First, Second, Third Order

The barcode arrived in 1974.

The first product scanned at a checkout was a pack of Wrigley's Juicy Fruit chewing gum, in a Marsh Supermarket in Troy, Ohio. The store manager later said the technology was "kind of slick, but mostly it just sped up the line."

For 30 years before that, the same can of beans, the same loaf of bread, the same bottle of Coke had been typed in by hand. The barcode replaced that with a beep. Checkout got about 30% faster. That was the headline story.

This is the first-order story. Faster. Cheaper. Fewer errors. It's the story that everyone tells about every new technology, because it's the easiest to measure and the easiest to explain.

It's also the smallest part of what actually happens.

The first-order effects of a barcode were efficiency gains for the cashier. The actual reason the barcode was the most important retail technology of the 20th century is not in the first-order effects at all. Most companies stop reading the story right here.

Second-order effect: a power shift between actors.

Every scan was a data point. Every data point told the retailer what was actually selling, in real time, at every store. Pre-barcode, brands had power: they told retailers what to stock, paid for placement, controlled marketing. Post-barcode, retailers had power: they could measure performance, demand standardization, and play suppliers against each other on actual evidence.

Walmart, which understood this in 1980 before almost anyone else, dictated terms to suppliers for the next 40 years.

Third-order effect: the entire industry restructures around the new power.

Brands that couldn't conform to retailer-defined SKUs, packaging standards, and pricing terms got delisted and went bankrupt. Manufacturing relocated globally because retailers could now demand cheaper. Whole categories — like local grocery stores, small-format retail, and brand-loyal consumers — declined. Global supply chains, just-in-time logistics, and the dominance of Walmart, Costco, and Target all flow from the third-order effect.

The barcode didn't make checkout faster. The barcode rewrote who held power in retail for two generations.

Now AI in 2026.

Look at how almost every enterprise frames it: "Our AI copilot cut support tickets 35%." "We saved 20% on legal review." "Our marketing team is 2× more productive."

This is the first-order story. It's true. It's measurable. It's also exactly the conversation that retail had in 1976 about barcodes: "checkout is 30% faster!" It captures almost nothing of what's actually about to happen.

Second order: who holds power shifts.

In every market AI touches, a new player is becoming the coordination layer. Whoever sees the most user intent (the recommender), or absorbs the most institutional knowledge (the agent), or owns the trust layer (the certifier) — that player extracts the rent that used to go to incumbents.

The vendor giving you the AI copilot is gathering the data your employees produce. Five years from now, that vendor will dictate terms to you the way Walmart dictated to suppliers. The 35% savings was the entrance ticket.

The first-order effects of automation make headlines. The second- and third-order effects redistribute the economy. The wealth never moves at the first order.

— From Reshuffle, Chapter 3

The strategic test for any AI initiative: can you name its second- and third-order effects, not just its first? If your AI plan reads like the 1976 newsletter about barcodes — "we'll save 30% on operations!" — your strategy is to be Kmart in this analogy.

The next Walmart is being built right now, in some category, by whoever is reading the second- and third-order effects of AI correctly. That category might be yours.

Sangeet on this in Chapter 3 ↗