It's Hard to Build a Durable Business Selling Durable Goods

Plus! Wrappers and Layoffs; The Movies; Evaporating Sanctions; Bonds; Errors in Translation; Diff Jobs

Quick note to new readers: The Diff has an associated AngelList syndicate, open to accredited investors. We are close to launching a new deal, so if you'd like to read more about it please sign up here.

It's Hard to Build a Durable Business Selling Durable Goods

One of Warren Buffett's more regrettable one-liners comes from Barbarians at the Gate, when he rhapsodizes about the tobacco industry: "I'll tell you why I like the cigarette business. It cost a penny to make. Sell it for a dollar. It's addictive. And there’s fantastic brand loyalty." All true, but perhaps an observation best kept to oneself.[1]

You can easily invert this to imagine a similar quote from Dark Buffett: I'll tell you why I hate the high-quality consumer durables business (e.g. appliances, gadgets, furniture): you sell it for a dollar but it costs 85 cents to make, the better a job it does the less likely your customers are to need another one, and if you want brand loyalty commensurate with that quality you'd better wait a generation or two to prove that the products you sell can last for a long time—meanwhile one or two bad years can sink you.

For example, Instant Brands (maker of the Instant Pot) put out an announcement last month headlined "Instant Brands Takes Action to Strengthen Financial Position and Support Long-Term Growth." Put more plainly, they filed for Chapter 11 bankruptcy. (Filing here.) There are some good high-level explanations: Instant Brands notes that their supply chain was disrupted by Covid lockdowns in China in late 2021, and later by Russia's invasion of Ukraine. They also note that US consumption shifted away from goods and towards services. Unfortunately, the 89-page bankruptcy filing does not have enough room to mention that their private equity owner refinancing debt, borrowing more, and using some of the proceeds to pay themselves a $245m dividend.

There are a few ways to frame the questions this raises:

  1. Why is it a particularly bad idea for a consumer durables business to lever up?
  2. Why, if they're such a bad business, do people in the developed world have fairly easy and affordable access to dishwashers, air conditioners, toasters, and other durables?

For any of these categories, it's easy to find brands that have led to enduring companies. In dishwashers, for example, Whirlpool has been around for a long, long time; they were founded in 1911, and own plenty of durables brands (Maytag, Kitchenaid, Amana, etc.). GE was in the appliance business for over a century, but sold it to Haier in 2016. Smaller brands like Le Creuset have also persisted for a long time, despite selling very durable durables indeed, offering a lifetime warranty for non-business use of many of their products. And there are newer contenders, like the aforementioned Haier, as well as LG, Samsung, and Panasonic.

We can divide these companies into two categories rather nicely: there are companies that started a very long time ago, often before the category was defined (Whirlpool's founder started the company to investigate the hypothesis that electric motors would make dishwashers more effective), and there are newer companies that started well after the category had been defined, assembled their products in a country with low labor costs, and won share by achieving rough parity on quality and competing on cost.[2]

This is a helpful split, because we see that more recent consumer durables brands tend to have a more dramatic history. It’s not just Instant Brands’ Instant Pot—there's also Traeger (shares are down 80% since their 2021 IPO), Hamilton Beach (down by two thirds since going public in 2017), and GoPro (down almost 90% in their nine years as a public company). And even the ones that are doing reasonably well today aren't in quite as good shape as they looked a few years ago, back when Instant Brands' private equity owners and their lenders were feeling so optimistic. For example, Yeti has actually put up good numbers since it became publicly traded in 2018, with a total return of 159% since 2018—but it's down by more than half from its peak valuation in late 2021.

Many of these companies are priced at a discount to the market, reflecting the volatility of the business. Some aren't, especially those that sell durable products to businesses; Rational, a German appliance company focused on the commercial market, has a very trusted brand name, but it trades at 6.5x sales, so the market is well aware of this. Similarly, Mettler-Toledo has built a great measurement instrument business (they sell scales and other measurement instruments—classic durable products) and trades at 8x sales. But they've built up their brand name over generations; the book Why Japan Was Strong, written in 1943, talks about Japan's industrial espionage, and uses their corporate ancestor, Toledo Scale, as a case study of the kinds of companies Japanese manufacturers were copying.

