when examining facts. Customers want to feel like a company is taking care of them and willing to pay a premium for it. If a company has an algorithm measuring the importance of popcorn machines, it has an impact.

When examining facts, it’s essential to speak with employees and determine what they think and feel. If unhappy customers are unhappy, telling them they are wrong won’t bring them back. Their perception is their reality, even if it’s false according to the data fed into the algorithm.

Read More @ Inc.

Suggested Reading