When Amazon Forgot to Drink Its Own Medicine

The company that preaches "customer obsession" took a $170 million loss building a product for itself.

Customer-centric companies listen.
Evidence-backed companies test.
The best do both—they listen, test, and learn before they build.

Amazon's Working Backwards process is a masterclass in that discipline. It's deceptively simple and brutally difficult to follow with rigor.

Before a single line of code is written, teams must draft a one-page press release written in the customer's words. Not the company's vision. Not the technology's capability. The customer's actual experience.

Then comes the FAQ—a document that pressure-tests every assumption:

→ What problem are we solving?
→ Why will this delight them?
→ How will their life be better once it exists?
→ Is the market large enough to matter?
→ Can we technically build it?
→ What assumptions need proof before we proceed?

That narrative becomes a hypothesis.
The product becomes the experiment.
The market provides the evidence.

When this process is followed with discipline, it produces transformational products. When it's abandoned, even Amazon—the company that invented it—fails spectacularly.

When the discipline worked: Kindle

Amazon released the Kindle on November 19, 2007, for $399. It sold out in 5.5 hours.

The vision was elegantly simple: "Every book, every language, in 60 seconds."

But the execution required obsessive focus on the customer experience. The team chose E-Ink technology for paper-like comfort, even though it meant slower refresh rates and no color. They built Whispernet for instant downloads anywhere, absorbing the wireless costs themselves. They sold the device near cost to remove barriers to e-book adoption.

Every decision started with the customer's reading experience, not Amazon's business model.

By early 2011, just four years after launch, Amazon sold 115 e-books for every 100 paperbacks. The Kindle wasn't just a device. It was a gateway that spun Amazon's flywheel faster—more readers meant more publishers, which meant more content, which meant more loyal customers, which meant more data to improve recommendations.

It worked because it began with empathy and was validated by evidence at every step.

When the discipline vanished: Fire Phone

Then came the Fire Phone in July 2014, and the rigor disappeared.

The phone launched at $199 with a two-year AT&T contract ($649 unlocked). It featured Dynamic Perspective—a 3D interface powered by four front-facing cameras that tracked head movements. It included Firefly, which could identify products to buy on Amazon. It offered X-Ray for movie information and Mayday for 24/7 video tech support.

The features were technically impressive. But they solved the wrong problem.

Instead of asking "What phone experience would customers love?" Amazon asked "How do we get customers to buy more from us?"

The phone was designed to make purchasing from Amazon easier. It was locked to AT&T. It used a proprietary app store with about 240,000 apps compared to more than 1 million in Google Play. It lacked Gmail, YouTube, Google Maps, and other flagship apps customers relied on daily.

The Fire Phone wasn't built for customers. It was built for Amazon's business strategy.

The results were devastating:

→ Fewer than 35,000 units sold in the first 20 days
→ Amazon slashed the price to $0.99 (with contract) within weeks
→ $170 million write-down in unsold inventory
→ Production ceased in August 2015, just 13 months after launch

The contrast couldn't be clearer:

The Kindle team asked: "What reading experience would customers love?"
The Fire Phone team asked: "How do we get customers to buy more from us?"

One started with empathy and validated it with evidence.
The other started with strategy and ignored the evidence screaming that customers didn't want it.

The redemption: Learning to drink your own medicine again

Most companies would bury a failure this public. Amazon did something different.

Bezos told the Fire Phone leader, Ian Freed: "You can't, for one minute, feel bad about the Fire Phone. Promise me you won't lose a minute of sleep."

Then he added the part that reveals Amazon's real culture: "If you think that's a big failure, we're working on much bigger failures right now—and I am not kidding. Some of them are going to make the Fire Phone look like a tiny little blip."

But here's what mattered more than the encouragement: Amazon went back to its own medicine.

The Fire Phone had developed sophisticated voice recognition technology. Instead of doubling down on the failed product, Amazon asked a different question—one that started with the customer: "How can we help people in their homes?"

That question became Alexa. The Fire Phone's voice technology became the foundation for Echo smart speakers. But this time, they followed Working Backwards. They started with the customer problem (hands-free help at home), validated the demand, and built the product accordingly.

The Echo launched in November 2014, just months after the Fire Phone failure. It became one of Amazon's most successful products—not because the technology was revolutionary, but because they returned to asking the right questions in the right order.

The lesson: Systems beat intentions

Amazon didn't fail with the Fire Phone because they stopped caring about customers. They failed because they stopped following their own system.

Empathy defines the story.
Evidence tests it.
Hubris breaks both.

The Kindle began with a question worth asking: "What reading experience would customers love?"

The Fire Phone began with an answer no one needed: "A phone that makes it easier to buy from Amazon."

The Echo began by learning from both: "How can we help people in their homes?"—then testing that hypothesis relentlessly before launch.

This is why transformation work must be grounded in systems, not slogans. "Customer obsession" is a nice value statement. Working Backwards is a testable system that makes customer obsession operational.

When you skip the system—even if you still believe in the value—you end up building products for yourself while thinking you're building them for customers.

That's the real cost of skipping discovery—and the reward of returning to it.