Why is an expert on disruptive technology “worried” about financial innovation?
Excerpt from "Chaos Is Hard to Predict”
Does technical innovation always end in displacement/ replacement? The professions being challenged include physicians, lawyers, consultants, and analysts. Algorithms and sensors could conceivably displace some subset. However, it’s not a certainty. One way to fend off automation displacement is to redefine and change the scope of the profession.
The classic example is from the birth of the Industrial Revolution. As machines replaced certain tasks, new jobs were created which required higher skills and hence edu- cation, leading to universal matriculation and eventually the popularity of higher education. Professionals need to “invent” new jobs for themselves as a means to keep disruption at bay.
In Apple's first 40 years it shipped 1,591,092,250 computers ((including Apple II, Mac, iPhone, iPad and iPod touch; excluding Apple Watch, Apple TV and other iPods. Includes Q1 shipments estimated at 63,597,000 Macs, iPhones and iPads)).
This shipment total is higher than any other computer company in its first 40 years. Actually there are no other PC makers that are 40 years old. One computer maker (IBM) is older but they only sold PCs for 24 years and what they still sell they don't sell in high numbers.
That does not make it the top seller in a given year. Looking at only the Mac, Apple's traditional form factor personal computer, Apple has only returned to the top 5 last year. Only if including the iPad it was the top computer vendor in 2011 and including iPhone, it was first in 2009.
After having a 40 year run and after selling more computers than all American and Japanese computer companies put together, how should we think about the next 40 years?
First, clearly Apple shifted from being a "computer company". It has already changed its name to exclude the word "Computer" but that has been interpreted as saying that it sells devices (which happen to also be computers.) The word "computer" is already archaic. We stopped using computers to compute in the 40s. We used them to make decisions, keep track of things, speed things up and then to communicate and then to entertain.
Devices, it seems, are what customers mainly use to do, well, everything. Computing has grown to encompass most activities we engage in. So is Apple then a device maker?
More broadly, irrespective of the name, is Apple at 40 still defined as a hardware company? Certainly Wall Street ((i.e. average people)) still think so. The conversations one hears clearly consider Apple's future to be tied to the sales volume of specific products. At one time it may have been the Mac, another, the iPod and now the iPhone. This thinking is again a legacy of where the company was but mainly it's a reflection of how revenues are reported.
We see a line item in the income statement that says "iPhone" and it has a big number next to it so we, of course, fixate on that big shiny number. But if we were to unpack that number and see why it's so big we would discover that buyers are buying something other than what financial statements report.
Buyers are buying experiences, I hear you say. Yes, but more than that. Much more. They are buying safety, comfort, convenience, simplicity, productivity. They are buying modules to combine with other modules that help to do something. Something better. They are buying hope, satisfaction, escape or vanity. In other words they don't know what they want but know when they have it. They trust products unseen which have a reputation of delivering that which they think they might like, without knowing what that is. By employing the product, and a whole lot more—effort, time, mental energy and repetitive actions—they learn to exploit a product to achieve satisfaction.
Products are enablers, as are experiences and software and ecosystems and services. They are modules used in a melee of activity to make the user feel they are making progress. And if that module is better than no module then it will command a price and if it's better than another module or modules then it will command a premium. And if most people agree then it will be popular with many. And only if all these are true will it create a big number on a report.
And so we must search for other names to call a company that delivers an enabler that may lead to progress. Crude categorization like the reporting of finances leads to self-deception and a loss of opportunity to understand. Firms are often victims of this self-deception because they start believing that customers buy the things they sell. They start to believe that what is on their financial reports is a reflection of the value they create. It's a simple mistake to make, but a mistake which leads to catastrophe. If its data is mis-categorized, by chasing numbers the company runs away from opportunity, leaving it to competitors otherwise unencumbered with knowledge of numbers.
Assuming Apple avoids mis-categorizing what it does, will it be a "solutions" or "services" or "brand" company? Is it, as I used to say, a "blockbuster manufacturing line"?
Yes, and still that's not all it could be. Nor is it enough to understand what will come.
