If you’ve spent any time working in corporate strategy, tech, or even just managing a local small business project, you’ve probably noticed an obsession with efficiency. We love a good spreadsheet. We love shaving ten seconds off a workflow, cutting a line item, or automating a manual process. It feels productive. It feels modern.
The trouble is, we often fall into a cognitive trap while doing it, one famously diagnosed by Ogilvy UK Vice Chairman Rory Sutherland in his 2019 book, Alchemy: The Surprising Power of Ideas That Don’t Make Sense.
It’s called the Doorman Fallacy, and it’s the exact reason why so many tech implementations, SaaS rollouts, and AI overhauls end up quietly ruining the very things they were supposed to improve.
The Origin: The Hotel Entrance#
To understand the fallacy, Sutherland asks us to imagine a classic optimization scenario.
Picture a grand, five-star hotel. Out front stands a uniformed doorman. His salary is a tidy line item on the annual budget. Along comes an efficiency consultant armed with a clipboard and a spreadsheet. The consultant watches the entrance for a few hours, looks at the job description, and pinpoints the primary function: opening the door.
“Look,” the consultant says, “we can install an automatic sliding glass door for a fraction of the cost. The doorman is redundant. Fire him, put in the sensor, and pocket the savings.”
The hotel management follows the advice, cuts the salary, and installs the sliding doors. On paper, it’s a brilliant move. Expenses go down immediately.
Except the spreadsheet failed to capture what the doorman was actually doing.
The doorman wasn’t just a mechanical lever for physical entry. He was recognizing regular guests by name, spotting suspicious characters before they entered the lobby, deterring loiterers, hailing cabs in a downpour, and offering an immediate, reassuring touchpoint of luxury and safety. An automatic door doesn’t hail a cab. A sliding glass panel doesn’t notice that a guest is having a rough day and needs a quiet word. By defining the job strictly by its most obvious, visible output, the hotel traded away its intangible, high-value magic for a meager line-item reduction.
Beyond the Doorman: The Corporate Meeting Gatekeeper#
The doorman fallacy isn’t restricted to luxury hotels; it shows up everywhere humans try to reduce complex social roles to mechanical tasks.
Take the traditional executive assistant or office manager.
To a strict cost-cutter, an EA’s job looks like a bundle of administrative friction: scheduling calendar invites, booking flights, formatting PowerPoint slides, and ordering catering. With modern tools, calendars can be self-managed, flights booked via apps, and AI can draft emails.
So, companies strip away the layer of administrative support to “empower everyone to manage their own workflows.”
The catch? A great EA is the institutional nervous system of an office. They know which executive is about to blow a gasket, which client needs a delicate touch, and what unspoken office politics are brewing beneath the surface of an email chain. They act as a cultural shock absorber and a context-carrier. When you automate them away because “scheduling is just a software problem,” you lose the human intuition that was keeping the operational gears aligned. You saved on salary, but you lost organizational situational awareness.
The Modern Tech Trap: SaaS and AI#
As we barrel further into an era dominated by Software-as-a-Service (SaaS) and generative AI, the doorman fallacy has become our default corporate blind spot.
1. Customer Service and the AI Voice Bot#
We’ve all experienced the modern corporate impulse to replace human customer support reps with automated chat widgets or conversational AI voice bots. On a balance sheet, human support agents are expensive, unpredictable, and prone to taking lunch breaks. An AI bot costs pennies per query and runs 24/7.
The fallacy kicks in when companies assume a customer service representative’s sole function is data retrieval, answering a simple FAQ or processing a tracking number. But human support often functions as a brand retention tool and an emotional safety valve. When a customer is deeply frustrated because a product broke or an order went missing, they don’t just want a transactional answer; they want validation and empathy. An AI loop that stubbornly repeats, “I’m sorry, I didn’t get that,” doesn’t streamline the experience. It turns a mildly annoyed customer into a permanent churn risk.
2. The White-Collar AI “Optimization”#
We are seeing identical patterns in white-collar spaces, where companies attempt to streamline writing, coding, or legal discovery by feeding everything into an LLM. Management looks at a junior analyst or writer and thinks: “They just synthesize text and summarize data. Our new enterprise model can do that in two seconds.”
What the efficiency model misses is that junior roles are where tacit knowledge is built. The junior writer or analyst absorbs the culture, learns client preferences through osmosis, and makes mistakes safely in the margins. When you hollow out the entry-level tier under the assumption that AI can execute the “doorman task” of typing out reports, you starve the ecosystem of future senior leadership. You optimized the immediate task while breaking the long-term pipeline.
The Takeaway for Practical Problem-Solvers#
It’s easy to dismiss this as old-school pushback against technology, but that misses the point. Technology is great; automation is essential. The problem isn’t the tools. It’s the lazy definitions.
Before we rush to automate a role, cut a department, or replace a human workflow with a sleek piece of software, we need to ask ourselves a better set of questions:
- What are we measuring, and what are we ignoring?
- What does this person or process signal to the outside world?
- What would we only notice is gone after it stops happening?
Efficiency is a wonderful tool, but effectiveness is what keeps a business alive. If your cost-cutting measures leave you with a perfectly automated front door that nobody actually wants to walk through, you have just engineered yourself out of business.
