Hey Reader - Thomas here.

Good day to my fellow compounders!

In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies.

Enjoy!

I. Agentic Commerce vs. the Average Shopper

The average American makes 47 payments a month, according to Federal Reserve data that Michael Morton walks through on Eric Seufert's Mobile Dev Memo podcast. Morton covers e-commerce and marketplaces at MoffettNathanson, and his breakdown goes like this:

10 are bills, mostly automated 10 are prepared food 6 are gas stations and transfers to friends

That leaves 21, and if you're generous and call half of those online, the average person makes about ten e-commerce purchases a month. His point is that Silicon Valley and Wall Street are extrapolating their own lives onto everyone else, and the average person isn't drowning in some "crushing burden" of purchasing decisions. If that's anywhere near right, these agents are fighting over a pretty small pie.

The other half of his argument is that marketplaces have moats. A little over a dozen operators account for 70% of US e-commerce, and 65% of their web traffic is people typing the address straight into the browser. Amazon's delivery network is bigger than UPS now, which he calls "the largest physical moat I can think of in my investment research career."

Where he does see change is product discovery. An LLM did in 30 seconds the research that took him 40 hours, and Mercado Libre has said its LLM-powered discovery means fewer clicks but higher conversion, with cost per click rising more than enough to offset the decline. He also explains why, if you're waiting for the marketplace selloff to end before buying, he "would be terrified of the Muse numbers for the near to medium term" (Muse being Meta's new personal agent).

​Click here to listen​

II. Growth Doesn't Fix Bad Unit Economics

Most startup advice says grow first and fix the economics later. Benjamin Hallen, Ron Tidhar and Kathleen Eisenhardt argue in Harvard Business Review that this is backwards, and they've got numbers. In 2025 they studied more than 1,000 startups that reached unicorn status over the past decade and found that fewer than 150 became prosperous public companies. I was surprised it's that low.

Unit economics just means the revenue and cost of each additional sale, including what it costs to win that customer. MoviePass had millions of subscribers paying a flat monthly fee for virtually unlimited movie tickets, lost money on nearly every transaction, and ended up in bankruptcy. The authors call the fix-it-later plan "grow and then gouge" and can think of very few startups that pulled it off.

Amazon ran losses for years too, but each retail transaction covered payments, warehousing and shipping, and the losses came from pouring cash into new businesses like AWS. That's the difference between unit economics and firm-level profitability. Both looked loss-making on paper, so the question is which of the two you're looking at.

Of their foundation-first examples, the one that stuck with me is Starbucks. Howard Schultz's rule was that a new store had to bring in $2 for every $1 invested to build it in its first year, and locations were checked against that before the company expanded. The essay also covers Klaviyo's three slow years before it raised serious money, Notion's near-death rebuild, and the roughly $3 billion Goldman Sachs lost on Marcus, its consumer banking venture.

​Click here to read​

III. Booking Holdings: Death by Agent, Round Two (SCIS Members)

This one's paywalled. It's my latest report for Steady Compounding Insider Stocks (SCIS) members, so I'll only say what it covers.

Back in February I wrote that the AI bear case on Booking Holdings has it backwards. Meta's Muse hadn't been released yet. This time round the Grim Reaper comes in the shape of Muse, and the market has once again decided that online travel agencies will get slaughtered by AI agents.

So the report starts with a premortem, a concept I discussed in my book. It's 2029. Booking's room nights are shrinking. What happened?

SCIS members: Click here to read​

Not a member yet? Click here to join​

That's all I have for you today!

Till next week.

Thomas Founder of Steady Compounding Connect with me: ​​​​

​