Playing To Win
The Magnificent Seven Stocks
Strategy Imperatives & Lessons
A client of mine, CEO of one of the world’s largest and best-performing pension funds, asked me a question last week about ‘The Magnificent Seven’ stocks that are described as such because they have powered and are continuing to drive healthy gains in the S&P 500. I thought my answer was worth sharing and elaborating on in this week’s post, which I call The Magnificent Seven Stocks: Strategy Imperatives & Lessons.
The Magnificent Seven Question
The client was not referring to the 1960 western (which was remade in 2016) but rather to the stocks of seven companies, Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla, which have been supercharging the performance of the equity markets lately. The S&P 500 had a great year in 2023, advancing 24%. But that paled in comparison to The Magnificent Seven (M7), which surged ahead 107%.
The client’s question related to the doctrine of diversification, which has always been a winning investment approach. But when so much of the market return comes from such a tiny number of stocks, have things have changed fundamentally? He wanted to know my view on whether the M7 was just transitory like the Nifty Fifty of the 1960s and 1970s, many of which, such as Polaroid, DEC, and Kmart, ceased to be very nifty.
My answer is that the M7 is a bit of both — partially fundamental and partially transitory.
The Strategy Imperative
In my view, the biggest change in business in past half century is the shift away from variable costs in the cost structure and toward more fixed costs, about which I wrote in a 2013 Harvard Business Review article,Rethinking the Decision Factory. In the article, I looked back fifty years (1962–2012) at the breakdown between cost of goods sold (COGS) and selling, general, and administrative expenses (SG&A) at Dow Jones 30 (DJ30)companies. COGS was my proxy for variable costs — the costs a company incurs as it produces another unit of its product/service. SG&A was my proxy for fixed costs because they are the categories of costs that don’t vary with producing and selling another unit.
The article pointed out that in 1972, COGS was 72% of revenue and SG&A was 13%. In the late 1970s SG&A began to grow as a proportion of revenue and soon thereafter COGS began to fall. By 2012, COGS were down to 51% and SG&A up to 24%. Indeed, as businesses have scaled, there has been an inexorable rise in the proportion of fixed costs, especially in areas such as R&D, advertising and IT.
This shift from variable to fixed costs fundamentally changes the strategic imperative. When a cost structure contains a high proportion of fixed costs, there is an imperative to scale. Low scale (relative to your competitors) in a high-fixed-cost business results in an uncompetitive cost position, even if you are pursuing a differentiation strategy. You can handle it in the short term by running with low/no profits, but in due course, it will be deadly.
In What is Digital Strategy Anyway? (a Playing to Win/Practitioner Insights Year I piece), I wrote about the enthusiasm for firms whose offering is entirely digital — like Facebook — and whose distribution is customer self-serve. It is because the marginal variable cost of production and sale of another unit of the product/service is close to zero. The closer you are to that model, the more the prescription for you is ‘Blitzscaling,’ to scale as fast as possible, which will spread your fixed costs faster than any competitor, existing or potential. It is perhaps the most popular concept in Silicon Valley — maybe other than OKRs — and the name of a best-selling Reid Hoffman book.
It is always interesting to see when old becomes new again. The true father of ‘blitzscaling’ is Boston Consulting Group founder, Bruce Henderson, who argued in the early 1960s that a company should price ahead of its learning curve to achieve a cumulative scale advantage so that it will always have a cost advantage over competition. It will hurt profitability in the short term, but it will set the company up for luxurious profitability later — exactly Jeff Bezos’ approach a half-century later at the helm of Amazon.
Great Variability
Each of the M7 is digital in some fashion, but the variability in variable/fixed ratio across M7 is huge. (I used the most recent year in Yahoo Finance for the COGS ratios below.) A lower variable cost percentage means that the company has a business model in which costs are more fixed and the leverage from scale is greater. A higher percentage, of course, means there is less leverage from scale. By this measure, the members of the M7 could hardly be more varied:
For the biggest portion of its business Amazon must buy the item for you and ship it to you — all huge variable costs. That is why Amazon loves Amazon Marketplace, its business in which a third-party seller handles product procurement and delivery (or pays Amazon for the latter).
Tesla’s variable costs are also high because it must piece together a whole physical car and get it to you. That is why, even though automotive is the world’s biggest industry, it is still highly fragmented with its biggest two companies, Volkswagen and Toyota, having just over 10% revenue market share. The high variable cost structure has dramatically slowed the consolidation seen in so many other industries — despite lots of M&A activity.
Apple is next highest because it must put together iPhones (and computers) and physically get them to you. But it has a gigantic service business (mainly the App Store) to counterbalance the products business. Its service business has a variable cost ratio of only 29%, while product is 64%.
