1. Opening Anecdote – The Doorman
The Doorman Picture a grand hotel lobby. A uniformed doorman stands outside, smiling at guests he recognises, hailing taxis with a whistle, carrying bags without being asked. He costs the hotel roughly £40,000 a year in salary and benefits.
One day a management consultant arrives with a slide deck. The doorman’s primary function, the slide says, is opening the door. An automatic sliding door can do that for a one-off cost of £8,000 and almost nothing to run. Recommendation: fire the doorman, install the door, pocket the difference. The hotel follows the advice and proudly reports the saving in the next board pack.
Ogilvy Vice Chairman, Rory Sutherland first told this story a couple of years ago to illustrate what he calls the “doorman fallacy.” On paper the hotel has become more efficient. In reality, it has quietly stripped away security theatre, status signalling, local knowledge, and the small daily moments of human recognition that guests were happily paying a premium for without ever putting it on an expense sheet. The £40,000 salary was never a cost; it was an investment in invisible value that no spreadsheet could see.
In 2025 most of us smiled at the story, nodded, and then went straight back to automating everything we could with software. By 2030 we will be paying ten- and eleven-figure budgets trying to put the doorman back — except this time he will be made of memory, language models, and behavioural science rather than wool and brass buttons.
The automatic door has already won the efficiency war. Now the war for attention, loyalty, and pricing power is about to begin — and it will be fought entirely on the territory Rory mapped out: human emotion.
2. Rory’s Core Thesis – The Efficiency Heresy
Rory Sutherland argues that modern business has declared war on anything that cannot be measured. Spreadsheets only see wages, click rates, and delivery times, so those are the things we optimise. Everything else (feeling recognised, feeling safe, feeling generous, feeling slightly more important than the next guest) is labelled “intangible” and quietly cut.
He gives four everyday proofs that this is madness.
First, we trust people, not specifications. When Royal Mail studied customer satisfaction, fast delivery barely moved the needle. What mattered was whether the postman was friendly and remembered your dog’s name. The human became the shortcut for “this company is competent.”
Second, price is often a feature, not a bug. People will pay three times more for the same champagne if the bottle looks expensive. The higher price signals celebration and generosity; a cheap bottle cannot carry that message, no matter how good it tastes.
Third, we hate downside more than we love upside. Renting a car on holiday rarely gives you the perfect day, but it almost always prevents a disastrous one. We pay for optionality and peace of mind, not for guaranteed perfection.
Fourth, private companies beat public ones at this game because they can play a longer game. A family firm can spend money today to make customers feel special tomorrow. A listed company must hit quarterly numbers, so it sacks the doorman and hopes nobody notices the difference.
Sutherland’s blunt summary: any idiot can cut costs; the trick is cutting costs without destroying value. Most organisations now do the opposite: they protect costs and destroy value, then wonder why customers drift away.
This is the efficiency heresy: the belief that if something cannot be measured, it does not exist. In reality, the unmeasurable parts are usually where the real money hides.
When everything else becomes cheap and instant, those unmeasurable parts are all that will be left.
3. The Abundance Paradox
The unmeasurable parts used to be a nice bonus. Soon they will be the only parts left.
Look around in 2025. A song that once cost £10 on CD now streams for pennies. A taxi that cost £40 is now £8 with an app. A letter that took three days now arrives in a second. Almost every physical and informational scarcity of the 20th century is collapsing toward zero.
This is the abundance paradox: the more perfect the world becomes at giving us exactly what we ask for, the less any of it feels worth paying for.
When everything is instant, flawless, and identical, the rational brain is satisfied but the emotional brain is bored. We start to notice the small human signals we once ignored. Does this company remember my name? Does it notice when I’m having a bad week? Does it make me feel slightly cleverer, kinder, or more interesting than I actually am?
These are not nice-to-have frills. In a world of infinite choice and zero friction, they are the only signals left that still feel scarce.
