The Unsung Hero in Apple’s Strategy

TL;DR version: It’s their A-series processors. However, I’ll lay out many of the reasons why this single component may be the source of Apple’s most significant competitive advantage.

In 2008, The year Apple bought P.A. Semiconductor, I was working with a company specializing in semiconductor industry IP licensing. The CEO became a good friend and he was a long-time exec in Motorola’s chipset business back in the day. Suffice to say, our meeting shortly after the news dropped was an interesting one. One of the key parts of this discussion was centered around how the strategy by semiconductor companies is to differentiate their chips by their designs. Meaning, how they put together the entire solution either onto a single chip or with many chips (co-processors) working together. Each vendor selling chips to others, generally, had a custom design and solution. What the smart semiconductor insiders knew with this acquisition of P.A Semiconductor was Apple was going to start designing their own chips. However, unlike others in the competitive chipset design business (Intel, Qualcomm, Nvidia, Broadcom, and a few others), Apple would not be designing these chips to sell to others but exclusively for their own devices and operating systems.

As I’ve studied the semiconductor industry for over 15 years now and the nuances between each chipset designer are fascinating. The way each one designs their chips to solve market problems and compete against other vendors is unique and creative. But all of these vendors have to design into their SoC or entire chipset solution the broadest of solutions in order to capture as many different parts of the market as possible. So a vendor like Qualcomm will design their architecture and then tailor it for different parts of the market. The initial design is broad, yet its applications can be made more specific. The same is true with Intel. They must cover the broad market with their solutions. Apple, on the other hand, does not, meaning every bit of their engineering resources, architecture designs, and nuances in optimization have a single purpose — to make iOS the most sophisticated, capable, and powerful mobile operating system on the market. Apple has been acquiring some of the best chipset designers in the world and they are tasked with making the most powerful and battery-efficient designs in the industry. These designs are exclusive to iOS while, at the same time, being competitive with the broader market chipset vendors as well. Without being a semiconductor industry insider, it is hard to internalize just how big of a deal the A-series processor design is in the grand scheme of things.

Prior to designing their own chips, Apple’s vertical hardware and software intergration stood apart because they could tune the software to the hardware, which contained many parts of the puzzle they did not own, and make specific software optomizations to accompany that hardware. This has only extended to a much deeper level now that they can tune/design the CPU/GPU and many other parts, like display drivers, motion processors, image sensors, and a host of other critical things, into the design and optimizations of other hardware and, most importantly, the software.

The results of Apple owning so many of the key components is not just the visual performance in terms of speed or graphics, but how fast Touch ID is, for example. Or their superior protection by being able to do hardware level security integrated into the chipset via the secure enclave, and locking in safeguards at the SoC level. So overall superior performance, security, and new software capabilities like 3D touch, are just a few of the things Apple can do to integrate and differentiate on their platforms. There are more to list and more will come, but there are two key points I’d like to draw out.

Other companies will be very slow to catch up. No vendor has quite nailed Touch ID the same way Apple has in terms of reliability and accuracy. The new Samsung devices I’ve tried are close, but experts have told me there are security concerns. They got one part close but not the other key ingredient. The real observation to be made here is what Apple has done to the premium hardware sector in smartphones, PCs, tablets, etc., in that it is mostly non-existent. In nearly every computing category Apple competes in — tablets, PCs, and smartphones — they own between 75% and 90% of the premium category. The impact of this reality is other vendors who compete in these spaces will find it more difficult to justify investments in leading edge components (even if they exist) because they won’t sell enough in volume. Samsung as an example, will lead in things like curved glass or other features which they own the supply chain for, but their ROI on those components won’t come from their handset business but from other companies. They key question here is, are there other customers for high end components other than Apple? If there are not, then even Samsung’s latest will be delayed due to lack of customers willing to pay premiums for curved glass or other things they focus on. And with the coming near-impossibility to sell an Android handset for more than $500, I do not forsee any other hardware vendors out there able to buy premium components en mass and commercialize them at the high-end where they must be due to their price.

