Subscribe
Share
Share
Embed
The Diff is a newsletter exploring the technologies, companies, and trends that are making the future high-variance. Posts range from in-depth company profiles, applied financial theory, strategy breakdowns and macroeconomics.
The Beloved Regional Grocery Store Chain Model
The grocery industry has gotten more concentrated over the last few decades, with the top four firms going from ~14% market share in 1990 to ~33% today. Weaker chains and independent stores have died off, while larger chains have grown either by adding stores and sales per store or, in the case of Walmart, Amazon, and Target, introducing groceries after achieving meaningful share in general merchandise.
One notable category of survivors is the beloved regional chain, a pattern that shows up in a couple places, like H-E-B in Texas or Heinen's in Ohio. These chains tend to be confined to a single state or a handful of states, and to have a general solid reputation for quality, affordability, and reliability. There are enough of these that the landscape of popular chains which locals swear by and which nobody else has ever heard of is material for viral tweets:
https://twitter.com/EudaimoniaEsq/status/1550328661626216456
The persistence of these companies raises some interesting questions about the optimal scale of a grocery chain. A single location runs some un-hedgeable risks, but larger chains are harder to manage. Larger stores do get the benefit of scale, and scale in groceries comes into play in many forms:
The grocery business is partly one of running a complex logistics network in which many products are necessarily just-in-time. Shifts in American eating habits are often driven by shifts in technology ("Iceberg lettuce" has that name because the way it was transported was in train cars filled with ice.)
Grocery buyers evaluating emerging food brands will sometimes ask for a free sample to test how well a product sells. A smaller store can't necessarily order in economical quantities, but a large store can get a few ~100% gross margin products every time it goes to a trade show. This isn't a scalable approach, but it does mean that the small stores pay to experiment and the big stores get paid to experiment.
Similarly, larger stores can negotiate bigger discounts from suppliers, sometimes forcing them to contribute to the grocer's own marketing. Since dollars are fungible, this spending is exactly equivalent to a discount.
Bigger stores can afford the fixed cost for third-party data, and for the infrastructure necessary to collect and analyze their own data.
A grocery store, even standalone, is actually a diversified company with multiple lines of business. There are different product departments with different margins and turnovers, of course, but there are also completely different kinds of operations. A Heinen's executive describes grocery stores as a combination of retail and manufacturing; for some goods, the extent of their processing is removing them from bulk packaging and putting them on the shelves, but for prepared foods they're also cooking. And sometimes the wasted output from one business line is an input into the next one: ugly vegetables that don't sell at the end of one day can become soup for the salad bar the next, and cut fruit can be a smoothie. Bigger chains can implement more rigorous measurements, and can ensure that their accounting reflects reality.
So there are good reasons to view groceries as a business that will ultimately be dominated by the largest-scale firms. There are, however, some benefits to small scale. Back when Walmart was a smaller chain, a big part of Sam Walton's job was visiting stores one at a time, looking for interesting new marketing techniques, and scouting out evidence of waste or sloppiness. (On weekends, he'd unwind by visiting competitors.)
The CEO of a bigger company has more responsibilities and less time, and also has more stores to visit. And this is one of the subtle advantages of being a multi-generational family firm. Passing a company down within a family always runs the risk of putting the company under the control of someone incompetent (The Diff has previously noted that Ferrari has a member of the founding family running their classics certification program). A grocery store's model can go wrong in many ways, but one of them is to let relentless cost-cutting start to impact the quality of the food they sell. And this is the kind of quality assurance that family-run companies excel at: TKTK
So H-E-B can operate with a third-generation family member as the CEO. And the Heinen family made the prudent diversification move to have twins, Jeff and Tom Heinen, who are co-CEOs. These aren't necessarily the best teams to own a high-growth retailer—though that doesn't mean zero growth, as H-E-B last week opened three new locations totalling 342,000 square feet in one day, wiht more planned.
Family ownership makes companies risk-averse, and if they're not levering up to take financial risks, that means focusing on operational risk. H-E-B, for example, started hearing about Covid around the second week of January 2020, and by February 2nd they were reviewing their contingency plans from the last pandemic scare and making adjustments. One thing that's convenient for family-run businesses as institutions and bad for the family members running them is that it's psychologically hard for an Nth-generation manager to ignore existential risks. A CEO can take positive expected value bets that have a chance of going badly because a high-profile failure doesn't preclude other opportunities in the same vein, but a family-run company is risking ending a multigenerational string of successes.
Over time, these risk-averse local chains build brand equity; when there's a natural disaster, they have stronger reasons to try to stay open and to help with the recovery, and in better times they're continuously building loyalty. This ends up being a self-fulfilling prophecy; one reason so many Texans shop at H-E-B is that other chains don't want to bother competing with them ("There are lots of weak competitors out there that we just don’t need to go poke that bear, so we won’t.")
A grocery brand can work nationally—Whole Foods has good national name recognition in part because there's a decent level of turnover from one affluent neighborhood to a similarly affluent one in an entirely different city, and in part because there's aspirational affluence. The company offers participation in coastal elitism to residents of flyover country. But the grocery business is also a hyperlocal one, and always has been. A grocery chain usually succeeds one location at a time. The highest-ROI form of branding might be a store that's sufficiently well-known in a single region that national competitors in the same category are reluctant to enter their market, allowing them to keep sales per store even higher and to acquire customers with weak brand affinity and convert them into loyalists.
