Longreads

• Jon Stokes as thoughts on AI content generation. This piece is the first in a planned series, and goes through some preliminaries, but they're very useful. Specifically, the framework that AI content generation is a search tool that searches the space of all possible results rather than all existing results explains a lot

• Patrick McKenzie has a great post on how banks think about where to locate bank branches, and I doubly recommend it if that sounds like a boring question. This post has it all: unit economics, the intersection of hyperlocal and national regulations, queuing theory, Chicago politics, and wisdom from real estate professionals. A bank branch is a monument to Hayekian thinking, the confluence of a bank's broad knowledge about its long-term lending and funding needs and a real estate developer's accumulated social capital.

• Mehr Nadeem in Rest of World writes about the $85m collapse of Airlift, the best-funded startup in Pakistan. Airlift was originally in transportation, but pivoted to rapid delivery during the pandemic. And rapid delivery companies seem to do better in developing markets: Getir, the Turkish delivery company, has raised $1.8bn, with its most recent round in March, and Jokr pulled out of the US to focus on Latin America. One easy-to-forget advantage of operating in a country with a large startup ecosystem is that there are lots of people who have experienced hypergrowth. In Pakistan, it was harder for the company to find employees with that kind of experience, which both meant that there was a mismatch between the pace at which the CEO wanted to move and the pace at which the company was capable of moving, and that there weren't enough people who'd had the experience of seeing a high growth company fall apart and who knew the warning signs.

• Michael Mauboussin at Morgan Stanley has a good piece on how market share affects companies' ability to earn higher margins, with, true to his usual practice, a nice twist halfway through where once you adjust for the cost of investing in intangibles, much of this perceived advantage goes away. One way to view this is that the rise of Big Tech as a share of market values is partly an accounting anomaly: they're really asset-heavy companies whose assets happen to be omitted from balance sheets. On the other hand, they've created a lot of value; they can't just be making money from piling up lots of assets if the mature ones don't need to raise capital and are often running buybacks. So one analogy might be to railroads: they do require a fair amount of money to get started, that money is unrecoverable if the demand isn't there, but if they connect the right points—whether it's searchers and advertisers or two big populous cities—the rewards are huge. (And that's a comforting thought for another reason: while tech has grown as a share of market cap, railroads basically were the US capital market in the late 19th century, with every other industry as a rounding error.)

• This 2018 piece by Reid DeRamus on understanding churn rates is a valuable one. The key point it makes is that churn is an aggregate number, but we really care about how individual cohorts are performing. A customer's churn rate is highest right after they sign up (for a product like a newsletter, there's a non-negligible level of same-day churn, from people who signed up to read a single story). You can look at the consequence through two scenarios: first, when a company's growth slows down, its churn numbers will improve as the average age of user accounts goes up. The users they have are, increasingly, the ones who will stick around. And conversely, when a company's growth accelerates, churn accelerates, too. Netflix this year is a great example: the best way to have a quarter of negative growth is to have a year of extraordinary growth first so your average subscriber age ticks down.

Books

• Money Games: The Inside Story of How American Dealmakers Saved Korea's Most Iconic Bank: In 1997, Asia's high-growth economies struggled through a financial crisis. This book is about distressed investing in the aftermath, specifically negotiations by a private equity firm to acquire a controlling stake in a Korean bank and restructure it. The deal was ultimately successful, and the book spends most of its time on the negotiations. For more than half the book, I thought each new chapter would start with "And so we'd finally closed the deal," but instead it starts with something like "But then, I got a disturbing fax—we had to start the negotiations over." This book is a great look at an unusually complex M&A deal, and at how governments deal with banks' quasi-public, quasi-private status.

• The Great A&P and the Struggle for Small Business in America: Before there was Amazon, there was Walmart, and before there was Walmart, there was The Great Atlantic & Pacific Tea Company. A&P once had 10% share of the grocery business at a time when food was almost a quarter of household expenditures. This book, by the author of The Box (also strongly recommended!) goes into detail on how A&P grew. It's also a story of constant pivots; they're descended from a leather company, then switched to tea, then to groceries—and in the grocery business, they moved from upscale establishments to small, cheap stores to supermarkets. One thing that stands out in the book is the continuous internal debate in the company over whether to focus on margins or return on investment. A larger store will usually have lower margins, but its turnover will be so much higher that it's more profitable. So the book is also a good history of how the US economy has broadly gotten better at identifying the right scale on which to operate. That's a process with winners and losers, and A&P wound up one of the losers, and was liquidated after its second bankruptcy. A great view of just how hard retail is, and at how many new debates over market power have been going on for a century or more.
  
  Disclosure: Long Amazon.

Open Thread

• Drop in any links or comments of interest to Diff readers.

• I've been doing more reading recently on the grocery industry, especially a) how grocery stores control costs, b) their data operations, and c) slotting fees and other not-from-selling-things-to-customers fees, and how those have evolved over time. Anyone with suggested reading, or who'd like to talk about the subject, is encouraged to get in touch!

Leave a comment

Reader Feedback

In a comment on last week's post on replacing one-off sales with subscriptions, Michael He asks why Nike doesn't have a subscription plan. This seems like an ideal case: Nike gets a fixed revenue stream, subscribers might get discounts and guaranteed access to some limited drops. And the broader question is: who else should be considering a shift to subscription, and why haven't they done it yet?

Leave a comment

A Word From Our Sponsors

Leveraging AI to drive deeper insights

Stream by AlphaSense provides investment research teams and analysts with unique, high-quality expert insights on demand. Sign up for a free 14-day trial here to make faster and more cost-effective investment decisions with AI technology that understands your sentiments.

With Stream, you’ll have instant access to:

• 20,000+ on-demand transcripts of expert interviews conducted by experienced buy-side analysts. Get the operator’s perspective, first-hand, on the companies and sectors that matter to you.

• The fastest-growing expert call transcript library in the industry.

• Hand-picked buyside interviewers (500+ hours experience).

Sign up for a free trial with Stream today!