There is a secret to running a good durables business! And that secret is to start early, ideally in a country where a) most households can't afford the product in question, and b) is getting richer by the year, such that the addressable market is growing the whole time. What kills many of these companies in their modern incarnation is that growth follows an S-curve, and the trouble with S-curves is that you can get a pretty good fit by applying an exponential function to the first half of the growth curve.

We can actually see this in action by looking at historical data on technology diffusion. Our World In Data has a wonderful dataset on this, with an accompanying chart:


Extracting some of the technologies that seem relevant to the case studies, we'll follow a straightforward process. First, we interpolate missing years, which is helpful for cases where we have missing/inconsistent time series data. Fortunately, diffusion of useful technologies is kind enough to follow this simplistic model quite closely, with r-squared values ranging from .88 (television) to .99 (flush toilets).


Next, we're going to look at replacement cycles. (Source: lots of Googling.) And we're going to look at household counts, via the US census. Changes in household count matter a lot in reference to replacement cycles. Consider the flush toilet. The average one lasts for 25 years, so if you flip that around roughly 4% of the installed base gets replaced every year. Which means that the difference between 1% annual growth in households and a flat household count is a 25% larger industry. That’s important because many of the products we're looking at took off at a time when US household count was growing fast, because of population growth and changing family structures. (Family structure took the lead after a while; the average household size in the US went from 3.3 in the early 1960s to 2.6 by the late 80s—from the appliance sector's perspective, that means a whole lot more toasters and coffee machines per capita!—and has since drifted down more slowly to about 2.5.)

What this gives us is a simple demographics- and market penetration-based model of organic demand over time. And what it tends to show is that from fairly early on, more than half of the organic demand for these products was from replacements, rather than from new ones. For example, central heating starts showing up in the data in the late 19th century, and by the 50s half of sales are from replacement; indoor plumbing hit 50% dependence on replacement in the early 40s; refrigerators, which weren’t include in data from the early 20s, already had most of their demand accounted for by replacement products by the mid-40s.[3]

There were several constraints on technology penetration in the early 20th century, the most important of which were 1) electrification, and 2) income. Electric power was in 20% of homes in 1915, 50% by 1925, and 80% in the mid-1940s. GE's affordable refrigerator model introduced in 1927 had a sticker price of just over $300, or around 1/3 of GDP per capita. In the 50s, when nominal GDP per capita had more than tripled, refrigerators' nominal prices were similar, at around $300. And today, with GDP per capita up more than 20x since the 1950s, listings for an apartment-appropriate refrigerator are still around $300.

So the appliance companies that started early had their growth naturally gated by the fact that many of their prospective customers couldn't afford their products, and many of the rest couldn't use them. Of course, those constraints went away, in part because a country that manufactures many useful appliances, and can export them or start foreign subsidiaries to amortize R&D costs, is a country that will produce sustained increases in GDP per capita and standards of living. Which meant that they slowly shifted from demand driven by growth to demand driven by replacements.

A new appliance company can't ride this gradual trend, at least in the developed world. Once their product starts working, the clock is ticking: if they try to expand slowly, someone else will clone the product and start taking market share. But if they grow at maximum speed, they'll inevitably reach a point where they've saturated demand, and they won't realize it until it's too late. Instant Brands accidentally calls this out in their bankruptcy announcement: "its iconic brands, found in approximately 90% of homes in the United States and in millions of other homes around the world..." That's good for bragging rights, but it also means that only one in ten households in the US hasn't yet bought one of these products.

Conversely, big appliance companies had the time to diversify; the air conditioner, refrigerator, and microwave are all ubiquitous today, but the start of their rollouts is separated by decades (1902 for the AC, 1923 for self-contained home refrigerators, 1947 for microwaves). Which tells us that older appliance companies could naturally diversify into new categories while their older ones were hitting a wall (and could reuse some components, as well as their marketing and distribution). So General Electric was, during the mid-20th century rise in American living standards and energy consumption, a much better business than Extremely Specific Electric would have been.