My simple proposal is to think of Apple (and actually any company) as a customer creator. It creates and maintains customers. The more it creates, the more it prospers. The more customers it preserves the more it's likely to persevere. This measure of performance for a company is not easy to obtain. It's not a line item in any financial report.
The closest figure we have is that today Apple has one billion active devices in use. We're not told of the total users or total customers because Apple cannot count people or wallets as accurately as it can count active devices. But as imperfect as it is, this number gives us a way to get close to counting customers.
So we know that at least Apple has created a large number of customers. Knowing total devices sold and number in use means that we can be sure that 63% of all the computers Apple sold are still in use. This means that most customers are recent customers. That is not necessarily good news. New customers are not well understood. We don't know how these new customers might behave in the future.
Will they come back for more of whatever Apple makes? If we look at the history of competitors or technology products in general, chances are they won't. The industries Apple has been a part of are filled with one hit wonders. Companies which had parabolic trajectories and built no loyalties. Statistically speaking, success in Apple's markets is always followed by failure.
But what if we did not seek correlation but looked for causation? What if we ignore statistics about companies other than Apple and focus on statistics about Apple's actual customers? Rather than assume all companies are the same, what if we assume that some can be exceptional? What if we we look for the anomalies, not the clusters?
We might want to ask about the causes of success or failure for a single company. We might want to sample its specific customers. We might want to observe how Apple's own customers behaved in the past. We might want to use theories of brand value or theories of network effects or theories of software ecosystems.
These measures, rather than those of "comparable companies" might lead to different conclusions: That customers are loyal, that they will value the brand and that they are invested in a network. In other words, that they will buy from Apple again.
If we then factor in the size and persistence of Apple's audience (customer base) we might conclude that there are factors which will allow it to grow rather than shrink. My analysis based on public data shows that the customer base has been increasing steadily.
Considering this audience and the revenues generated, we can categorize the revenue by active device as follows.
Each bar represents that product's revenue per day in a given quarter (from Q4 2012 to Q4 2015). Each year is therefore represented by a different color. Dark green for 2012, light green for 2013, burgundy red for 2014 and dark red for 2015. For example, each iPhone in use was generating exactly $1/day in revenues (including services and accessories) during the first quarter 2013.
This allocates all of Apple's revenues to these products and by proxy to customers.
In summary, the "average revenue per Apple device per day" is currently $0.68. If we assume one unique user per device then Apple collects 68 cents from each customer each day. ((In reality the figure may be closer to $1, as there are likely to be more 1.2 to 1.5 devices per person.))
Combining the history of customer creation and customer preservation with the value obtained from each customer implies that the next 40 years will be about creating another large tranche of customers whose willingness to spend on whatever Apple creates will be relatively unchanged.
We don't know the limit. One billion was hard to imagine even one year ago. We might see two billion devices in short order. Perhaps not. Perhaps as we have a multitude of devices about us all day the number will become less meaningful. But if the number of middle class customers grows and as Apple keeps its products within their reach, there is no reason to think that there will be a reversal of the last 40 years in terms of customer acquisition.
This analysis contradicts the current perception of the company as reflected by its stock price. That price assumes the audience will wither ((With a P/E ratio of 11, without a change in revenue per user, the only justification would be the audience shrinking by more than 90% over the next 20 years.)) That is not insanity. It is what we could expect if Apple were unexceptional.
We could surrender to correlation and ignore causation but statistics show that such an approach rarely works.
-------
Incidentally, this and a lot more will be discussed at the next Subscribe conference.
How Apple is managed is one of its enduring mysteries. The idea that a company with $235 billion in sales is managed with a single P/L ((Profit and Loss statement is an income statement usually applied for a subset of a larger company which allows subdivisions to be managed individually.)) is fascinating in many ways. Not least of which is how it allocates resources. The fundamental question of which great idea gets to be funded and which great idea gets to be ignored is the core of every manager's dilemma. The Apple problem is at scale when each decision's consequences are so momentous. In the case of Apple there are so few projects that reach the market and their impact is so great that one wonders how they can be sure they are doing the right thing.