Because its offering is digital and largely delivered self-serve, most would guess Alphabet would have a very favorable variable cost ratio. But that is not the case, and a big reason is that it pays so much in Traffic Acquisition Costs — 17 cents per dollar of revenue — plus big, though not disclosed separately, content costs for YouTube and other Google platform content. Even though when it sells another dollar of revenue, one would think the marginal variable cost would by low, it isn’t. It is 43 cents.
At 31%, Microsoft is an impressive new-economy machine. This is a big reason why its market cap is so stratospheric. The leverage it gets from increasing scale is impressive.
Nvidia is arguably even more impressive. Its business requires it to make a chip to sell it. Yet its variable costs are only 27% of revenue, partially because sales are on fire, bringing down the variable ratio from 38% over the previous three years.
But most striking on this front is Meta at 19% — selling nearly 100% digital product delivered self-serve. It has the greatest leverage from increased scale of any of the M7, by a healthy margin.
Thus, there is great variability across the M7 in terms of strategic economics.
Where-to-Play/How-to-Win (WTP/HTW)
Leverage by way of digitally driven low variable costs is not the only factor at play. And that is why blitzscaling is not a substitute for strategy (just as I have previously and controversially argued that OKRs are not a substitute for strategy). You can sell your digital product below cost, build scale, and lower your scale-driven costs for a while. But eventually customers will decide whether your offering is compelling — or not — versus competition and/or substitutes. Is it differentiated from competitors in a way that is valuable to them? Or is it as compelling as competitors, but you can produce it at lower cost? For many tech start-ups, these questions aren’t asked because they are too busy blitzscaling — into oblivion.
On this front, Meta has a less rosy story than its low variable cost ratio would suggest. For better or worse, TikTok’s WTP/HTW is superior to Meta’s. TikTok has pulled even with Facebook in minutes used by US adults in 2024, has been ahead of Instagram since 2022, and will be substantially ahead of Facebook by 2025.
The scary thing to understand about companies with low variable costs (like Meta) is that while when they grow, the growth drops directly and substantially to the bottom line, when they shrink the opposite happens — the shrinkage falls directly and substantially to the same bottom line. Hence, the imperative for these high operating leverage businesses is growth — and that means that they need a powerful WTP/HTW that won’t be trumped.
But two M7 are being trumped. Meta is by TikTok. And Tesla is in the process of being trumped by BYD, which has already seized global EV market share leadership from Tesla despite operating in far fewer markets — yet!
And it is not a great thing for three of the other seven (Amazon, Google, Microsoft) who are going head-to-head in what is arguably the best business for two of them (Amazon/AWS and Microsoft/Azure). None of the three is either used to or particularly good at going head-to-head with well-resourced and powerful competitors, which each is doing now. As the cloud services market continues to slow — and it will — competition is going to be brutal because price competition in low-variable-cost businesses is particularly intense. Just ask the airlines. The variable costs associated with getting one more passenger into a seat on a plane is nearly zero, which is why steep discounting is so pervasive.
As of right now, I like the WTP/HTW strategies of Apple and Nvidia best of the M7. Apple has stuck assiduously to unique WTPs and has built HTWs that are perfectly matched. And Nvidia appears to have made a number of choices over the past number of years that are paying off now in a big way — and as it powers ahead, its competitors, such as they are, are struggling to catch up.
I hasten to add, this is not a buy recommendation. If Nvidia’s stock price has been bid up so much that its current price exceeds the discounted cash flows that its great strategy will generate, then it is terrible buy with a terrific strategy.
Practitioner Insights
Be wary of constructs like M7. Groups of companies are rarely as homogenous as they seem. The Nifty Fifty had Polaroid, DEC, and Kmart, but it also had Coca Cola, Pfizer and P&G. Lists like this never last forever — including the DJ30, which changes at an average rate of one company per year.
I gave the most pertinent advice on this issue in a Medium piece which I published before I started the PTW/PI series. It was called Avoiding the Pitfall of Our Same-Different Impulse. I was learning the platform at the time and didn’t realize that a 16-minute read would be too long. So, I later created eight-minute version and published it as Battling Our Same-Different Instinct during the short hiatus I took between Years I and II of PTW/PI. I didn’t make it a part of the PTW/PI series and can’t remember why — I should have.
It is well worth the read but my summary advice in it, and applicable here, is whenever you look at a set of things and your first reaction is same, immediately ask yourself: what is different? And if your initial reaction is different, ask: what is same? Appreciating both samenesses and differences will deepen your understanding of any phenomenon.
On that note, there are important elements of sameness in M7, and it is valuable to understand them to imagine how those elements will produce a similar future. But there are also important elements of difference and as they play out, I predict M7 won’t feel as tightly clustered as now — in fact, we will stop calling them M7.