It’s possible that Instant Pot saw this demand cliff coming, but the signs were more subtle. Take a look at Google Trends:


A few years ago, there was the holiday search volume spike that's typical of any good that makes a decent gift: people start searching on Black Friday, and there's a frenzy of Googling in the last days before free shipping isn't an option. But what's striking in that graph is that in the first four years, peak search volume actually happens after the holiday season, because people are searching for new recipes. That's a strong indication of high usage. But the size of that post-holiday spike died down over time, and last year it was smaller than the shopping spike. This kind of indicator actually leads demand, since it reflects customer (or, really, gift-recipient) satisfaction with the product. But that feeds into demand if it creates a narrative that the Instant Pot has gone from a great gift to a redundant one.

There are still ways to escape the generally challenging economics of durable goods with an unknown peak market penetration. One popular option is to turn them into subscriptions or to pair them with complementary, and less durable goods; easier for something like exercise equipment, and challenging but not impossible for food. (Traeger, mentioned above, does some of this by selling wood pellets. But not enough to fulfill the market's expectations from their IPO.) Another option is to have a sufficiently diversified portfolio such that something new is taking off for every product in decline. Instant Pot did attempt to do this, but, perhaps in part because their dividend left them undercapitalized, wasn’t able to find their next hit in time.

Of course, retailers want these complementary durable goods to exist, and are willing to accept lower margins on some hit products simply because they expect a decent attach rate on other goods they sell. But they aren’t partial to specific brands. Which presents another obstacle to manufacturers getting recurring revenue: in this Tegus interview, a former marketer at Williams-Sonoma uses the Instant Pot as an example of a product that gives the retailer information they can use to market to that customer. Instant Brands isn't going to capture as much information—and Williams-Sonoma has a wider range of follow-up products to sell. So they'll implicitly subsidize companies like Instant Brands, up to a point, but ultimately they're fine with abandoning them; there will be another Instant Pot to send foot traffic to their stores and web traffic to their sites regardless.

But even if some of these problems are self-inflicted, the difficulty of modeling demand for durables in a developed economy is universal. It becomes an information-asymmetry problem: what customers are really doing is paying for a certain number of use-years for the product, and it may be the case that the highest-quality version is, on that basis, also the cheapest. But proving that a product will be useful for a decade actually takes a decade, so if the market can get to saturation within the useful life of the first iteration of the product, sellers will not be able to price it at quite what it’s worth. Having a product that nobody wants is a problem, but having a product that everybody wants is an only slightly less problematic, because failing to grow into that demand means ceding the profits of a new product to a more risk-tolerant fast-follower, while trying to capture all of them means eventually having warehouses full of unneeded pressure cookers that were trendy a few years ago—which is why so many customers bought them—which is why nobody needs another one.

  1. Weirdly enough, it’s not the only example of a future billionaire showing off by making an observation about cigarettes that they later regretted. ↩︎

  2. Whether they're "better and slightly cheaper" or "slightly worse but bizarrely cheaper" depends on the company and time period; there was a time within living memory when 'made in Japan' was pejorative; there was an urban legend for a while that Japanese manufacturers assembled some products in the city of Usa in the Ōita Prefecture in order to label them "MADE IN USA," but this is untrue. ↩︎

  3. One fair critique here is that the useful lives of these products were relatively short early on, because we didn't know how to manufacture them well, but improved later on as the products were perfected and quality standards went up, but have since declined as more companies compete on price rather than quality. The last bit isn't quite relevant to this model, but the first bit is worth thinking about: there's an opposite effect, which may actually be stronger—standards of living were much lower in the early days of technology deployment, so even if the cars of 1920 were fairly shoddy because we didn’t know how to manufacture them well, the car buyers of 1920 would be more likely to page through their Model T owner's manual and order some spare parts than they were to replace the whole car. ↩︎

Diff Jobs

Companies in the Diff network are actively looking for talent. A sampling of current open roles:

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.