Apple Suppler Leverage. Once Apple becomes the main customer for all high-end innovations, they acquire the upper hand in supplier and supply chain negotiations. Take this latest move to dual-source the A9 from Samsung and TSMC. Samsung is on 14nm and TSMC was on 16nm. Samsung had the capacity to make all of Apple’s chips but Apple chose to dual-source to not only keep their suppliers honest, but to give business to TSMC so they can use the money to invest in their next process node at 10nm. TSMC has prioritized Apple to the point of aggravating other very well known semiconductor companies, causing delays for others to get their latest chips to market. Should TSMC make the jump to 10nm and the process is mature, Apple would be the only customer for it, as they are the only ones who can afford to pay the premium. This gives Apple a great deal of leverage. This is just one example of many I hear from my contacts in the supply chain where Apple has the upper hand.

When I put all this together, what stands out to me is, in this race, Apple is sprinting and everyone else is walking. Others in the market may get there but Apple will get there first. The markets they compete in really are theirs to lose and owning all the key parts of the hardware and software and increasingly, services chain, means it’s hard to see how they lose it any time soon.

Published by

Ben Bajarin

Ben Bajarin is a Principal Analyst and the head of primary research at Creative Strategies, Inc - An industry analysis, market intelligence and research firm located in Silicon Valley. His primary focus is consumer technology and market trend research and he is responsible for studying over 30 countries. Full Bio

37 thoughts on “The Unsung Hero in Apple’s Strategy”

  1. If you were to apply Clayton Christensen’s theories to this observation, it would be a classical example of integration vs. modularity. The advantages of integration vs. modularity and vice versa are exactly as you describe.

    Christensen suggests that the forces that pushes a market towards integration or modularity are whether the product/technology has reached “good enough”. If the technology has matured to the point where sufficient performance can be obtained with a generic design, then the market trends towards modularity. On the other hand, if the extra performance that can be squeezed out by carefully optimising for the whole product is meaningful, and customers are willing to pay more for it, then the market will favour integration. In terms of cause and effect, the balance between integration and modularity is defined by whether the product is “good enough”. The ceiling is what consumers will accept as “good enough”.

    If you believe in this view, then the reason why Apple can sprint while everyone is walking is not only because Apple has the legs, but also because Apple’s customers appreciate the extra features, whereas Android users somehow do not. Apple can somehow push its own ceiling higher.

    This is what I find fascinating. Apple somehow manages to make the features it chooses to introduce, features that require a high-level of integration to implement, into must have features that customers are willing to pay hundreds of dollars for. On the other hand, Android OEMs are mostly unable to do this and often find themselves wasting money on features few people will pay extra for.

    1. I think I’ve mentioned this before, but the birth of a true global pure consumer market is teaching us some new things that Clay’s theory need to account for, and have had trouble accounting for. The largest being, using the word “market” to be wholly representative. In consumer tech there is no market but really lots of little markets within a market. This is why Apple can do things, others can not, and even if Android vendors tried to integrate it would not work. Because for them modular dynamics are at play. Apple is going after a different market, one that is large enough thanks to the size of global consumer bases, that allows them to be vertical and slice the market in their favor.

      The one thing that always hit Clay’s theory, in my mind, was size. Most the examples he used were within very small TAMs. As I’ve told Horace, and he agrees, this pure consumer play is a bit of a blind spot because again the market is not one market but many. The challenge for his theory is to address a market like CPG. CPG teaches us quite about the mini markets that make up markets and strategies that work and don’t work within them.

      I’ve always thought the lessons consumer tech companies need to learn should come from CPG. That’s the most mature consumer area to study by far.

      1. Yes. Actually that was what I was hinting at by saying Apple users appreciate advanced features while Android users don’t seem to as much. The smartphone market is clearly not a single market.

