The Critical Path #130: Determinism vs. Probabilism

Horace and Anders discuss Apple Watch pricing, Consumerism and Planned Obsolescence. In the closing segment Horace presents a new dichotomy of company values.

via 5by5 | The Critical Path #130: Determinism vs. Probabilism.

  • stefnagel

    “Reassuringly expensive” Good description for both the gift and the status market sectors.

    In book publishing, you can aim at selling a million books at buck apiece or one book for a million bucks. It does matter which pole you aim at. My mentor used to say, more usefully, “do a good book, price it to do another good book.”

    Or you can just do The Good Book:'s_Bible

  • Walt French

    From a former portfolio manager, take the note that high diversification à la Google (more so: Amazon) is great in the absence of frictional costs.

    But communications costs are huge in the smartphone marketplace: learning how to reset privacy preferences, copy over music & address books, manage multiple email inboxes, etc., can be a HUGE time suck, worth hundreds or more in lost time.

    Amazon’s attempt to apply its low-risk, low-margin “bazaar model” to smartphones is not so much an attempt at disruption as a simple failure to understand why people buy other products from them. The “probabilistic,” “law of large numbers” approaches just don’t match products with a lot of brand identification and non-genericized product info.

    • stefnagel

      Jumping thee minnow?

    • I totally agree that the probabilistic approach will be much less effective for certain products. In addition to brand driven and complex products, I would like to add that new category devices are also difficult with a probabilistic view.

      The probabilistic view will probably work ok if you are copying or acquiring a product that has already gained acceptance in the market. Communication has already been done by the pioneers.

      Applied to Google, it explains why search, maps, gmail, android worked well because these had successful precedences. It also explains why Chromebooks are taking so long because there were none.

      Google TV vs. Chromecast is another good example. Google TV failed miserably as an attempt at a new category, but Chromecast was ok because they were following the lead of Apple TV.

  • Jacob Williams

    Here’s a scenario:
    – The iPhone name is retired and the next would be iPhone is called a Mac Nano with retina display. It has OSX and IOS modes.
    – Apple comes out with the standalone 5K monitor that only works with the Mac Nano and Mac Mini.
    – Apple charges a premium on the 5K monitors because people already have the Mac Nano and $2,200 for a 27″ 5K computer is a steal.
    – Apple sells a lot of 5K monitors.

    This is a great way to get the massive pool of iPhone users on a Mac. It’s also a means to get current Mac users without 5K to make the jump.

    Most importantly it puts Mac on the two year upgrade cycle. The iPhones are “good enough” as you say for phones. If our desktop performance was contingent on our iPhone’s performance, than we’d upgrade them every two years. It’s a clever way to get our desktops subsidized.

    Does anyone see unintended consequences with this strategy?

    • Bruce_Mc

      The iPhone brand name is much better known than the Mac brand name. Killing it off is wasteful.

      Single point of failure. Would it be nice to have a desktop computer that stops working when my phone is in the other room or the battery runs down? No.

      Do you change the phone processor to Intel and make all of the iPhone app developers recompile their apps, or keep the phone at ARM and make all of the Mac developers recompile their apps. No, and no.

      “Most importantly it puts Mac on the two year upgrade cycle.” Is extracting the maximum amount of cash from each user as quickly as possible Apple’s most important priority? No.

      • Jacob Williams

        Each point:

        The “i” prefix is also a good brand and they didn’t use it with Apple Watch.

        The monitor plugs into the bottom of the phone and charges it while it’s plugged in.

        The computer wouldn’t stop working if the phone left the room. The monitor would just go to sleep. The phone is the computer and will still be on and working… in the other room. Which by the way, is good for workplace privacy.

        If you get a text message or phone call they’ll pop up on the desktop.

        As far as recompiling… Apple is a software company. Such concerns didn’t stop apple from switching Macs to Intel in 2005.

        And yes, making money is apple’s priority. Making good products does that. But so does market strategy.

