Did you know Uber Eats drivers share strategies on Reddit about which orders to skip, cancel, or pass on to less savvy drivers?

It’s almost a science — some might say pseudoscience — focused on spotting “tip baiters” and maximising efficiency.

After spending weeks roleplaying as an Uber Eats driver (and stealing everyone’s food), I uncovered an uncomfortable truth: many of these behaviours stem from gaps in Uber’s design.

Perhaps part of the solution lies in product psychology — the art of designing systems that prevent bad behaviour before it happens.

Here are three examples of how Uber could use “upstream thinking” to stop theft, reduce delays, and make drivers happier.

P.S., this is a summary from my own article here.

1. The “That’s Not My Door” Problem

When drivers deliver food, they’re sometimes required to upload a photo as proof.

But there’s no image recognition — so you can upload a picture of a brick wall, a bush, or even your shoe, and it’ll pass.

To prevent this, Uber could ask customers to upload a photo of their door ahead of time.

Drivers’ delivery photos could then be cross-checked against it.

If the system flags a mismatch — say, a blue door instead of a red one — the driver could get notified before leaving the order in the wrong spot.

A small tweak like this doesn’t just prevent errors; it also discourages intentional fraud.

2. The “Tip Baiting” Dilemma

Here’s how tip baiting works: customers promise a generous tip upfront to get their orders prioritized.

Then, after delivery, they reduce or remove the tip entirely.

Why does Uber allow this? Because the design focuses on clicks, not consequences.

A better approach would be:

• Show the driver’s effort, such as miles driven or time spent waiting.
• Add emotional context by showing the driver’s name and photo during the tipping flow.
• Frame what this means for the drivers.

By humanizing drivers and creating a sense of accountability, Uber could reduce this exploitative behavior.

3. Time-Sensitive Delivery Failures

If a driver can’t find the customer or get a PIN for delivery, they’re allowed to abandon the food after a short timer.

The app notifies customers with a generic call — but there’s no urgency.

Imagine if the customer’s app mirrored the driver’s countdown timer. F

lashing notifications and prompts like “Your food is about to be abandoned — tap here to connect” could nudge the customer into action before the driver gives up.

Simple changes like this reduce frustration for both parties and build trust.

The Real Cost of Reactive Design

Uber’s current approach to bad behavior is downstream and reactive: they punish drivers or refund customers only after something goes wrong.

But preventative design — built on principles of product psychology — could save costs, reduce theft, and create a better experience for everyone.

The lesson is clear: prevention always beats punishment. So why isn’t Uber experimenting with it?

If you’re digging this, you can read the full article here. It shows what happens when drivers literally just steal your food and leave it in a bush.

How Product Psychology Could Stop Uber Drivers From Stealing Food was originally published in UX Planet on Medium, where people are continuing the conversation by highlighting and responding to this story.