        I’m not sure if the smartphone market is too complex for disruption theory to tackle. As you have written in Techpinions before, if you separate the Android market from the iPhone market, then what has happened to Samsung closely follows what disruption theory would predict. In fact, his “law of conservation of attractive profits” is also playing out quite well as a explanation for the improved profits of Samsung’s semiconductor arm and Sony’s component business. Hence despite the underlying huge complexity of the market, maybe it is sufficient to simply separate Android and iOS. Of course the sheer size of iOS and its impact on the market as a whole makes this separation challenging sometimes.

        Because I look at this from the demand side and not the tech side, what fascinates me is how Apple is able to create the demand to warrant an integrated approach. How Apple whets their customer’s appetite for more advanced features, even as the Android side struggles as customers find their mid-, low-range products “good enough”. I think it’s much more than design or luxury, and actually goes far back even to the Apple II era, when Jobs realised that there are more software enthusiasts than hardware ones. But that’s a different topic.

        Once you get the market sorted out, yes, the global scale enables Apple to be much more vertically integrated than before, as you explain very well in your article.

        My question would be, isn’t separating the Android and iOS markets sufficient in terms of granularity to apply disruption theory successfully, especially on the Android side? Further granularity might provide more insight, but can significantly increase complexity and make data harder to come by.

        1. It probably is wise to separate Android and iOS in order to apply disruption theory. I’ve said for along time Apple essentially has no competition (most of the tech industry views Apple’s approach as wrong and I don’t see any others truly interested in following a similar path). When I look at what Apple is doing, it’s all about the user experience, and when viewed through that lens, the user experience can never be good enough.

          Most of the criticism of Apple seems to be a denial that there’s anything different about the user experience Apple delivers, but that isn’t good analysis, it’s just a denial of reality.

          1. Of course, the problem with separating markets for the purpose of analysis is that you could end up with a lot of separate theories for separate occasions, which means that you don’t have any real theories at all but simply a collection of explanations/observations. This is basically a case of overfitting.

            https://en.wikipedia.org/wiki/Overfitting

            That is one of my strong objections to the ideas that for example, Ben Thompson have put out. Hence I also try hard to see how we can understand the market without a separation (i.e. simplify the model).

            I tend to think that iPhone and Android markets do interact, and that Android was a low-end disruption to the iPhone. Android clearly enabled many more people to own touch-based smartphones and to connect to the Internet, etc. Android’s contribution to the world cannot be understated, and in this sense, it was a hugely beneficial low-end disruption. However, whereas low-end disruptions like the PC improved to the point where they displaced mini-computers and workstations entirely, Android is having difficulty catching up with the iPhone, and in fact the gap may be widening. Low-end disruption did occur, but only up till a certain point.

            This in itself is not unheard of, and I think the resurgence of Swiss watches relative to quartz is another good example where the entrants have failed to capture the high-end.

            The interesting thing is that with Samsung’s case, we are seeing a disruption within a disruption so to speak, where low-end disruption is happening within the Android market, but is not affecting the iPhone market. It seems that disruptions can be nested (in computer language speak), and one might even make the observations that low-end entrants are much more fragile to successive low-end entrants, whereas some high-end incumbents seem to be isolated.

            It would be interesting to understand what makes some high-end incumbents more resilient to the final stages of low-end disruption. It is actually very interesting from an East Asian perspective, because although Japanese and Korean products have been very good from even a quality perspective, we have often failed to become one of the high-end incumbents that are resilient to disruption.

          2. “It would be interesting to understand what makes some high-end incumbents more resilient to the final stages of low-end disruption.”

            I think it’s the user experience, when the value you deliver to the consumer is woven into the fabric of the user experience, and this happens in a situation where the buyer is also the end user (consumer markets). When this is the case I think it’s very hard to disrupt. Apple could falter on the user experience of course. If my experience of value re: Apple’s products and services is no longer good, I will look for other solutions.

            I wonder if Apple’s position is unique. What other industry even allows for the kind of integration Apple is doing? And does any other industry have such a high touch relationship between user and vendor? I can’t think of any off the top of my head, but perhaps they do exist.

