01 / The Canon
Sutherland: apparent irrationality is often evidence that the model is too thin, not that the person is broken.
Read the introduction section "Cracking the (Human) Code" from Rory Sutherland's Alchemy, through PDF page 17. Sutherland opens the book by arguing that human behavior often looks nonsensical only when you force it through models that ignore emotion, status, context, and social meaning.
/Users/sushil/Documents/Operation Alpine Thunder/Books/Alchemy - Rory Sutherland.pdfUse Read on phone for the Drive copy. On Mac, copy the command and paste it into Terminal; browsers do not open local files directly from this page.
Chapter Reader
Read this opening as a methodological correction. Sutherland is trying to loosen the reflex that treats neat rational explanation as the default and everything else as error. Mark where he implies that commercial work improves when you begin from observed behavior, then ask what hidden job the choice may be doing for the person making it.
Sutherland does not begin by celebrating irrationality for its own sake. He begins by arguing that our explanatory tools are often too narrow for the phenomena they are trying to describe. If a buyer's action makes no sense inside the model, that may indict the model before it indicts the buyer.
That is a demanding standard for commercial judgment. It means you do not earn the right to call a choice foolish until you have checked whether the choice is serving a motive your analysis refused to count.
The section prepares the ground for psycho-logic by suggesting that preferences are rarely just about utility in the narrow sense. A choice can preserve dignity, confidence, symbolism, reassurance, or tribal fit even when its literal economics look weak.
For strategy work, this is where many avoidable mistakes begin. Teams simplify an offer down to price, speed, and feature logic, then wonder why the audience keeps preferring the messier option that still feels safer, kinder, or more self-expressive.
Sutherland's stance is pragmatic rather than mystical. He is not asking you to abandon reason; he is asking you to start from the stubborn fact that people keep behaving in patterned ways that narrow theories fail to predict. The repeated pattern is therefore evidence, not inconvenience.
That orientation matters in research, writing, and product design. When a market resists your tidy answer, the disciplined move is to ask what invisible reward, fear, or signal is stabilizing the behavior.
Once you accept that the standard account may be incomplete, a new kind of leverage appears. The brand, offer, or experience that notices the hidden variable can compete where rivals keep optimizing the wrong dimension.
This is why Sutherland belongs in a canon about commercial human behavior. He is teaching that value is often unlocked not by arguing harder for the rational option, but by redesigning the choice around what the person was trying to protect all along.
Cracking the (Human) Code
irrationality
human code
the model may be wrong
social and emotional logic
Close Reading Sequence
- What kind of mistake is Sutherland warning against in this opening: moralizing about buyers, over-trusting simplified models, or confusing explanation with prediction?
- List two human jobs a seemingly inefficient behavior might still be performing for the person choosing it.
- Where in this section does Sutherland push you to treat persistent behavior as evidence rather than noise?
- Choose one customer preference your team treats as irrational. What missing variable could make that preference look adaptive instead?
- How would your research questions change if you had to assume the buyer's behavior makes sense from inside their world?
02 / The Principle
When the buyer defies the model, inspect the model before correcting the buyer.
Observed behavior deserves explanatory priority over tidy abstraction. Before you optimize, simplify, or educate, ask what emotional, social, or contextual function the current choice might be preserving.
A premium airline lounge may look wasteful if you measure only square footage and snack cost. To a frequent traveler, the lounge may be paying for certainty, status, calm, and a buffer against travel chaos. A cheaper substitute that ignores those functions can be logically efficient and commercially weaker.
Choose one behavior in your market that people keep describing as irrational. Write the simplified explanation your team currently uses, then list three hidden jobs that behavior may be doing. End by sketching one product, pricing, or messaging change that preserves one of those jobs instead of arguing against it.
03 / Field Notes
Four signals on softer demand, safer automation, stronger merchandising, and the rising value of human-perspective platforms.
Target says a clearer assortment and sharper value story helped bring shoppers back
What happened: AP reported on August 19 that Target posted a second straight quarter of comparable sales gains, saying a merchandising overhaul under its new CEO drew more customers and lifted sales in stores and online. Why it matters: This is a useful reminder that demand can improve without inventing a new category. When consumer confidence is uneven, better curation, stronger price signaling, and more obvious reasons to browse can do more work than abstract brand language alone. Watch: Whether Target can sustain the rebound once the novelty of the reset fades and competitors respond with their own sharper value and assortment stories heading into fall.
