The secrets of Amazon’s leadership
Inside Amazon’s Leadership System: What Most Companies Miss About “Customer Obsession”
For years, Amazon has been described in shorthand.
Customer obsession. Long-term thinking. Data-driven decisions.
The phrases are familiar — almost cliché.
But clichés only emerge when something real gets copied badly.
In this conversation with Rich Wasserman, former Technical Director at Amazon (2004–2018), what emerges isn’t a sloganised version of leadership — but a disciplined operating system. One that explains why Amazon can scale into wildly different domains without losing coherence, and why so many organisations fail even when they borrow the language.
The Question That Dominated Every Room
According to Rich, nearly every serious discussion at Amazon — from strategy to technology investment — came back to a single test:
Is this good for the customer now, and in the long term?
What’s striking isn’t that the question was asked — many organisations ask it.
It’s that Amazon refused to accept false trade-offs.
Price or service?
Speed or quality?
Choice or relevance?
The default response was not to choose — but to reject the framing entirely and search for a solution that raised all three. Compromise was treated as a failure of imagination, not realism.
Jeff Bezos had a line that captured this posture perfectly:
“If we wanted to solve easy problems, we wouldn’t hire smart people.”
How Amazon Really Prioritised Work
Like every complex organisation, Amazon always had more initiatives than it could possibly execute. The difference was not a lack of ambition — but clarity about what actually mattered.
Teams typically knew:
What was unacceptable for customers right now
What single improvement would create the next meaningful step-change
Where investment would compound over time, not just patch symptoms
Crucially, Amazon invested heavily in systems that would still matter years later, even when that made life harder for engineers in the short term. Technical elegance was secondary. Customer impact wasn’t.
Some systems, Rich admits, were held together with “duct tape and baling wire” — not because of carelessness, but because customers came first.
When “Customer-First” Requires Doing Less for the Customer
One of the most revealing examples involved Amazon’s Subscribe & Save programme.
Originally, items shipped exactly when customers ordered them — resulting in dozens of fragmented deliveries. From a surface-level customer perspective, that felt responsive.
From a systems perspective, it was a disaster.
Splitting shipments drove enormous cost, which ultimately constrained Amazon’s ability to offer lower prices — the very promise of the programme.
The counter-intuitive decision was to delay some deliveries and consolidate them — while being radically transparent with customers and giving them control.
The lesson is subtle but profound:
Amazon distinguished between what customers say they want and what serves them best over time.
Leadership That Actually Knows the Work
One of the most consistent themes in the conversation was proximity to reality.
Senior leaders at Amazon were expected — not encouraged, expected — to understand the actual work. Deeply.
Executives routinely:
Visited fulfilment centres and call centres outside their remit
Asked detailed, sometimes uncomfortable questions
Treated anecdotes as signals, not noise
Distrusted data that didn’t align with lived experience
Rich captured it simply:
When the anecdote doesn’t match the data, you probably shouldn’t trust the data yet.
This wasn’t anti-data thinking. It was data in service of understanding, not authority.
Decision-Making Without Ego
When disagreements arose — and they often did — Amazon didn’t resolve them by hierarchy or persuasion.
Instead, leaders asked:
What would have to be true for each option to be right?
Can we test this safely?
Can we observe the work directly?
In one meeting, a senior executive literally dialled into customer service live to see what customers experienced — rather than debate hypotheticals.
Truth was allowed to interrupt opinion.
Experiments — But Not Everywhere
Amazon ran experiments obsessively — but not indiscriminately.
The company invested heavily in robust experimentation platforms, causal modelling, and long-term cohort analysis. But it also recognised that not everything was up for debate.
No one needed an experiment to prove customers dislike higher prices or worse service.
The key distinction was between:
Two-way doors — reversible decisions that should be tested aggressively
One-way doors — fundamental choices requiring deep thought and caution
Failure was accepted — even expected — when learning was real and risk was contained.
Why Amazon’s “Leadership Principles” Actually Worked
Amazon’s leadership principles weren’t motivational posters. They were operational constraints.
Teams documented their tenets — explicit beliefs about trade-offs — so decisions didn’t have to be re-litigated endlessly.
For example:
Buyers and sellers are both customers
When interests conflict, buyer experience wins
That single sentence represented hundreds of hours of debate — distilled so that people closest to the work could act decisively without escalation.
Tenets weren’t slogans. They were decision accelerators.
Targets Without Target-Driven Behaviour
Amazon used goals — but cautiously.
Most goals were:
Close to customer experience
Observable and functional
Paired with input measures teams actually controlled
Crucially, promotions and compensation were not tied to hitting targets. That removed the incentive to game numbers or defend broken metrics.
If a goal turned out to be wrong, changing it was seen as learning — not failure.
That alone explains why so many organisations struggle to replicate Amazon’s approach.
The “One Thing” — Without Simplifying the Business
Amazon did many things at once — but it didn’t manage them that way.
The breakthrough was the concept of the single-threaded leader: someone accountable for an end-to-end domain, with authority across product, engineering, science, and economics.
Not a coordinator.
An owner.
This structure allowed Amazon to identify and work on the true constraint in each domain — without drowning in cross-functional committees.
As Rich put it, most organisations don’t need a restructure to benefit from this — they need shared end-to-end goals and clarity of ownership.
Why This All Matters
Amazon’s success wasn’t luck.
It wasn’t charisma.
And it certainly wasn’t slogans.
It was the cumulative effect of:
Leaders who understood the work
Decisions grounded in customer impact
Data used as a lens, not a weapon
Systems designed for learning, not compliance
And an organisational structure that respected focus
Most companies borrow the words.
Very few adopt the discipline.
And that’s the real difference.
To watch the full interview go here.
Stuart Corrigan was voted most influential consultant 2026 and is an Amazon No1 Best Seller
https://www.loweffort.com/