            Anyway, the short version might be that when your product is the user experience, ‘good enough’ simply isn’t a factor since the user experience can never be good enough. And this of course self selects the high end of any market because as anyone in sales knows, the buyers prioritizing price are the majority of a given market.

          3. I actually don’t agree to that explanation which I often see online. I think that view fails to expand the idea beyond Apple and hence only explains Apple. Unless you can apply the same idea to similar and independent observations, than the validity of that argument as a generally applicable theory is questionable at best.

            There are other examples of markets where the low-end disruptor has failed to displace the incumbent which fled to the high-end. Unfortunately I do not study industries extensively and can only bring up a few examples, but watches and automobiles come to mind. In both cases, low-end entrants (many from Japan) entered the market and especially in the case of watches, disrupted the market. However, they failed to capture the high-end market. However in neither of these cases was the high-end indisputably better in terms of user experience. In the case of mechanical Swiss watches, you have to adjust the time every once in a while and they are heavier and bulkier. In the case of high-end German automobiles, I would bet that Toyotas are more reliable and require fewer repairs.

            Looking at these examples, my observation is that user experience per se is probably not the answer. I lean towards differentiation, brand and a symbolic representation of values, all of which are criteria satisfied by Apple but not Samsung. But again, I have not studied enough examples.

          4. “However in neither of these cases was the high-end indisputably better in terms of user experience.”

            I don’t agree with this. The differences may be subtle and only appreciated (or valued) by a segment of the market, but the user experience was likely better enough that the low-end disruptors could not entice the high end buyers.

            The user experience is not simply the functional aspect. Design, craftsmanship, support, quality, trust, results, and much more, are all part of it. These aspects of the user experience can be hard to quantify, and many of these aspects are not valued by many buyers. This leads many to the erroneous conclusion that these same aspects have zero value.

          5. I appears that a good analysis of why people buy what they buy is in order. A jobs to be done analysis.

            Without that, any theory including mine, has only very weak foundations.

          6. Agreed. When I’m considering jobs-to-be-done it is not only the what that matters but the how. Looking at only the what is why we have many comments on this very site about Device X from Company Y offering the same capabilities as some Apple device (false equivalence) at half the price and years earlier. Although when the buyer doesn’t value the aspects of the user experience that are in fact different, for that buyer we could say it is true that the two products are the same. But for a different buyer the two products are not the same at all. It depends on the what AND the how.

          7. True. It does seem that brand and symbolic representation of values are the differentiation behind the high-end car industry. I think the same would apply for high end fashion.

            Another aspect in some industries is customer service, and maybe Apple leads there(not sure,i’m not from the u.s.).

            Also let’s not forget, phones and operating systems are unique in their ability to lock-in users, so that’s something we probably won’t see in other industries.

          8. Although I’m not absolutely sure since I don’t own high end cars, watches or fashion myself, the service that my Toyota dealer is giving me is very good. Although I’m sure that there is a home turf advantage, it seems that here in general, people who value service, quality, reliability and cost performance choose domestic cars. The reason they may buy pricier imported cars lies elsewhere. Hence while I think that Apple’s service is very good, I don’t see customer service as being an factor that we can apply generically to other markets to understand why the high-end is resilient to low-end disruption.

            Also regarding lock-in, my point of view is that it is overstated in smartphone platforms. The apps that most people use (Facebook, Twitter, Line, Whatsapp etc.) are all cross platform and free. The ones that are not are mostly iCloud apps (iMessage, calendar), which are interestingly the least respected. What we may see though is a general reluctance to change what is working. If I remember correctly, statistics for switching carriers for example, show that people don’t switch much. Carriers seem to be much better at lock in, even without owning operating systems.

            http://www.wsj.com/articles/SB10001424127887323971204578630202425653338

            Getting to the true reason why some high-end products remain resilient to low-end disruption is very hard in my opinion.

          9. * i agree that people hate change. So now you have to change a phone you have to:

            1. Learn the OS and get used to it.

            2. Learn a new design language expressed via apps.

            3. Install all your apps. Lose data for apps that only store data locally. Change some apps – because certain apps aren’t exactly the same over platforms.maybe some other apps don’t exist. need to buy apps again. lose bookmarks. probably some other small details.