This is a comparatively boring and straightforward route to success, or, to put it more precisely, it's a model where the interesting decisions are interesting to grocery store obsessives and boring to everyone else. A profitable grocery store is a succession of good decisions around product mix, store layout, marketing, and optimizing selection so it drives foot traffic and then maximizes basket size once those feet are in the store. And that tends to happen one 1% price tweak or coupon campaign at a time. On the other hand, it's a repeatable model, and it's evidence that in some businesses, there are working models at many different scales.
Further reading: Michael Ruhlman's Grocery is the source for many of the Heinen's-specific anecdotes here, and is full of useful information on how the grocery business really works.
Disclosure: long AMZN.
Elsewhere
The Microcap Manipulation Boom Continues
The Diff has been periodically covering the phenomenon of tiny companies seeing giant price spikes, a trend that has slowed down a bit as the Nasdaq has started carefully scrutinizing IPOs from small companies in Asia. But the existing stock of thinly-traded microcaps has still led to opportunities for strange behavior. Huadi International Group, a maker of stainless steel products in China, went public in early 2021. In the last five trading days, the stock went up some 777% on news that they plan to enter the battery industry. Shares that had been trading at $30 at the start of November reached a high of $192.88 on Friday. And then the company used a preexisting shelf registration statement to place 1m shares at $25 with outside investors. Yesterday, shares were down 91% to $15.81.
One question this raises is: who's taking advantage of whom? The company might have planned to hype its stock in order to raise money, but it ended up raising money at a lower valuation than its share price before the spike, and the stock's response to the capital raise was to go even lower than that. Haudi has $25 million more cash than it used to (less whatever fee it paid its bank, which for a risky rush job like this may have been substantial). But it took less than a week for the entire market value pop from the battery announcement to be erased.
Persistence
Verdad has a good piece in their series on persistence in financial metrics, looking at how well future margins can be predicted based on current ones. Unlike growth, which mean-reverts fast—between the S-curves of growth businesses and the sine waves of cyclicals, that makes sense. Margins are much more of a long-term characteristic of a business: it's hard for software margins to compress to the level of grocery stores or vice-versa. But what's especially notable is that one of the predictors of returns is listed is the ratio of free cash flow to EBITDA. While backwards-looking signals aren't necessarily forward-looking, and the market is always getting more efficient, this is notable for two reasons:
In many sectors, investors still default to EBITDA as a rough measure of profitability. It strips out lots of noise, and makes it easier to evaluate a business rather than a financial structure. And cash flows tend to be lumpier. But companies that optimize for EBITDA may end up making suboptimal capital allocations on a cash-on-cash basis.
This implies that one of the most valuable things a company can do is to buy assets that depreciate faster on paper than their depreciation schedules imply, that avoid leverage, and that optimize their taxes as much as possible (EBITDA is closer to free cash flow when interest and taxes are zero).
Bad Hedges
The total return from buying the S&P and also buying 5% out-of-the-money put options to protect downside has been identical to the total return from just buying the index over the last year. This kind of strategy is one of many ways to get equity exposure without taking as much risk (the easier one is to allocate less money to equities). But it makes certain assumptions: buying options is generally expensive over time, which is another way of saying that options sellers collect a risk premium from periodically losing a lot of money in a crash. But short-term slightly-out-of-the-money hedges do worst at a time when the market is gradually dropping but never gets around to crashing. The market declines make investors more nervous, raising the cost of options, but those options don't necessarily pay off: looking at monthly returns, the S&P's one-month drawdowns in excess of of 5% this year have totalled 9.7%, while the current cost of buying one-month protection against a 5% decline is 80 basis points. Annualize that cost, and the insurance works out to costing 10.1% annually. There are few free lunches in finance: if you're hedging by making a bet that pays off in a crash, you're overpaying to hedge against a more gradual drop.
Repurposing Assets
When Covid impaired the value of lots of real estate assets, it also made those asset owners think hard about how much they were paying for and how much they were getting. The owner of a building is paying for 100% of the cost regardless of where occupancy goes, and even as (some) companies return to the office and (some) people relocate back to cities, they're rethinking what all that square footage can earn. AMC has started a deal with Zoom where they're renting out theater space during the day for meetings. AMC's fixed costs keep ticking even outside of peak movie viewing hours, so the marginal cost of this is low. Meanwhile, Zoom knows that one of the risks to its business is the demand for in-person rather than virtual meetings. A meeting at a theater is at least a partial compromise there.
The other notable thing about this deal is that AMC was hurt by Covid on the revenue side, but helped on the financing side, since it acquired an army of dedicated retail investors. They may have decided that their company is structurally long the pandemic and its second-order effects, and have found a way for this to affect their fundamentals as well as their stock price.
Comebacks
John Foley, founder and ex-CEO of Peloton, has gotten together with some other recently-departed early Peloton employees to start a new business offering custom rugs. They've raised $25 million, which is a lot for an early-stage DTC rug business, especially in the current funding environment. This funding comes from some early Peloton backers. A CEO whose company grows and then crashes offers a valuable but mixed signal: they obviously know something about increasing revenue, but may know less about managing expenses and investor expectations. Early investors tend to have a sense of what management is like in many different situations, and may be betting that either the previous company ran into a perfect storm or that it taught its founder an expensive but valuable lesson.