  • Bruce_Mc

    Apple’s watch could appeal to collectors. People collect Nike shoes and Casio G-Shock watches, both priced at under $350.00. Making small batches of aluminum or steel cases with engraving, different finishes, or different colors might not be too difficult. It may take a few years of Apple selling all they can make of the current 3 watch lineup before they decide to add limited editions.

    Probabilism lets your competitors learn from your failures because they fail in the marketplace. Determinism keeps the failures secret because they never leave the company.

  • Stefan Popescu


    Fabulous analysis of Determinism vs Probabilism and the impact on how people and companies think and act based on that!

    That makes me rethink the whole question around why Apple is doing the watch in the first place.

    Fundamentally, Apple has been set by Steve Jobs to run based on deterministic methods but that was done essentially by drastically reducing the number of probabilities through focusing on just a few things. As such the number of probable outcomes is artificially constrained to a very low number. This may be the key in understanding how the deterministic approach would be applied and it works in business. In essence, and by following your logic, the deterministic approach would be a particular case of probabilism.

    From a fundamental physical standpoint, the phone product in general is a product that is suited very well to the deterministic approach. The device is sturdy and resilient enough and it’s supposed to interact inside universe that’s entirely man made, tightly controlled and deterministic in itself: connections, networks, apps, user interaction, etc. So no wonder it’s been and it is so successful. Everything around it it’s deterministic.

    You can probably say the same thing about the PC and the tablet.

    Your whole theory of Probabilism vs Determinism in business basically explains why Apple as a company is so successful.

    But the watch is a different story. Trying to understand the logic of Apple doing this move, two things come to mind:

    One: the watch is a moat. It has no major influence on Apple’s portfolio other than keeping the competitor’s away from the Apple ecosystem. Apple basically wants to make sure Apple user won’t wear a Microsoft or a Samsung watch therefore leaving the door open for these competitors to create nests of opportunity into its own ecosystem. That includes bands and watches and alikes.

    Two: the watch is a fundamental shift from the jobsian approach towards probabilism. That based on the premise that the main job for the watch is to constantly interact with the human body and constantly capture health data.

    From a fundamental physical standpoint, unlike the phone, the watch is a device under two constant physical major stresses: mechanical actions (the watch worn on a wrist is supposed to be banged, smashed, scratched – much more that a phone would be) and fluid actions (sweat, water splashes from washing hands, taking showers, swimming, etc). The existing watches have come to a point where they can sustain these types of physical stress and still do their jobs very well. But none of the watches out there are collecting health data from the human body. The way the current sensor technology works for wrist devices requires accurate measurement of electrical micro-currents from the skin, temperature, color, chemical variations, etc.

    This type of environment is highly probabilistic and if you add the possibility of fluids interacting with the sensors (I am assuming that the device core in itself is waterproof and immune to fluids).

    So what is that going to give us? Unless the sensor technology becomes 99% immune to such disturbances, the watch may remain a niche device with limited applicability. The major problem is that for the health application, a large number of data points is required. If the sensor is not accurate even for a limited period of time, then the whole data collected becomes useless.

    Unless there is something else that I can’t discern here and based on the above, I believe the watch will ultimately be of lesser importance that the Ipod was.

    I am also thinking that a third possibility is that the watch will be a “bridge” product in the same way the Ipod helped Apple ultimately create the the Iphone. But this is pure speculation …

  • Bruce_Mc

    One more thing on collecting Apple products. Beats is now selling Hello Kitty headphones.

  • santoscork

    On planned obsolescence – I have to think Apple does indeed engage in this strategically. Let me give you a few examples, mac mini, is crippled, not even RAM is upgradeable, non-5K mac, same, RAM capped and soldered. Why would Apple do this? Another example; the Retina MacBook Pro, a professional computer, users today wish they had 16GB and later on they will wish they had 32GB RAM. The MBP has processors that will still be capable 2 or 3 years down the road but with RAM being capped clearly has my blessing on what dictates planned obsolescence.

    It’s a pity that this is the case but this will certainly have people upgrading their machines not by adding RAM, that’s not an option, it will require a whole new machine.

    For all Apple is made up to be, this is one area that simply deserves an explanation by Ive’s design team and extended engineering teams. I believe this is marketing driven.