            4. Lose the ability to use imessage. i heard many on reddit.com/r/android complaining that that’s the thing that keeps them on IOS.

            5. If you’re not a mobile hobbyist, just user – you actually don’t know how much change this entails. so there’s risk.

            6. If you have itunes music library built over years, you’ll lose it.

            If you add together all the cognitive price that this demands, and consider that people spend a lot of time via phones, this seems like a much bigger change than changing carriers(just investing 1-2 hours). Also , it does cost money(itunes+apps).

            * regarding service – my point wasn’t that service is a differentiation in cars. Maybe it isn’t.

            But maybe it is a differentiation in phones – Apple uses margin/volume for better service, while android competition and low-margins makes this harder.

          10. My point is that lock-in due to operating system differences is not necessarily high compared to switching costs in other industries. A simple 24-month contract scheme seems to be at least as effective, maybe more. I think the lock in effects of operating systems is far over-rated.

            Regarding service, if you are satisfied with with an ad hoc explanation for each market that you discuss, then your reasoning makes sense. If you’re looking for a generally applicable theory, then no. You have to either dig deeper or look elsewhere for an explanation.

          11. Regarding service: the way to test this will be to look at industries where there’s a meaningful difference between the service levels between the high end offers and others. Maybe the car industry isn’t such an industry. But there are other industries that use better service to charge higher prices. The hotel industry for example.

            Also it’s just intuitive – some people value a better service. And Apple markets(and offers) it’s service in the u.s. better than others. Sure it’s not the only thing , but it’s a factor.

            As for the lock-in in contracts, which definitely are a strong lock-in mechanism, probably much stronger than you’ll find in most industries. But that doesn’t mean the OS lock-in is weak.

            Also didn’t the move to installments decrease the lock-in for carriers greatly ?

          12. That way, you’re basically only saying that “service matters when it matters”. If you really want to test if service makes a difference, you collect all cases and see if service levels positively correlate to resilience to low-end disruption.

          13. We know that high quality of service matters in luxury products. At least that’s how they train luxury marketers. We also know that customer service is a highly potent tool to “enforce” brand values. That’s also what’s recommended to marketers.

            So maybe it’s value is mostly about helping to strengthen the brand?

    2. From what I understand of Clayton Christensen’s theory that he got wrong is that he is looking at the world as if it was the way it was in the 1980’s and 1990’s. The problem as Ben Bajarin and Ben Thompson understand is that now consumers are driving the market. This means that things that Clay’s theory does not account for at all that are quite important to consumers that do not matter at all to business buyers. Things like ease of use, design, style, fashion, privacy, security, support and easy updates.

      For consumers these things are quite important as they do not have their own help desk or IT staff to call on when something goes wrong along with guide them along with what to do and what not to do. Plus as we rely and trust more of our lives to these smartphones and the marginal costs of getting the best are so small, why would you not get the best smartphone you can assuming you can afford it.

      1. I am aware of the arguments mainly laid out by Ben Thompson but I find them rather disingenuous. By far my main gripe is that disruption theory has not been limited to B2B markets at all, and that there are many, many examples of it being applied to B2C markets. A simple review of Wikipedia will show you many consumer market examples and you can bet that there are many many more. A further gripe is that he mainly bases his discussion on the Apple anomaly and very few other examples.

        https://en.wikipedia.org/wiki/Disruptive_innovation#Examples_of_disruptive_innovations

        To conclude that disruption theory does not apply to consumer markets based on the ideas of a few pundits who can only point to a few examples, is totally amazing.

        Of course, there will always be anomalies and who knows, it’s even possible that the theory it totally wrong in many important aspects. However, to dismiss a carefully built theory that has many examples in many fields, I expect the same level of care and thoughtfulness. I haven’t seen that at all.

  2. Thanks Ben.

    It’s nice to read a piece that evaluates the big picture surrounding tight integration. There’s no shortage of tech articles giving the impression hardware engineers just shuffle around off-the-shelf parts at the bequest of marketing.