Conventionally, product development is filtered through a sieve of metrics, market sizing and impact on top/bottom income lines. These "financial" measures of success are considered prudent and optimized for return on equity (also known as the maximization of shareholder returns).
But this can be a toxic formula. The financial optimization algorithm always prioritizes the known over the unknown since the known can be measured and is assigned a quantum of value while the unknown is "discounted" with a steep hurdle rate, and assigned a near zero net present value. Thus the financial algorithm leads to promoting efficiency at the expense of creation. Efficiency may be the right priority when times are difficult and resources are scarce but creativity is the right priority in a time of plenty. And abundance is what being big is all about.
To allow for some creation large firms create divisions. Some divisions are tasked with "core products" which are measured by a set of firm metrics and some are tasked with "emerging products" which are given a set of loose metrics. This leads to obvious resentment and war between divisions when resources need allocating.
Ominously, the core divisions tend to always win. At the root of divisional power struggles are measurements--the roots of financial metrics. Managers fight with data. "You can't manage what you can't measure" they're taught. But if you have something that can be easily measured and something that is difficult to measure, won't the easily-measured be managed and the hard-to-measure be ignored?
There is a general principle at work here: Managing by measurement is fraught with the pitfall of measuring the wrong thing. Making sure that you're measuring the right thing becomes the value-adding role of the manager. A role that is increasingly being neglected.
The mass phenomenon of measuring the wrong thing because it's the easiest to measure is called "financialization". Financialization is the process by which finance and finances (rather than creation) determine company, individual and society's priorities. It comes about from an abundance of data that leads to fixation on what is observable to the detriment of awareness of hazards or obstacles or alternatives. This phenomenon is more likely when the speed of change increases and decision cycles shorten.
Financialization is creeping into all aspects of society and the extent to which it infects companies is the extent to which they suffer from early mortality.
So is Apple avoiding financialization? How can anyone avoid the tyranny of mis-optimization?
The unified P/L might be a clue. It obviates the need for divisional rivalry and power-based or financial politics. Without divisional P/L, management can be organized functionally with the obvious benefit of de-politicization. The singular P/L does create another problem however: the absence of an alternative resource allocation algorithm centralizes the decision. By centralizing decisions at the highest level, few decisions can be taken and that means each decision has to be right more often. We swap a distributed but financialized process for a central but capricious one.
So how is the central decision process made fair? What guides the allocation process?
I believe that Apple does not avoid quantified decision making. I believe the logic of their decisions is based on a nuanced quantification of the company's customer base. Opting to think about how many customers they have rather than how many things they can sell could offer the Apple manager a better decision tool, avoiding financialization and mis-optimization.
The guiding principle is putting the customer rather than the firm at the center of the decision process. The idea of this change of orientation extends to a re-definition of the purpose of the firm. Apple calls it its "North Star" and it places "making the best product" at that location.
The challenge is how to measure "goodness" of product. It's easy to overshoot "best" and become too good. What regulates greatness is another measurement: absorption of what you create. The true orientation (which Apple won't confess to) is absorbable greatness.
Absorbable greatness is measured through other proxies: customer creation, customer satisfaction, and customer retention. By adding and preserving great customers you can be sure that what you create is indeed the "best". Setting the goal, and indeed the purpose of the firm, as the creation and preservation of customers seems to be the unwritten constitution of Apple.
The idea that the purpose of the firm is to create and maintain customers is not new ((Customer creation is in contrast to the notion that the purpose of the firm is to maximize shareholder returns. The notion of customer creation as primary purpose was popularized by Peter Drucker. The notion of optimization through shareholder value creation was popularized by Jack Welch.)) but it is relatively rarely practiced. The reason is that the data is harder to obtain. The data that comes from sales is crisp and concrete. The data about customers is muddy and soft. In a world where spreadsheets are used as weapons, the crisper the data, the better the ammunition.
Clarity wins arguments but it may not be where salvation or truth can be found. Good judgement in a cloudy reality is a better compass than a clear road map to a fantasy world.
