Left-Digit Bias at Lyft.

This article investigates the presence and implications of left-digit bias—a behavioral pricing anomaly where consumers disproportionately react to changes in the left-most digit of prices—in the pricing strategy of Lyft, a major U.S. rideshare company. Using observational data from over 600 million Lyft sessions and a large-scale natural field experiment involving 21 million users, the study finds significant discontinuities in demand at round dollar prices, with passengers perceiving prices just below a dollar threshold (e.g., $13.99) as substantially cheaper than those just above (e.g., $14.00). The estimated inattention parameter, reflecting this bias, is approximately 0.5, indicating that a one-cent price drop below a dollar value is perceived as roughly a 50-cent reduction. The field experiment confirms that implementing a 99-cent pricing strategy increases Lyft’s profits and customer usage, with potential annual profit gains estimated around $160 million if such pricing were broadly adopted. The study also discusses why Lyft’s prior machine-learning pricing algorithms did not exploit this bias and situates its findings within the broader behavioral economics literature.