Moving fast in hardware: how the iPod shipped in 6 months
I’m really interested in building hardware fast. I was reading through Patrick Collison’s list of big projects that were completed quickly when I saw this entry:
iPod. Tony Fadell was hired to create the iPod in late January 2001. Steve Jobs greenlit the project in March 2001. They hired a contract manufacturer in April 2001, announced the product in October 2001, and shipped the first production iPod to customers in November 2001, around 290 days after getting started. Source: Tony Fadell.
Here’s the detailed timeline according to Tony Fadell:
Jan (1st week) – First phone call with Apple
Jan (3rd week) – First meeting with Apple
Jan (4th week) – Become a consultant leading iPod investigation, “what is possible?”. There’s no team, prototype, designs, nothing.
April (2nd week) – Become full time employee
April (3rd week) – Find contract manufacturer somewhere in Taiwan/HKG/Korea
May (2nd week) – Hired first employee on team
May (3rd week) – Pitch to Steve Jobs, project greenlit at the end of the meeting
October (4th week) – Launched iPod to the world
November (1st week) – Shipped first iPods to customers
Apple sold approximately 125,000 iPods by December 2001. How on earth did they go from hiring the first employee to work on it in May to shipping in November? This is an almost unheard of timeline for consumer hardware, especially given the volume. The answer: this timeline is highly misleading. But it still contains great lessons in fast hardware development.
Apple had been investigating the possibility of an mp3 player since as early as 2000, but hadn’t been able to find the right components to make something really compelling. Then Toshiba showed them a new 1.8″ hard drive. Jon Rubinstein, an Apple hardware executive, immediately recognized that this was the missing piece needed to make an iPod. Apple signed a purchase order for $10M and negotiated exclusive use.
Soon after, Rubinstein hired Tony Fadell to lead the iPod program. A handful of engineers were hired to design the architecture of the device and select the major components. One of the first things Fadell did was partner with PortalPlayer, a startup that made processors specifically for mp3 players. PortalPlayer had a reference design for an mp3 player based on their chip. Copying this existing reference design immediately got Apple about 80% of the way to the final iPod electronics. But Apple wasn’t the only company working on an mp3 player using PortalPlayer chips. In fact, there were 11 or 12 other companies also working on their own versions. But Fadell managed to convince PortalPlayer to stop supporting all of their other clients and focus 100% of their effort on helping Apple. About 280 employees at PortalPlayer dropped everything to develop the iPod. Apple requested support for things like larger playlists, AAC files, and an equalizer, and PortalPlayer implemented them.
The higher level software came from another startup called Pixio. Their OS was tweaked to become the iPod GUI. The rest of the electronics, like the power supply and the display driver, were based on previous designs Apple had used in other products.
But what about the iconic form factor? Contrary to popular belief, it was not summoned from the ether by Jony Ive. The basic shape was defined by the size of the hard drive. Phil Schiller, Apple’s head of marketing, was actually the one who came up with the iconic scroll wheel. He stole the idea from his home phone, made by a company called B&O. Apple immediately patented this “idea”, and later ended up licensing that patent back to B&O!
I don’t care how amazing your resources are, getting a product from an idea to putting 125,000 of them on shelves in 290 days is a major achievement. And it’s clear what laid at the heart of Apple’s engineering success: design re-use. Practically every part of the system was already de-risked. Apple could have started with a general purpose processor and written all the code needed to make an mp3 player, but instead they found a purpose-built part that they knew already worked. On top of that, they copied the application note! To some engineers, this might sound like cheating. “I’m not doing anything clever! This is someone else’s design!” they will say. That’s true, and that’s good. Because the end result for Apple was the iPod, done, in 6 months. And remember: design re-use doesn’t make building the product automatically trivial. The engineers that worked on the iPod worked extremely long hours and had no work-life balance. But no matter how hard they worked, they wouldn’t have been able to ship so quickly without relying on designs from PortalPlayer and Pixio.
Stop trying to re-invent everything. Use known working blocks from previous designs, and use application notes and reference designs. This will often result in designs that aren’t at their peak performance. In fact, the iPod almost didn’t ship because the battery life was so bad. This is where they applied some “original” engineering and patched the problem so that it was good enough for version one. Peak performance is not the point, especially for the first version. It’s so much more important to get things out the door. Doing this requires humility.
If going fast in hardware interests you, you might like my book Designing Electronics that Work.