Optimizing around customer acquisition rather than equity returns leads to a new set of metrics. What would these metrics look like for Apple?
The company publicly offers three separate sets of quantity of customers and quality of customers.
In terms of quantity we have:
Number of iTunes accounts
Number of iCloud accounts
Number of active devices
In terms of quality of customers we have:
Average selling prices for devices (as a proxy for willingness to pay)
Customer satisfaction (as a proxy for loyalty )
Services and accessory revenues (post-sales and recurring value)
The following graph shows the historic data for service accounts and corresponding services revenues.
The following graph shows the average price per hardware product indicative of loyalty through stability in the face of growing competition.
The following graph shows an estimate of growth in active devices as a proxy of customer base.
Unlike product shipments and revenue data which follow rises and falls in individual products due to the cyclicality of technology and adoption, these graphs show a smoothness and predictability of customer base behavior.
Looking at Apple this way gives the impression that the company has a large number (nearly a billion) of repeat (subscribed to service) customers. These customers can be seen to spend a predictable recurring amount (average selling price) on Apple's brand. The repetition of these metrics in public commentary suggests that customer acquisition and retention are among Apple's goals.
The key question is whether this is Apple's primary goal.
Seen this way each centralized resource allocation question can be assumed to be prefaced with "In order to create/preserve customers should we...?"
This leads to answers quite different from questions that start with "In order to sell/profit more should we...?"
For instance, offering education products, retail store experiences, healthcare products, green energy initiatives, accessories, content and services make sense even if they may have poor financial "returns". These initiatives score highly on customer retention and satisfaction. They may lead to engagement and repeat sales. Managers working on these initiatives would gain the respect of their peers, leading to more esteem and higher quality output. Finally, and most importantly, these efforts would be preserved regardless of cyclical downturns.
The assessment of analysts should therefore be couched in similar terms: "What is the impact of a new initiative in customer creation and preservation?" rather than "What is the impact of a new initiative in driving revenues and profitability".
We can see immediately why the second question is popular while the first isn't: it's easier to assess an answer to the second because we have the data that allows for its calculation.
This is why asking the wrong questions is popular.
Horace and special guest Ben Bajarin of Creative Strategies discuss the Glance conference. They investigate the job of the Watch and touch on the changing responsibilities amongst the upper ranks at Apple.
Horace and Anders discuss the iPhone analog audio port, the lightning port's use by Apple Pencil and Apple's viewpoint on privacy and the environment. Horace rounds the episode out with a definition of disruption.
In my quoting of the "Cook Doctrine" I cited the primary criteria for Apple to enter a new market:
We believe that we need to own and control the primary technologies behind the products we make, and participate only in markets where we can make a significant contribution.
These criteria, often repeated, were certainly in force when Apple chose to enter the watch market. Apple has sought and achieved a significant market share and did so while owning and controlling the primary technologies behind the product.
I now turn to the significant contribution criterion to study the possibility of Apple's entry into the car industry. ((Control of the primary technologies behind the product is a topic for another post.))
The significance test shifts the speculation from whether Apple would build a car, to how many cars is could build. Making a few cars is easy (see first commandment of The Entrant's Guide to the Automotive Industry). Making lots of cars is hard and hence significance in the automotive market (as in watches and phones) means achieving some degree of adoption, a higher degree of usage and a very high degree of profitability.
So what does being significant in the car business mean? Does it mean becoming the next Tesla? The next BYD or the next VW? How quickly?
Fortunately, we have something to compare an Apple entry to. Apple has made a "significant" market entry in phones and others have made entries in cars. If we contrast the rate of growth of Tesla, EVs, and Hybrids ((Of which Toyota has 70% to 80% share)) to the rate of growth of iPhones in their respective US markets, we obtain a test of significance:
The graph shows the percent of sales for the alternative car technologies (and Tesla) vs. the percent of US phone users using iPhones (comScore). Here are the conclusions:
Hybrid cars were introduced in 2000 in the US. They managed to gain 2.5% within 8 years but have not increased to a consistently higher level since.
