Apple Silicon and Org Structures
Hello and welcome to the second edition of Hubble. Yes, this newsletter has a name now. Cause why not.
Also, few folks pinged me asking why I did not publish a newsletter last week. Well, in the previous edition I promised a "max weekly frequency", with the key operating word being "max". I don't plan on publishing more frequently than once a week. But I'm yet to reach a model of fixed frequency. So expect few weeks in between each post.
In this edition, I’m going to look at Apple’s decision to switch to its own chips for Macs and how org structure has a significant part to play in the execution of this. I’m going to follow a chain of thought and link out to various sources, particularly Stratechery, for more in-depth articles on various topics.
Read on!
Apple Silicon
Apple has long been touting the advantages of a hardware + software integrated product, right down to a custom built chip. They have been doing this on the iPhone for quite some time. But it didn't start with that on day-1.
The initial version of Apple iPhone, launched in 2007, was a fabulous device but slow. The phone was apparently put together with components used in Samsung DVD players, including a chip. In 2008, they started working on their own chip, with the belief that a custom chip is needed to offer the best experience.
Apple’s decision to design and build its own silicon was made sometime in 2008. At the time, the company had a mere 40 engineers working on integrating chips from an assortment of vendors. Then, in April of 2008, Apple bought a chip startup called P.A. Semi for $287 million. That increased the total number of chip engineers to about 150 and brought home expertise on what matters most on a phone: power efficiency. The fruits of this group’s labor were first revealed to the world in the iPad 4 and the iPhone 4. Those devices were powered by a processor named A4, which was a modified version of a chip design from ARM Holdings. The A4’s primary focus was to make the Retina displays shine.
Apple has been using custom chips to enable bunch of features.
Face ID needed A13 chip to be customised to support neural engine
Airdrop, Find My Device are powered by U1 ultra-wide band chip
Device encryption, SecureBoot and TouchID on Mac are powered by a T2 chip
To be honest, it's not like other hardware manufacturers do not have custom chips. Many of them have and are able to provide same feature sets. There's no major technical challenge why others cannot match feature-set and quality with their own hardware. Hence this ends up being a pretty debated topic. And I don't intend to kindle this debate.
Adding link to a nice read on Apple’s A13 Bionic Chip.
Apple Silicon in Mac
In WWDC 2020, Apple announced that they are moving to their own silicon on MacBooks too. This is super fascinating to me. And not due to the "integrated hardware software benefit" or what features it enables but due to the sheer audacity of it and the organisational setup needed to pull something off at this level.
Unlike software features or a hardware improvement like upgraded speakers or touchpad etc., changing the internals of the laptop is only possible through a multi-year strategy. Multi-year strategies in itself aren't anything new, but the speed and thoroughness of the execution across such a wide domain is impressive.
To understand the complexity of this, we need to understand the impact of this change. Changing the processor basically means changing how the entire system works. This means
Developing the new processor
Rewriting the OS
Rewriting app layer
Rewriting Mac's default apps for new architecture
Migrating 3rd party apps to new architecture
Supporting older apps which haven't been migrated to new architecture
Custom builds to support virtually running Windows/Linux which are written for Intel's x86 but not for Apple Silicon, and will never be
Oh and one more thing: iPad apps can now run on Mac with new Apple Silicon without requiring any changes.
All this requires coordination across a massive product development team which isn't for the faint hearted. And I think it has a lot to do with the organisational structure of Apple.
Steve Sinofsky, ex-President of Microsoft Windows, wrote a post on how Apple is making this possible which I highly recommend. Some excerpts from the post:
My jaw dropped when Tim Cook discussed the transition to Apple Silicon (ASi) as a two year journey. First, that’s like no time at all. Second, that’s an incredibly long time to tell everyone how long it will take and that they should be patient. Seriously.
7/ But really that is incredibly brave when so much could potentially change, more importantly could go wrong. Every big company does multi-year planning (I did) but everyone knows those plans mean little after a fiscal year. Apple is entirely different in that regard.
8/ The big thing about this is how Apple’s overall model of enables this to work. Every aspect of the system has to come together to create an environment where choices can be made AND supported that allow these plans to have integrity.
What I mean by Apple’s model is not about its direct to consumer business or vertical integration, but the culture of having a “point of view.” Apple makes products that customers love and are delighted by, but it makes them by studying technology, the market, and usage to arrive at plans and strategies. Unlike what you read in textbooks, Apple is much less about responding to micro changes, hype cycles, or even “feedback.” In fact you can see often how Apple’s model does not work so well when it rushes products to market or listens too closely to hype (eg Home Pod). Apple is a company that has a point of view — when the point of view lines up with a great product people love, it can become an unstoppable force.
Apple’s Org Structure
Apple is an exception in the FAANG list. Have a look at their leadership.
Notice something? While all others in FAANG are structured into divisions, Apple follows a unitary / functional organisation form. This means that there's no "Head of Mac" or "Head of iPhone". Instead, there's a "Head of Software", "Head of Hardware" etc. The alternative is a divisional org. For eg, Google has YouTube division with Susan Wojcicki as “CEO of YouTube”, Cloud division with Thomas Kurian as “CEO of Cloud” etc.
There are some very clear benefits of functional orgs but are much harder to pull off at scale.