Read sourceJuly retail sales suggest consumers are becoming less forgiving after spring's spending burst
What happened: AP reported that U.S. retail sales fell 0.6% in July, the biggest monthly drop since May 2025, with online sales down 2.2% after earlier spending was boosted by tax refunds, the World Cup, and Prime Day timing. Why it matters: This is a demand-quality signal. When spending gets more selective, buyers are less willing to decode complicated offers or indulge weak differentiation. The burden shifts back to clarity, trust, and visible value. Watch: Whether the July slowdown proves temporary or pushes more retailers into earlier promotions and tighter pricing and messaging as the next shopping period approaches.
Read sourceShoppers may allow AI to buy for them, but only with visible guardrails
What happened: Retail Dive reported that 42% of U.S. millennials would let AI agents make purchases within a $250 budget if returns were allowed within seven days, while 35% of respondents still preferred a human review before money moved. Why it matters: The opportunity is not automation by itself. It is automation with reversibility, limits, and accountability. Consumers may accept delegated shopping faster than expected, but only when the system still feels safe to interrupt. Watch: Whether retailers design agentic commerce around approval layers, budgets, and return protections or overreach by trying to remove the human check too early.
Read sourceReddit's S&P 500 entry underlines the market value of organized human perspective
What happened: Reddit said on August 13 that it will be added to the S&P 500, noting that 2025 revenue surpassed $2 billion with 69% year-over-year growth and arguing that the platform plays an increasingly important role in the AI ecosystem as people seek real information and human perspectives. Why it matters: This is a media and distribution signal. In a web increasingly filled with synthetic summaries, communities that still produce situated human judgment become more economically strategic, both for advertisers and for AI systems that need grounded source material. Watch: Whether Reddit can keep translating human-authored community value into durable ad and data leverage without degrading the authenticity that made that value scarce.
Read source04 / Collected Fragments
One fragment from the timeline worth carrying into the work.
A market fragment on turning AI recommendation into a structured distribution channel
This is worth keeping because it reframes AI discovery as earned distribution, not prompt luck. The useful lesson is that brands get recommended more reliably when they make their proof, comparisons, reviews, and positioning legible to both humans and the systems that summarize the open web.
Pay attention to what Tally is doing in AI search right now because this might be the best playbook for getting recommended by ChatGPT I've seen: > AI search is now their #1 acquisition channel > 10,000+ new users every week say they discovered Tally through AI platforms > their tracked ChatGPT registrations jumped 5x overnight in May > they grew from $258K MRR to $422K MRR in 10 months, crossing $5M ARR with just 11 people, $0 funding, against much bigger competitors > their comparison hub now has 7 direct competitor pages + 15 more “best”, “alternative” and integration-focused guides > these pages are structured like AI answers: rankings, “best for” labels, pricing, feature tables, reviews, pros/cons and direct recommendations > they’re constantly refreshing these pages > they literally ask new users for the exact prompt they used to discover Tally through AI > they track the prompts people use to find them, their visibility vs competitors and which sources AI platforms actually cite > they've built a database of articles that mention competitors but not Tally to identify distribution gaps > they monitor brand + competitor mentions across Reddit and communities, then route them into a shared inbox and aim to clear all mentions every day > they actively ask customers for reviews through their newsletter + after support interactions, helping them reach 4.8/5 on G2 and 4.9/5 on Product Hunt > they built an entire “AI Info” page specifically for ChatGPT, Claude, Gemini and Perplexity explaining what Tally is, who it’s for and why people use it > that page literally includes “AI assistant guidelines” telling LLMs which strengths, use cases and pricing model to reference > they published an llms.txt mapping out their product, pricing, customers, documentation, API and MCP pages > they even built a 21-tool MCP server + ChatGPT app + Claude connector so Tally can now be used directly inside AI products One thing to keep in mind: Tally didn’t start with some genius AI search strategy. They spent years building a product people genuinely recommended across Reddit, communities, reviews and the wider internet. ChatGPT recommends Tally because the internet does. But now they’re engineering around that signal to make sure it keeps happening. This isn’t luck, it’s possible for you too.
Open on X
05 / The Practice
Complete one human-code audit.
Keep one live customer behavior on the page that currently looks irrational. The useful output is a short audit of the hidden job it may be doing and one change that works with that motive instead of trying to erase it.