Electric vehicle sales re-sarted in earnest in 2008 and have gained 0.4% share in eight years. They are still growing, partly helped by significant federal and state subsidies.
Between sales start in 2008 and 2015 Tesla achieved 0.15% share.
The iPhone ushered in a new paradigm within the phone market ((Incidentally the number of phones in use and cars in use in the US is comparable.)) in mid 2007. In the time since, Apple achieved the equivalent of about 35% share of unit sales. ((Significantly more in terms of sales value and over 90% of profitability.))
Between 2002 and 2015 the share of smartphones as a percent of all phones reached over 80%.
To summarize, the range for entrant technology share capture spans between 0.4% (EVs between 2008 and 2015), 2.4% (Hybrids between 2000 and 2015) and 80% (Smartphones between 2002 and 2015).
Within a similar time frame, the range for entrant company share capture spanned between 0.15% (Tesla as percent of US car market) and 35% (Apple iPhone percent of US phone market). ((Toyota's capture of 2% of the market with hybrids may or may not be considered entrant--I chose to consider Toyota an incumbent which took share from others with the Prius))
The differences thus span two orders of magnitude (>100x). Put another way, if Tesla's car entry was equivalent to Apple's iPhone entry they would have delivered about 5.5 million cars rather than the 50.5k they delivered in 2015.
This wide gap in the impact of entry in the two markets has much to do with how modular are the systems being engaged by the entrants. An electric automotive system requires new infrastructure, new user behavior, and, in the case of Tesla, new distribution network and new post-sales support network. Offsetting this partially is a government subsidy.
In contrast, the smartphone communications and computing system leveraged existing infrastructure ((And vice-versa)), consistent behavior, at least initially, and the same distribution and support network. In addition, smartphones benefited from significant hardware and software and services ecosystems which encouraged third parties to add value to the system.
The difference in collaborative development and conforming user experiences (i.e. modularity) between the two innovations allowed the speed of change to be two orders of magnitude faster. The result is what I would consider significance for the iPhone vs. insignificance for electric vehicles to date.
Which leads us back to the question of why Apple would enter this market. If the absence of modularity prevents rapid change, that implies Apple cannot achieve any significant impact from entry.
We could ask the question differently: If they do enter, they could achieve significance only if they change the degree of modularity in the system. Think of it as allowing infrastructure leverage, consistent user behavior, wide distribution, wide support network, ecosystems, and component architecture; allowing rapid scale production and lower costs, etc.
Therefore the only way that Apple could justify entry would be if they "rationalized" the system into a sufficiently modular alternative. Only if we assume that change is possible could we then target share above the 0.1% insignificant change that has been demonstrated. They might even perhaps aim at 1% impact within a few years of entry. 1% was the iPhone target within the first full year of sales.
If system-wide change could be brought about and made uniquely (through some form of sustainable asymmetry) then the foreseeable share possibility quickly reaches the 10% range.
The last time that happened was when Toyota changed the way cars are designed and built. ((VW before Toyota and GM before VW. Incidentally Toyota, GM and VW are now nearly tied as the top three automakers globally.))
We consider the landscape that Apple's purported Titan project will address in a few years time.
Horace discusses the pattern of disruption powered by Moore's law. We turn to the transportation sector and consider the "reimagining" of the car. We further consider scenarios, from sustaining where the current players grow, to new entrant opportunities.
The conversation diverts a bit into the regulatory and taxation regime, specifically US road funding is largely tied to fuel taxes. We note the odd situation where an entry level car driver pays fuel taxes while a luxury Tesla driver does not.
We speculate on Apple's possible "meaningful contribution" to transportation and the required product, customer experience, sales channel, price and financing options.
Horace and Anders discuss viewing Apple's business as a recurring revenue model, then touch on the iPad Pro somewhat unsuccessfully and host a very special guest.
Apple Watch was released April 10th 2015. Eight months later we are holding the first Apple Watch conference. To kick off the discussion, here are my favorite myths about this new product:
