The Doorman Fallacy

How Short-Sightedness Burns You in Business

Mar 20, 2026

Making decisions from the doorman fallacy can make short-term actions simple but will cost you in the long run.

What do retail banking, the Vietnam War, high-frequency trading, the rollout of AI tools, and high-end apartment doormen have in common? This might seem like an odd question, but they all have the same, simple answer: Each one will punish leadership for handling them the wrong way.

Let’s roll the clock back to last year. I was driving home after a long day, listening to Rory Sutherland talk to Chris Williamson on his Modern Wisdom podcast, where Sutherland shared a powerful lesson from his 2019 book, Alchemy. The lesson has been dubbed the doorman fallacy, and I believe it establishes a pattern through history that we’re seeing played out again with the AI boom.

What’s With The Doorman?

The doorman isn’t about the door.

That’s the through‑line that runs from a five‑star hotel lobby, to Great Society social policy, to Vietnam, Wells Fargo, modern market structure, and now the AI gold rush. Once you see it, you can’t unsee it.

We keep making the same mistake:

we define a role by the thin slice of its work that’s easy to describe and measure, then optimize that slice to death—and only later discover that the real value lived in everything we weren’t counting.

The Original Doorman

Rory Sutherland’s parable is simple: a consultant looks at a hotel, sees a uniformed doorman, and decides this is waste. The doorman’s “job” is to open the door. An automatic door is cheaper, never calls in sick, and works 24/7. On the spreadsheet, this is genius.

In real life, it’s vandalism.

Because the doorman was never just opening doors. He was:

Providing visible security and social filtering.
Recognizing regulars, remembering preferences, smoothing frictions before they reached the front desk.
Hailing cabs, answering questions, and signaling: “This place is high‑status and safe enough to employ people to care about you.”

The automatic door performs the notional function flawlessly. However, the hotel quietly loses the actual function. Guests can’t quite put their finger on what changed, but the place feels cheaper, less safe, less special. Bookings drift down. The “cost savings” ends up driving down revenue, brand, and resilience.

That is the Doorman Fallacy in miniature:

Define a rich role in terms of a single obvious verb.
Replace that verb with a cheaper mechanism.
Accidentally erase all the invisible value that made the role worth having.

It’s not just limited to fancy hotels though, we’ve seen this play out at brutal, massive scale.

The Great Society: When Metrics Ate the Mission

The Great Society programs were, in many ways, the doorman problem at national scale. It was good intentioned in theory: eliminate poverty through government programs and support. In reality, the doorman fallacy was in full swing.

Planners defined the “job” of social policy in narrow, countable terms:

Dollars spent.
Units of housing built.
People enrolled in programs.
Benefits delivered.

Those are the doors opening.

But the actual work that holds communities together is harder to measure:

Dense local networks of trust.
Family stability and informal support systems.
Community institutions that create norms, expectations, and meaning.

Urban renewal is an especially stark example. Bulldozing “blighted” neighborhoods of single family homes and replacing them with new towers looked like progress on paper: old units out, new units in, federal dollars deployed, ribbon cuttings achieved.

In reality, the programs often shattered the thick, invisible lattice of social capital in those neighborhoods. The people weren’t just living in old housing; they were embedded in webs of mutual obligation and informal control. When those webs were destroyed, the stats said “more housing,” but the lived reality was atomization, crime, and dependency. Some places even doubled-down on the fallacy so much that they used “hall monitors” in city projects, looking for families in the projects that had a father in the home, something which the Great Society punished (Higgins, Investing In U.S Financial History)

Policy makers optimized for the visible functions of welfare and housing, the automatic door. The real “doormen” of social order, the informal institutions and relationships, were treated as expendable. The costs came later, framed as “unintended consequences,” but they were baked in at the moment roles were misdefined.

Unfortunately for the US, the doorman fallacy was so “successful”, that we took the same ideas into one of the most hated wars in American history: Vietnam

Vietnam and the Body Count

Vietnam was the Doorman Fallacy with a body count.

Robert McNamara’s Pentagon brought a systems‑engineering mindset to war. If you can’t measure it, you can’t manage it. So the war was redefined into a set of metrics:

Bodies counted.
Sorties flown.
Tonnage dropped.
Villages “pacified.”
Territory “secured.”

The “job” of the military, in this abstraction, became eliminating enemy capacity faster than they could replenish it, mostly by stacking more bodies. A war of spreadsheets.

But legitimacy, morale, and political will are not easily measurable. The true “doorman work” of war: understanding local alliances, narratives, and perceptions of legitimacy; recognizing how operations play in the minds of both villagers and American voters, barely showed up in the dashboards.

So the system optimized for higher body counts, more sorties, more villages painted green on maps. It fired the doorman, local knowledge, political nuance, human intelligencem, understanding what our “win” condition was (thank you Dr. Paine for teaching me that), and installed an automatic door: operations that produced favorable numbers.

It “worked,” until suddenly it didn’t. The underlying political structure collapsed, and all those impressive metrics were revealed as theater. The visible function had been maximized; the invisible function had been hollowed out. (We’ll be talking more about this in a later article, so be sure to subscribe if you haven’t)

Wells Fargo: When Cross‑Selling Became the Door

Fast‑forward to modern finance, and the fallacy appears again.