    As a disclaimer, I have been exclusively using Macintosh since 1987. My first Mac was a Iix.

    My comments are regrettable because I think Apple can and should build more expandable products; in the process delivering much more sustainable devices that can render a longer useful life for the user. Now I know that overtime, products get replaced and their energy footprints do drop, new products are just more energy efficient;. comparing the first iMac to the current iMac is a strong example. One might argue that it’s more environmentally sound to replace the machine when more RAM is otherwise required because a newer machine uses less energy so the energy offset will make up etc.

    I encourage any journalist who might have an op to interview Cook, to ask him about these shortcomings and for an explanation on this strategy of deliberately removing the option(s) of upgrading basic components like RAM. What is the strategy?

    • Walt French

      Not to be too Pollyanna-ish on this, but EVERY issue you raise can be due to OTHER factors than planned obsolescence.

      I’m writing tonight on a 2010 17″ MBP. Yes, it’s limited to 4GB of RAM, and its CPU speed is matched by the iPhone6+ in my pocket. But (after replacing its optical drive with a SSD, and upgrading its spinner to a 1.5TB disk), what makes me yearn for a new machine is mostly that it lacks BTLE for Continuity and the Retina display that allows more pixels on a much smaller screen. I’ve switched to bicycling for my commute and 17″, nice as it is for my senior-discount eyes, isn’t as good as the new products.

      This was (almost) top of the line in 2010 and I’ve replace the hard drive twice, and put in a third-party SSD twice. (The original failed but OWC is good to their warranty and Time Capsule saved my bacon.)

      I don’t see how Apple, in 2009 when they designed this beast, could have foreseen Intel’s product map (not that any laptop I know of in the past 5 years has a socketed CPU), GPU advances, software needs for RAM, BlueTooth, Thunderbolt, USB/3, etc etc. Laptops are a big part of Apple’s X86 mix and it would’ve been virtually impossible and/or hugely expensive to provide for expansion. (Per AnandTech, the iPhone6 has a 128-bit path between the CPU and RAM, unheard of in desktops of this machine’s era, and providing for a 32GB memory space would’ve required a costly 3—or 4—extra address traces to address, also one of the classic cost trade-offs.)

      Newer machines have the same tradeoffs. The ability to replace the CPU is costly, as is the likelihood that a casual user improperly uses too much thermal paste, frying the CPU. Expansion RAM requires more connections. Etc.

      Meanwhile, there is a robust market for used Macs on eBay. Buy a new machine; copy your files to it; wipe the old disk and re-install an appropriate OS, then sell. Done.

      The Olde Days invited lots of user fiddling, as hardware was fabulously expensive. The 2014 days invite zero user fiddling because your time is much more valuable than the hardware. Buy hardware for perhaps a 3–5 year horizon, and know that you didn’t have to buy cutting-edge tech at 5X its replacement cost. And, it was easy and environmentally sensible since the replacement tech used so much less power and other resources because you didn’t buy unproven, early tech.

      Yes, all this works in favor of Apple selling you new machines instead of some third party selling upgrade chips/boards/disks. As it happens, that’s in your favor, almost always.

      • santoscork

        Hi Walt

        I agree on many things you are saying. With respect to socketed RAM in particular, this is really the most relevant point I was aiming for. Firstly, I don’t mean to bash Apple. There is no angst I have, they deliver a superb experience, one I’ve relished for decades.

        When I take a look at the decisions engineers at Apple take on the topic of RAM, it seems irrefutable to me that they know exactly what it means to take away RAM upgrade options. They are necessary and are required for longevity and usability of the hardware longterm.

        That being said, it’s entirely possible that their software roadmap doesn’t necessitate more RAM. Obviously, the new RAM management schemes in 10.10 coupled with SSDs may well reduce the effects of less RAM but if an OS upgrade is ever restricted due to a RAM requirement not being met then the engineering team will have conspired to make otherwise upgradeable machines not be eligible for the same and that to me is irresponsible on Apple’s part especially when most hardware today is 64bit from top to bottom.

        That’s my issue. Of course a mac does not loose its usefulness once it is no longer eligible for an OS upgrade but it begins it’s inevitable march towards obsolesce earlier than it would otherwise should.

        Part of the problem is the word planned in ‘planned obsolesce’, this term engenders conspiracy — I just see it as ill planning.

    • reliability

      Most people never replaced the RAM, yet they had to pay the cost of decreased reliability due to socketing.

  • Secular_Investor

    It is worth noting that

    * most profits in the mobile handsets industry are generated in the high end sector

    * Samsung’s catastrophic profits collapse began even BEFORE the full impact of the iPhone 6 and 6+ was felt on Samsung Galaxy phones sales. This can be seen in the US where buyers evidently held back from buying Galaxies in anticipation of the launch of the iPhone 6. According to Cannaccord Genuity the new Galaxy S5 only held top selling spot on the 4 major US carrier for JUST TWO MONTHS following launch – May & June 2014. Thereafter the iPhone 5s took back its top selling position on all four carriers in July and August 2014 – despite the fact that the iPhone 5S was some 10 months old, and most buyers would have known that the iPhone 6 was about to be launched.

    * Since the launch of iPhone 6 and 6+ things appear to have gone from bad to worse for Samsung. In September & October 2104, Apple took the TOP THREE SELLING POSITIONS on ALL FOUR CARRIERS (No. 1: iPhone 6, No. 2: iPhone 6+, No. 3: iPhone 6+).I cannot recall Samsung ever not talking any of the top 3 positions on all 4 carriers.

    * Further confirmation of the Samsung’s increasingly dire position now comes from the new UBS four country survey of buying intentions of 4,000 smartphone owners, which shows that an amazing 30% of those intending to buy iPhone are coming from users intending to switch from Samsung.

    This is going to greatly accelerate Samsung’s horrendous profits decline – in fact their smartphones are likely to start causing large losses as their lower end phones also lose share in the very low profit low end segment.

    Also with this collapse of sales and profits it is unlikely that Samsung will be able to afford to maintain its huge $14 billion marketing, advertising and promotions budget, which could lead to even more downward pressure on sales. Add to this the loss of work from their largest customer Apple, and it seems that Samsung is about to join the ranks of Blackberry, Motorola, Nokia etc.

    Interestingly, the UBS survey also shows iPhones gaining new users from just about all their competitors’ whose users intend to switch to Apple: 4% of those intending to buy iPhones are Huawi owners, 4% Nokia, 3% Xiaomi, 2% LG, 2% Sony 2%, 2% HTC, 2% Motorola and 7% from others

    Interestingly, only 44% of those intending to buy iPhones are Apple users, indicating a huge gain of market share for Apple with an incredible 56% of those intending to buy iPhones switching from competitors’ brands.

    Another very significant finding of the UBS survey shows that Apple remains by far the most “sticky” smartphone brand, with 84% of Apple users responding that they intend to stick with Apple, compared to only 55% of Samsung owners, the most of any of Apple’s competitors, but indicating that fully 45% are unsatisfied with Samsung and intending to switch away. This foretells a huge loss of market share of Samsung.

    The survey is also bad news for Windows, with only 49% intending to stay with them i.e. 51% intending to switch away.

    The survey also shows Xiaomi also has very low customer loyalty with only 30% intending to stick with them i.e. 70% of owner intend to switch away.

  • Great podcast Horace! There seem to be big advantages to the deterministic approach:

    (1) You still get the benefit of product funneling/iteration (the funneling/iteration is just internal, instead of funneling/iterating through multiple product releases that use paying customers as test dummies).

    (2) Fewer released products mean the company’s efforts/talent are focused on fewer things, which helps the company create better products. An obvious point.

    (3) You’re in a better position to support the product and the customer over the long term.

    Those three things help build a brand and create customer loyalty.

    The probabilistic approach seems to ignore the customer — products that don’t work well get dumped, leaving the customer with a discontinued product. This can have a big negative impact on tech products, since tech products require vendor support.

    A probabilistic approach with products sold to consumers — where user experience and customer loyalty are so important — makes little sense to me.