The spreadsheet looked great

What gets cut when nobody can find a number for it.

There were two doormen at the entrance of the Stephen’s Green Hotel in Dublin. Top hats, long coats, the whole thing. They greeted people coming in, hailed cabs, made small talk with someone wearing a welsh rugby jersey who’d clearly stayed there before. The lobby behind them had a sound to it that hotels with sliding glass doors don’t have.

It is rare now. Most hotels at that level removed their doormen years ago. The math to replace them always made sense. Around $50k a year per door, give or take. A sensor never calls in sick. Nobody wrote a memo about what was being lost, because on paper nothing was being lost. A salary line went to zero and the spreadsheet looked great.

What the doorman was actually doing was reading the room. He registered the suitcase, the pace, the look on someone’s face after a delayed flight, and adjusted. He remembered the guest from three visits ago. He produced a tiny social signal, you are in the kind of place that notices you, that the building itself could not produce. Rory Sutherland wrote about this years ago. I have been noticing it run, in different clothes, for most of my career.

The cut always makes sense on the model. But what if the model is measuring the wrong thing?

The first time I watched it happen, I was working on Ireland’s first digital bank. Two-factor authentication then meant a small device that looked like a calculator. From a product point of view it was friction. I hated it. It was added cost and I wanted it gone. Then we asked customers, and they told us, clearly and repeatedly, that they liked it being annoying. This was their savings account. Most of them put money in at the start of the month and pulled it out by the end. The device gave them a few seconds to think. Many even lost it and never looked for replacement because it meant they could never access their savings, so their savings kept accumulating.

We ended up running marketing campaigns that called the friction a feature, money at arm’s length. The thing the spreadsheet wanted to delete was the thing customers were paying us for.

Years later, in fintech in MENA, the line item under pressure was events. Sponsorships, exhibitions, the cost of being in the room. The attribution was awful and never straight forward to measure. Almost no deal could be tracked back to a stand without some serious pseudoscience. Easy cut, on the model. What we kept finding was that businesses we’d met at events came back six months later, after a chain of conversations that started because they’d seen us there. The act of paying budget to be present was itself the trust signal. It build credibility.

The spreadsheet saw a cost with no leads closed in the first month post the event. What was actually being bought was the right to be considered credible in a market where credibility is the entire product.

A different version of the same pattern, running in reverse, in real estate in Dubai. We tried to build a platform to sell property fully online. Cut the real estate agent, cut the showroom visit, cut the trust gap, cut the cost. It sounded amazing and music to many ears. Also looked good on a slide. It was dumb and it didn’t work at full ticket sizes. People did not click “buy” on a house. The version that did work was a hybrid, a deposit online, with a window to come and see the place in person before committing. We had to put friction back in. The friction was not in the way of the trust. The friction was the trust.

I am watching the same conversation happen now in almost every company I talk to. The new voice AI bot with a generic voice will replace the friendly human onboarding call. The generic writing automation will take care of new marketing blog posts. I’ve even sat in a meeting where someone proposed AI to interview all new candidates. The numbers make sense again. The thing being removed is, again, hard to put on a slide. Some of those teams are going to find out, in 18 months, what was actually keeping the customer there or what was attracting top talent.

Some of what builds a business has never made it onto a dashboard, and probably never will. The doorman, the calculator device, the event stand, the showroom visit, none of them survive a financial review on their own merits, because their merits are not the kind of thing a financial review can see. The people making these cuts are usually smart. And in a meeting, only the model gets to speak. So the cut goes through, every time, until the trust is gone and nobody can say exactly when it left.

The companies that kept the doorman looked stupid at the time. Now they’re the ones that feel premium.

See you out there.

Martin

P.S. My writing soundtrack The Doors - Break On Through