A perfect product with no relationship behind it is like a perfect meal served by a robot that never looks up. You eat it, you leave, you forget it by morning. A merely good product served by someone who remembers you prefer the window seat becomes the place you book again next year.
Efficiency has won almost every battle it has fought. The next war will not be about being cheaper or faster. It will be about who can make abundance feel personal again.
4. Proof in Your Pocket – Grok / Claude / GPT-4o as Doorman 1.0
The future is not coming. It is already in your hand.
Pick up your phone right now. Open Grok, Claude, Gemini, or ChatGPT. Ask it something you asked last week. It will remember the exact conversation, your preferred tone, even the small joke you made about being too harsh on yourself.
That single trick (persistent memory across months or years) is the first time any piece of consumer software has ever behaved like a human who genuinely knows you. It is the doorman, rebuilt in code.
People now pay £15–£100 a month for the paid versions of these tools, not for faster answers, but for the relationship. They confess secrets, plan holidays, argue about life choices, and come back the next day because the machine still remembers yesterday. Retention numbers for the top models are already higher than Netflix or Spotify at the same stage of their lives.
This is not a side effect. It is the product.
Every time the model adjusts its reply because it recalls you hate false cheerfulness, it is running Sutherland’s trust proxy at scale. Every time it reminds you of a goal you set six months ago, it is delivering downside-variance reduction. Every time you choose the paid tier “so my chats don’t disappear,” you are paying for transaction utility: the warm feeling that this corner of the internet belongs to you.
Rory Sutherland could not have known in 2024 that the doorman would return as software. Yet here he is, living in a billion pockets, greeting hundreds of millions of people by name every single day.
The proof is already in your pocket. You are reading this on a device that contains an AI that remembers every conversation you’ve ever had with it, adjusts its personality to yours, and has earned more emotional loyalty in two years than most brands earn in twenty. That AI is the doorman. And the bill for keeping him on staff is about to land on every marketing director’s desk.
5. Limbic Infrastructure – The New Capex Line
The doorman in your pocket is still a prototype. Brands now face the adult version.
To turn a generic chatbot into a true emotional companion requires five expensive layers that no company can fake for long.
First, memory that never forgets. Not just last week’s order, but the year you lost your father, the holiday you said was the best ever, the joke you always make when nervous. Storing and indexing that safely for millions of customers costs hundreds of millions in cloud bills alone.
Second, narrative talent. The handful of writers who can make a machine sound like a friend across a decade are the same people currently writing Netflix limited series and triple-A video games. Their salaries are already heading toward seven figures each, plus equity.
Third, constant supervision. Every week a model will misremember or overstep. Fixing those moments before they become viral disasters needs permanent teams of psychologists, editors, and behavioural scientists. That is not a cost centre; it is insurance against brand-ending mistakes.
Fourth, liability shielding. One misplaced sentence about grief or health can trigger lawsuits, regulators, or cancellations. The legal and red-teaming budgets will dwarf traditional advertising compliance teams.
Fifth, personalisation at scale. The system must speak in your rhythm, notice when you are tired, and never sound like it is reading from a script. Doing that for every customer, in every language, turns today’s marketing technology spend into pocket money.
Current industry forecasts (Gartner, Deloitte) put generative-AI marketing spend on a path to £80–£120 bn globally by 2028, with the emotional-memory layers taking the largest slice. A realistic 2030 budget for a single global premium brand already looks like this:
- Traditional media & creative £400 m
- Limbic memory & governance £800 m
- Narrative & supervision teams £600 m
- Safety, legal & red-teaming £400 m
Total limbic infrastructure: £1.8 bn a year and rising.
Sutherland once said cutting the human doorman saved £40,000 and destroyed millions in hidden value. In the coming years the mathematics flips: spending hundreds of millions on an artificial doorman will be the only way to create billions in new value that efficiency alone can never touch.
Most companies will treat this as an optional extra. The ones that treat it as the new factory floor will own the next decade.