Wrappers and Layoffs

Kitchen gadgets are not the only sector where companies have an incentive to grow too fast. Two AI companies, Jasper and Mutiny, have just started layoffs, in Jasper's case less than a year after their $125m Series A. One way to look at this category is that it exists in the gap between what models are capable of and which problems customers realize can be solved by AI. And there is a lot of room to create value in closing that gap—but there's also an incumbency advantage to doing it as an existing service provider who has one more tool to sell, instead of as a new business. The new company can grow fast, but in the best-case scenario its product roadmap is constantly getting shifted by changes in the capabilities of the product it monetizes, while in the worst case its suppliers will end up launching a tool that renders the intermediary obsolete.

The Movies

Harrison Ford and Tom Cruise are astounding monuments to the human body's ability to stave off entropy when there's enough money and fame on the line, and are also, of course, symptomatic of the movie industry's growing reliance on established stars and franchises. But after Indiana Jones had an opening weekend 40% weaker than the last iteration of the franchise and Mission Impossible also put up disappointing numbers, it's becoming apparent that the indefinite franchise extension model has some weaknesses. In a sense, this model is one of the biggest zero interest rate phenomena around: the lower rates get, the more valuable an indefinite stream of income from constant recycling of material becomes. (Especially if you can achieve Disney's generation-bridging strategy of getting parents to introduce their kids to franchises ($)—though the single best example of this is Lego stores where sub-$50 sets based on new Star Wars IP are at kids' eye levels while the $850 Millennium Falcon set is clearly positioned so nostalgic parents will see it.)

It's interesting that this phenomenon in legacy media coexists with an ever-faster content and meme cycle in new media. As in many other sectors, the middle ground of high production values and some baseline of originality has been hollowed out. The most interesting question here is whether this presents an opportunity for a new media company, or for a legacy one looking for a revival. When distribution is changing fast, the media winners are the ones who figure out the new distribution model, but when that matures, the advantage tilts back to the companies with better content. And, as Ben Thompson pointed out this morning, changes in distribution also lead to labor conflicts, another mechanism by which legacy companies can fall behind.

Evaporating Sanctions

The gap in exchange rates between physical foreign currency and electronic representations thereof has collapsed in Russia as the Russian economy adapts to sanctions. This happens both through changing consumption patterns (more trade and tourism with countries that accept rubles) and through more effective evasion of sanctions. Keeping the effect of sanctions fixed means constantly tweaking the implementation. (And another way to phrase this is that retaining sanctions over time tends to mean, in effect, gradually relaxing them.)


Citadel Securities is expanding into electronic bond trading ($, FT). (The Diff covered the shift from voice to automated trading last week (Diff, $).) One vision of liquidity is that it's all about creating a platform that connects buyers and sellers so they can efficiently transact without some middleman interposing itself into the trade. But another vision is that it's actually driven by exactly those middlemen, who get paid for providing liquidity. In the bond world, with ~100x as many products available to trade as the US equities market, the odds of any one buyer and seller wanting to trade the exact same product at the same precise instant are quite small, so partly- or fully-automated liquidity providers become a bigger catalyst for growth.

Errors in Translation

The Diff has noted a few times that American companies are very fortunate in that their home market is roughly a quarter of the global economy, so they can use a profitable home base to subsidize international expansion more effectively than a company in, say, the Netherlands (~1% of global GDP), which has to decide early whether it's global by default or a smaller-scale local business. The tradeoff, though, is that success in the US market trains companies with certain expectations that don't apply elsewhere. Bloomberg has a good piece on the difficulties Vanguard has faced expanding its model to Europe, where a decade and a half of local operations have gotten them just 4% of the firm's overall assets under management. Part of the problem is that Vanguard had a unique marketing pitch when it started in the US, and made its brand synonymous with index investing. But the idea of index investing spread globally before Vanguard itself could. It sometimes happens in the technology industry that a company effectively privatizes some intangible product that was previously owned collectively (like the social graph, the link graph, email, or protocols for broadcasting short mostly text-based messages in near real-time). When that happens, the protocol in question gets used more, because a profit-seeking entity can subsidize its expansion. Vanguard in Europe shows the inverse of this: when a strategy like indexing is just part of the broader toolkit rather than something being pushed to retail investors by one dominant company, the indexers are less profitable and indexing itself is less popular.