Functional orgs generally have significant benefits of having coherent and focussed strategy and strong collaboration which leads to a lot of benefits. It makes it easier for Apple to cannibalise its own revenue from iPods with iPhones. Imagine taking this decision when the job of "Head of iPod" is in the line! In fact, this was a real problem for Microsoft when they launched Teams at the cost of Skype and Lync. Teams was built by a completely separate team not involving Skype, Lync orgs. And the product superseded both of these.
Steve Jobs has even credited org structure as a reason behind their great products:
[iPads] are post-PC devices that need to be even more intuitive and easier to use than a PC, and where the software and the hardware and the applications need to be intertwined in an even more seamless way than they are on a PC. We think we have the right architecture not just in silicon, but in our organisation, to build these kinds of products.
Divisional vs Functional Organisation
The lure of functional organisation is pretty high, to the extent that Microsoft tried to move to functional org from divisional org in 2013.
“we are rallying behind a single strategy as one company — not a collection of divisional strategies” and that the changes “will enable us to execute even better on our strategy to deliver a family of devices and services that best empower people for the activities they value most and the enterprise extensions and services that are most valuable to business.”
Source for this is a memo and strategy doc that Steve Ballmer wrote in 2013:
Ben Thompson was extremely critical of it and wrote about it in detail. In the article, Ben also talked about Divisional vs Functional org structure. Key excerpt below:
In a divisional organisation, different products are companies unto themselves. They have their own marketing, their own engineering, and their own finance. There may be some centralised functions, such as legal and HR, but everything that makes money for a product lives within that product’s organisation. Most crucially, each product has its own profit-and-loss statement (P&L). The performance of each division is thus clear to everyone from the CEO to the division presidents to Wall Street, and accountability is usually directly tied to the P&L. So is motivation: senior division leaders have much of their income tied to the performance of their division.
This model scales to a very high number of products. Consider General Electric, the classic example of a divisional company. It has twenty-five different businesses, ranging from finance to jet turbines. The competitive advantage of such companies is usually in their management acumen and capital reserves, and the preferred employee is a generalist, able to quickly master any job with a refined set of skills.
Functional organisations are the exact opposite: each function is siloed, and products cut across functions in a matrix-like fashion. This significantly expands the role of the CEO and his leadership team in all products, as that is the main point of coordination.
In this model, there is usually no ownership for a product’s P&L. Instead, everything accrues to the company’s all-up P&L, including compensation. In addition, the ideal employee is a specialist, not a generalist.
Most functional organisations are single-product companies; to put it another way, almost all companies start as functional organisations. In fact, the moment a company switches to divisions is as good a marking point as any for the end of startup-hood.
Apple is the exception that proves the rule; they are functionally organised, but that absolutely does not mean a functional organisation is best. Understanding why almost all corporations are organizes by divisions (and how Apple manages a functional organisation) illustrates why I think Ballmer is making a grave mistake.
Source:
Functional orgs are extremely hard to scale and hence not for all companies. It's definitely not the universal answer. For a lot of companies with multiple product lines, it’s significantly easier to manager a divisional org vs a functional org.
Andy Grove, former Chairman and CEO of Intel and the author of the classic book, High Output Management, has this to say on this topic:
Countless managers have tried to find the best mix of the two organisational forms. And it’s been no different at Intel, among senior management and throughout the ranks of hundreds of middle managers, who from time to time attempt to improve the organisation of the groups they supervise. But no matter how many times we have examined possible organisational forms, we have always concluded that there is simply no alternative to the hybrid organisational structure.
He goes to the extent of putting forth a law:
Grove’s Law: All large organisations with a common business purpose end up in a hybrid organisational form.
Making functional organisations work at scale is really, really hard. And that's one of the key reasons why Apple's structure is incredibly unique.
How Apple Makes Functional Org Work At Scale
Ben Thompson is a phenomenal writer. While I have read a whole bunch of articles, papers, books on org structure, Ben manages to summarise it the best. And it’s no surprise that he has written on this too.
As per Ben Thompson, here’s what you need to make a functional org work:
Functional organisations rely on true collaboration
Functional organisations rely on effective, visionary leaders
Functional organisations need to be focused
Functional organisations rely on experts
Head over to Stratechery to read more on this:
Overall, I believe that functional org works for Apple because of the common customer base across products and focussed product line which leads to highly integrated products. To the extent that the integrated product is actually a core part of their business strategy. But that’s not enough and Apple has the people structure needed to make the functional org work.
For companies that sell to different customer bases, divisional org is the preferred way to go, to the extent that functional org will be detrimental to their success.
For companies in between, odds are that going with divisional org is the best bet for them. Getting functional org wrong is easier and can be catastrophic.
That’s all, folks
Org structures is a fascinating topic. I can only sit and speculate from the outside on the impact of Apple’s org structure on its strategy. But damn is this interesting!
As a software dev 5 years ago, I liked to believe then that the company’s output was dependent on great functional work across product, engineering, sales, marketing. But increasingly I have realised that the output goes through a bunch of modifiers, a major one being org structure. And you really don’t want to ship the org structure.
(Greg Brockman is the Chairman and CTO of OpenAI and ex-CTO of Stripe)
Post Script
I would love to hear your thoughts on this post. Is this a topic that you found interesting? Any pointers on how I can make this more useful for you? Any suggestions of topics you think I should particularly write about?
Feel free to write back with feedback, thoughts, suggestions. If you got this in your inbox, just hit reply and I’ll get the mail. If you are reading this online, subscribe for future updates and shoot me a mail on hubble@smehta.io.
Until next time!