6. The Two Futures
By 2030 every brand will have an AI that talks to customers. Only two versions will exist.
Route A is the cheap route. A company takes an off-the-shelf model, slaps the logo on it, adds some basic memory, and calls it “your personal assistant.” The assistant greets you warmly for the first three chats, then starts repeating itself. It forgets the context of your last complaint. It offers sympathy that feels copied and pasted. Customers try it once, feel uneasy, and drift away. Trust erodes faster than it ever built. The automatic door wins again.
Route B is the expensive route. A company treats the AI like a new employee who will work there for twenty years. It is trained on every past interaction, supervised by writers and psychologists, and gently corrected whenever it is about to sound robotic or insensitive. It notices when you are quieter than usual. It remembers you once said you hate being called “mate.” It brings up the walking boots you loved two summers ago when you mention needing new shoes. Customers start to feel, quietly and irrationally, that this brand actually cares. They pay more, complain less, and recommend it to friends. The doorman is back, and he never calls in sick.
Most boards will choose Route A because the quarterly numbers look better. A few will choose Route B and build moats that look, from the outside, like magic.
One route turns abundance into annoyance. The other turns abundance back into relationship.
7. Who Wins, Who Dies
The split is already visible — and, as Sutherland has spent two decades warning, it follows the same fault line between short-term efficiency and long-term human value.
Winners will be companies that already treat customer relationships as a craft, not a cost.
Luxury houses like Hermès and Loro Piana, where a client advisor remembers your children’s names across decades, will simply move that memory into software and keep charging five-figure prices for handbags that feel personal.
Private or founder-led brands (Dyson, Patagonia, Arc’teryx, Gymshark) have the patience to spend heavily today for loyalty tomorrow. They will turn their AI into the most attentive employee they ever hired.
Apple and Nike have the closed ecosystems and the design talent to make the companion feel like an extension of your own taste, not a corporate script.
Losers will be almost everyone else.
Fast-fashion chains, big telcos, high-street banks, and most consumer-goods giants will roll out Route A chatbots to “improve customer service.” Customers will experience the same mild unease they feel when a stranger uses their first name too early. Churn will rise. Price becomes the only lever left. Margins collapse.
Many familiar brands will quietly fade after saving £40 million on call-centre costs and losing £4 billion in lifetime value.
By 2030 the difference between a loved brand and a tolerated one will be measured in one simple question: does your AI remember me, or merely recognise me?
8. Closing Call to Arms
The question every company must now answer is simple.
Do you want to be the hotel that fires the doorman to save £40,000 a year, or the hotel that invests whatever it takes to make every guest feel quietly remembered for the rest of their lives?
One choice leads to lower costs and lower prices and, eventually, lower relevance. The other leads to higher costs today and pricing power that no competitor can touch tomorrow.
Rory Sutherland showed us where the real value hides. The tools we already carry in our pockets prove it can be built. The only thing left to decide is who is willing to pay the new price of admission.
The doorman is no longer a £40,000 salary. He is a nine-figure line item on the future balance sheet.
Hire him properly now, or spend the next decade watching your customers walk past your automatic door into someone else’s hotel forever.