    1. I’m sure our resident Anti-Apple-Know-It-All will enlighten us on the subject of Apple doing Nothing Special At All ™.

          1. And here I was hoping it was me… Sigh… 🙂
            Regarding the other person, being I agree with them more than I disagree, I suppose it’s close enough. 🙂

      1. Yep. I know exactly who you are talking about. The one who seems to be confused as to what Apple’s mission is and who their customers are. Hint: It is not him!

  3. Running vs walking….

    Do you believe that the A10 and A10X will again nearly double the performance of the previous gen A9 chips as the the A9s doubled the previous A8s?

    1. I think Apple is starting to run up against technological limits. If you compare the iPhone 3GS to the 6s then the sources of performance improvement are:
      1) clock speed went from 600 to 1850Mhz. The fastest Intels run at 4000Mhz but run rather hot and are not battery friendly, so using current manufacturing technology you’d probably not go much beyond 2500Mhz (see various Androids).
      2) we went from 1 to 2 cores. Another doubling to 4 cores is certainly possible but there are not many workloads that I can think of that benefit from say 8 cores in a handheld device. That said, Intel appears to scale a bit better in things like Geekbench tests (i.e. the Apple 2-core speed is only 70% faster than the 1-core speed).
      3) instructions per clock have roughly quadrupled, which includes the move to 64 bits. Low-clock Intel chips like in the new MacBook can do double the amount of work per clock when compared to the A9 and the fastest Intel chips in the iMac. In practice, the fastest Intel processors deliver maybe double the A9 performance per core. Still, Intel’s amount of work per clock has increased about 10%/year.

      So in theory each of these dimensions has the potential to double, but they cannot all double at the same time because of trade offs. Several wildcards that are hard to predict:
      1) Apple controls the A-series and the Swift programming language, which may allow them to create some clever optimisations.
      2) Apple may be able to dedicate additional silicon to CPU intensive tasks.

      In short: a doubling in performance is possible but now that most of the low hanging fruit is gone, I would suspect that the performance increases will be spread over a few generations.

      1. Thanks. I agree the low hanging fruit is gone. But more is on the way! (I hope.)

        TSMC will be making 10nm A9s next year, good for 30% if the architecture doesn’t change, but Apple has more than a 1000 chip designers paid to change things. So, next year, a tick_tock year (architectural AND process node changes) might see 50, guessing that the architecture upgrade will provide a 20% boost plus the 30% boost of going to 10nm (Thanks TSMC.)

        So, 50% would be sweet, but is there anything that gets us to 80 or 90%?

        1. It seems this year was the tick as they made the jump from 20 to 14 and 16nm. I don’t think TSMC will be at 10 next year but maybe the year after.

          Apple is able to squeeze a lot of performance out because they tightly integrate the software to the hardware. I do expect heavy optimizations but the perf gains may very well be in graphics, and importantly in power. Being able to get some power gains but a lot more battery life will be an upside of the tock next year.

          1. Yup. I crammed two years into one. it’s A10 still on 14/16nm. A11 on 10nm if all goes as planned.

    2. “Do you believe that the A10 and A10X will again nearly double the
      performance of the previous gen A9 chips as the the A9s doubled the
      previous A8s?”

      Apple might be able to squeeze more optimizations out of next year’s CPUs. However, they’ve hit at least one wall — up until now, they were always a step or two behind the bleeding edge in terms of the process node they made the CPUs on. Which means they were always able to improve energy efficiency by simply upgrading to an improved process. This year, for the first time, they are on the current cutting-edge process node, so they will have to wait along with everyone else for the next node to become available (in 3 years or so) to get any benefits from there.

  4. Ben

    Out of interest…what are the security concerns regarding the version of Touch ID implemented by other vendors?

    I’m assuming it’s related to the security enclave within the iPhone which holds the fingerprint ‘representation’?

Leave a Reply

Your email address will not be published. Required fields are marked *