Inside Wells Fargo, in the 2000s and 2010s, retail banking was defined by a deceptively simple metric: products per customer.

“Eight is great.” If you want growth, the thinking went, you deepen relationships by selling each customer more accounts.

Branch staff were now “doors” instead of “doormen” whose job, officially, was to open more accounts. Performance systems, incentives, and culture were wired tightly to that single number.

Everything that didn’t feed that number: the trust that gets someone to deposit their life savings, the sense that “my banker is on my side,” the reputation of the institution, was treated as background noise.

U.S Attorneys Office Press Release, February 2020, justice.gov

The result:

Employees opened millions of unauthorized accounts to hit their targets.
Customers were harmed by fees, credit score damage, and betrayal.
Fines, regulatory shackles, and reputational damage wiped out any short‑term gains.

On the spreadsheet, cross‑sell metrics looked spectacular right up until the scandal broke. The automatic door was gliding open on time. The hotel, meanwhile, was quietly filling with pickpockets.

Once again:

Define the job as a narrow function: “open accounts.”
Optimize that function to the exclusion of all else.
Discover the hard way that your real asset was the trust you weren’t measuring.

Wall Street wasn’t immune either, as we saw with the Flash Crash of 2010.

High‑Frequency Trading and Ghost Liquidity

The Doorman Fallacy also shows up at the level of market design.

After the advent of electronic trading, regulators and exchanges decided that “good” market structure could be captured in a few tidy metrics:

Tight bid‑ask spreads.
High displayed liquidity.
High trading volume.
Fast execution.

If you hit those numbers, the market must be healthy. So incentives were tuned to favor participants, especially high‑frequency traders, who could deliver those stats.

Market makers were implicitly treated as doormen whose job is to “stand by the door and open it quickly.” High‑frequency strategies excelled at that visible job: tight spreads, lots of quotes, plenty of turnover.

But true market health depends on deeper, harder‑to‑measure functions:

Liquidity that stays when volatility spikes.
Participants willing to warehouse risk instead of disappearing at the first sign of trouble.
Investor confidence that the price on the screen will still be there by the time their order arrives.

On May 6, 2010, the Flash Crash, those invisible functions failed in public. Under stress, many algorithms that had been providing beautiful liquidity metrics simply pulled their quotes. Prices crashed downward. Liquidity turned out to be a mirage.

SPY Volume and price on the minute, *Finadings Regarding The Market Events of May 6, 2010,* sec.gov

Again, the visible function had been optimized. The system had “fired” the old doorman, slower, human market makers who carried inventory and reputational obligations, and replaced them with an automatic door that works flawlessly, except when you really need it.

You would think we would learn our lesson, but in the age of AI, the doorman is under threat once again.

AI and the New Doorman

In 2026, we’re seeing the doorman fallacy all over again.

Executives look at customer service reps, loan officers, content creators, analysts, and see a list of verbs: answer, process, categorize, draft. They define the job as “respond to tickets” or “produce content at scale” and conclude, reasonably enough, that a large language model can do that faster and cheaper.

But that list of verbs is the automatic door view.

The real “doorman work” in those roles is:

Reading the emotional state of a customer and deciding when to bend a rule.
Sensing when a small issue hints at a larger, systemic problem.
Carrying institutional memory about what worked last time the system broke.
Signaling status, care, and accountability to counterparties and regulators.

Companies that go all‑in on “AI replacements” discover this the hard way. They save headcount. They celebrate response times and ticket throughput. And then:

Customer satisfaction and loyalty erode, because interactions feel generic and unhelpful.
Brand trust takes a hit, especially in high‑stakes domains like finance and healthcare.
Edge cases pile up, creating failure modes that are more expensive to unwind than the salaries that were cut.

Don’t believe me? You can ask Klarna how well the “oops, all AI” strategy worked out

The pattern is exactly the same:

Take a complex human role and compress it into a single measurable verb.
Deploy an “automatic door” (AI) that does that verb extremely well.
Only later notice that the real value was in the unmeasured functions you quietly removed.

AI isn’t uniquely to blame here. It’s just the newest, shiniest automatic door. The underlying failure mode, reductive definitions of work and short-sightedness in the name of efficiency, is likely as old as business itself.

Why We Keep Falling for It

It’s tempting to frame all of these stories as “unintended consequences.” That makes it sound like bad luck: who could have known?

But each case shares a deeper structural bias:

We gravitate toward what’s legible and quantifiable.
We distrust or discount the value of soft, relational, contextual work.
We reward managers for visible cost savings now, not invisible value preserved over years.

The doorman is easy to cut because the line item is clear and the savings are immediate. The things that disappear with him, safety, status, warmth, trust, show up only later and only indirectly, in a drift of complaints and a slow leak in revenue.

Great Society planners, McNamara’s whiz kids, Wells Fargo’s product‑per‑customer evangelists, market‑structure engineers, and today’s AI transformation offices are all wrestling with the same challenge: how to manage systems whose most important stabilizers are often the ones you can’t easily see.

The Doorman Fallacy is the name for what happens when you pretend those stabilizers don’t exist.

Thank you for reading the article!

If you found this article valuable, then subscribe to the newsletter on LinkedIn or Substack. I am on a mission to make economics, finance, and business easier to understand for the average person, and each subscription helps!

WealthNWisdom

Discussion about this post

Ready for more?