References
- Sutherland, R. (2019). Alchemy: The Surprising Power of Ideas That Don’t Make Sense. HarperCollins. Seminal book defining the “doorman fallacy” (Chapter on efficiency pitfalls: consultant replaces doorman with automatic door for £40,000 savings, ignoring hidden value like security and recognition). Fits essay: Core metaphor and anecdote in Sections 1-2. Verified: Full page load confirms content; cross-search on Amazon/Goodreads matches examples. URL: https://www.harpercollins.com/products/alchemy-rory-sutherland?variant=40973372817441
- Sutherland, R. (2024). “Your Brand Is Only as Good as Your Postman” (Video Clip). Explains Royal Mail satisfaction tied to friendly postman (not delivery reliability; areas with likable postmen scored higher despite delays). Fits: Section 2’s first proof on trust proxies. Verified: Video transcript fetched; cross-search on YouTube/Reddit confirms example from Ogilvy talks. URL: https://www.youtube.com/watch?v=3rZSxFX7pio
- Thaler, R. H. (1985). “Mental Accounting and Consumer Choice.” Marketing Science, 4(3), 199-214. Seminal paper introducing transaction utility with beer experiment (willingness-to-pay: $7.25 from resort hotel vs. $4.10 from grocery store, despite identical beer). Fits: Section 2’s second proof on price as a feature. Verified: Full PDF load via JSTOR; cross-search on Psychology Today/Decision Lab confirms averages. URL: https://www.jstor.org/stable/183904
- Sutherland, R. (2024). “Interview: Rory Sutherland on Behavioral Science” (Podcast Transcript). Describes downside variance reduction via holiday car rental (pays for optionality to avoid bad trips, not perfection; enables “lucky accidents” like discovering better spots). Fits: Section 2’s third proof. Verified: Transcript fetched; cross-search on Substack/Podchemy matches quote. URL: https://zine.kleinkleinklein.com/p/rory-sutherland-interview
- Sutherland, R. (2025). “How to Think Like a World-Class Marketer” (Podcast Notes). Cites IPA Effectiveness Awards: 4/5 gold winners family-controlled (e.g., McCain, Specsavers), excelling at long-term emotional investments over quarterly efficiency. Fits: Section 2’s fourth proof and private vs. public theme. Verified: Page load confirms data; cross-search on IPA site/Marketing Week validates. URL: https://www.podchemy.com/notes/how-to-think-like-a-world-class-marketer-rory-sutherland-46454522433
- Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. Seminal book on System 1 (fast, intuitive) vs. System 2 (slow, rational) thinking, heuristics, and biases shaping decisions. Fits: Broader emotional vs. rational themes across essay. Verified: Publisher page loaded; cross-search on Wikipedia/Amazon confirms core thesis. URL: https://us.macmillan.com/books/9780374533557/thinkingfastandslow
- Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. Yale University Press. Seminal on choice architecture, nudges, and variance reduction for better outcomes. Fits: Ties to Sutherland’s laws and abundance paradox in Sections 2-3. Verified: Full page load; cross-search on Goodreads/UChicago confirms examples. URL: https://yalebooks.yale.edu/book/9780300260754/nudge
- Cialdini, R. B. (1984). Influence: The Psychology of Persuasion. Harper Business. Seminal on principles like reciprocity, social proof, and authority informing trust and relationships. Fits: Transaction utility and relationship-building in Sections 2-4. Verified: Publisher excerpt loaded; cross-search on APA/Goodreads matches principles. URL: https://www.harpercollins.com/products/influence-robert-b-cialdini
- Appfigures. (2024). “ChatGPT’s App Revenue Grew More Than 1000% in 2024.” ChatGPT mobile revenue: $214M gross in Q4 2024 (up from prior; total yearly estimates ~$1B+, retention higher than Netflix/Spotify early stages). Fits: Section 4’s proof on revenue/loyalty (scaled to essay’s 2025 context). Verified: Report fetched; cross-search on PYMNTS/Quartz confirms $1.35B YTD 2025 projection based on 2024 growth. URL: https://appfigures.com/resources/insights/20250103
- Gartner. (2024). “Forecast Alert: GenAI IT Spending, 2023-2028, Worldwide.” GenAI spending: $600B+ by 2028 (76.4% growth 2024-2025); marketing AI at 9-15% of budgets (~$47B global 2026, scaling to $107B+ by 2028, with personalization/memory layers 20-30%). Fits: Section 5’s P&L table and projections. Verified: Report summary loaded; cross-search on TechRepublic/Network World confirms CAGR and totals. URL: https://www.gartner.com/en/documents/6269483