Just now
Recently, a group of Uber and Lyft drivers noticed something curious: fares offered for similar trips weren’t identical. They did indeed experience slight variations in fares, but they drew the wrong conclusion about the reason why.
There are several reasons why fares for similar trips might vary slightly when drivers see them around the same time / side-by-side, which we’ve written about in the past and may include:
- GPS Discrepancies: Fares are calculated using estimated times of arrival based on GPS location, which doesn’t always perfectly match a driver’s physical location.
- Surge Pricing: If a driver has recently traveled through an area with higher Surge, their fares will reflect that higher rate, not the Surge amount at the trip’s starting point.
- Dynamic Updates: Surge heatmaps refresh continuously at regular intervals, so drivers may see different adjustments even seconds apart.
- Feature Testing: We regularly test new features and models to improve the app, which can result in temporary fare variations across groups. These tests are never based on protected characteristics.
- Promotions: Additional earnings like Boost+ incentives vary by city, time, and week, and are displayed directly on the offer card.
In the video from September, the X-factor was in fact the first bullet: GPS discrepancies.
This underlying issue is not specific to Uber. GPS accuracy continues to improve, but it’s not always perfect. The US Government’s official website on GPS data says that the margin of error for GPS is 16 feet, which means that two phones physically positioned next to each other could potentially show GPS points 32 feet apart. And the margin of error actually gets worse indoors, or near buildings, bridges, and trees.
The Role of GPS in Fare Calculation
The fares offered to drivers before they accept a trip depend on multiple factors, like estimated time and distance to the pickup location. But here’s the key: our app uses the driver’s location, as determined by GPS, to calculate these estimates.
In the case of the drivers in the video, Uber’s technology predicted that the drivers’ cars were parked at different spots on nearby streets. Even small differences in starting points can affect the time and distance estimates (especially if there are traffic lights), leading to different fares.
Visualizing the Variations
Here are some examples from the video:
We were able to use the information displayed in the video (such as pick up and drop off locations, fares and approximate time frame) to find the trips offered to the drivers in the video and analyze what was happening.
The green dots mark the locations captured by the driver devices’ GPS readings within the apartment building. Uber’s technology then estimated the on-street locations of the drivers’ vehicles, which is shown with the yellow lines. The slight variations of estimated on-street locations led to Uber’s technology using slightly different starting points. (You’ll notice the GPS is accurate enough to not estimate any starting points on “Street C”, farther to the left.)
Fares varied slightly because of these different starting points. For example, on one trip, the rider was south of the building. The drivers who were assumed to be starting on “Street B” had a slightly shorter route (and one less traffic light to navigate) to get to that rider than the drivers who were assumed to be on Street A. The estimated time from the riders’ pick up to the destination were all the same, further illustrating that the drivers’ starting points was the only factor contributing to the variations.
Here is a representative example of the different fares offered to drivers in the video:
We reviewed all trip requests offered to the drivers in the video that we could identify (using the images from the video as a starting point) and saw the same pattern: there were slight variations in estimated starting points that led to slight variations (often less than $.50) in the fares offered to drivers.
Why GPS Can Vary
Indoor GPS inaccuracies are just one reason for location discrepancies and other issues will exist even when drivers are parked near each other in an outdoor environment. The US Government explains that other factors include:
- Signal Reflection: GPS signals bouncing off walls or surfaces can confuse receivers.
- Device Differences: Phones have varying GPS hardware and algorithms, leading to inconsistent results. In fact, US Government testing shows that GPS accuracy varies based on the type of device (iPhone v. Android) and model of device (iPhone 13 v iPhone 12) that is used.
- Assisted GPS (A-GPS): Phones combine GPS, Wi-Fi, and cellular signals for location. Differences in Wi-Fi networks or signal strength can affect accuracy.
Uber Does Not Individually Personalize Upfront Fares
Upfront Fares are not personalized — our fares algorithms do not use information based on an individual driver’s personal characteristics (like their gender, race, or ethnicity) or past behavior on our platform (like their rating, past trips, or acceptance behavior) in formulating a fare offer for that driver. Our algorithms also do not use the type of device or operating system to calculate fares differently.
In certain cities, Uber offers promotions and rewards programs (such as Advantage Mode) that may be based on things like acceptance and cancellation rates. These programs apply the same to all drivers in the promotional or rewards group and are not personalized to individual drivers.
Riders May Experience Similar Fare Variations Due to GPS Discrepancies
Riders are not immune to GPS inaccuracies and may also experience similar fare variations if, for example, two riders request a trip to their “Current Location” standing side by side.
Closing Thoughts
We understand why drivers might compare fares and question variations. GPS inaccuracies, especially in situations like this, often explain these small differences.
Platforms like Uber have a responsibility to provide our users — drivers, couriers, restaurants and their customers — with transparency around how the platform works, including how fares are formulated. As we continue to innovate, we’re working toward the same north star — expanding access to good work and reliable transportation, hand in hand.
* This blog post describes Uber’s practices at the date of posting.