Relentless Improvement Engine
Give every team a measurable function and compete on learning speed
- Difficulty
- Expert
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 98%
Khosrowshahi describes Uber as a collection of teams, each organised around improving a specific function such as payment success, fraud, or app conversion. The whole company advances when every part owns a measurable target and keeps changing. The relevant benchmark is competitive: getting better is insufficient if rivals improve faster. Execution speed matters because it compresses learning time. A team taking two meaningful shots can receive two data points while another takes one, increasing both its information and its opportunities for success. But activity alone is not the engine. Khosrowshahi pairs speed of change with correctly identifying the opportunity. The method therefore links team-level goals, rapid attempts, observed learning, and comparison with the external rate of improvement.
Origin
Extracted from The Diary of a CEO. Dara Khosrowshahi describes Uber teams as individually targeted on improving their functions every day, with the company competing on both opportunity selection and speed of change.
Core principles
- 01Every part of the system should own an improvement target
- 02Improvement must outpace competitors, not merely beat the previous baseline
- 03Faster execution creates more attempts and more information
- 04Opportunity selection and speed of change jointly determine progress
How to run it
- 1
Map the Functions
Break the operating system into functions with clear owners. Define the customer or business result each function influences.
Pro tip Choose boundaries that let a team act without hiding cross-team dependencies.
- 2
Set Improvement Targets
Give every team a measurable outcome to improve, not merely a list of tasks. Track the target consistently while watching for gaming.
Pro tip Pair a performance measure with a quality or customer guardrail.
Watch out The transcript acknowledges that goals can be set badly or gamed.
- 3
Increase Meaningful Shots
Shorten the cycle from idea to tested change when risk permits. Ensure each attempt can produce information useful to the next one.
Pro tip Count learning-producing attempts, not raw activity.
Watch out More shots are harmful when failures carry unacceptable consequences.
- 4
Read the Data
Use results from each attempt to update the next decision. Stop work that is not improving the function or clarifying the opportunity.
Pro tip Record what each attempt changed in the team's belief.
- 5
Benchmark the Pace
Compare the team's improvement rate with competitors and market change. Accelerate, redirect, or change ownership when the function repeatedly falls behind.
Pro tip Benchmark trajectory as well as current level.
Watch out Do not sacrifice legal, ethical, or safety constraints to win on speed.
In the wild
Khosrowshahi names payment types, payment success, fraud, and mobile conversion as examples of functions with teams organised to improve them. He says the company expects every part to advance and frames faster attempts as a way to obtain more data and more chances of success.
→ Continuous improvement becomes distributed team ownership rather than a general executive slogan.
Common mistakes
Measuring Activity Instead of Improvement
A high volume of work does not show that a function or the company's judgement is getting better.
Benchmarking Only Against Yourself
Khosrowshahi argues that competitors are improving too, so an internally positive trend can still lose ground.
Speeding Through High-Cost Failure
The shots-on-goal mechanism should not be applied where rapid failure would create unacceptable harm.
Is it for you?
Best for
It is best for measurable, iterative operations where teams can run frequent changes and observe results.
Not ideal for
It is not ideal where speed would compromise safety, legality, quality, or the time required to observe a reliable outcome.
From the transcript
“every team is set up and organized for and goaled on improving”
“We have to get better faster than our competitors”
“every one shot that you take, I can take two shots”
From the episode
Uber CEO: At Uber, If You Don’t Perform, You’re Out! Uber Was Losing $3b A Year