Note: This piece was originally shared internally at Lufthansa Group (March 2025) and has been slightly updated (in May 2025) for broader publication. It is also shared on LinkedIn.

The glint of sunlight on the Lufthansa fuselage. That image, taken years ago in Buenos Aires, instantly brought back a wave of… well, let’s call it travel regret.

While I was writing this post about Argentina and tariffs, I was looking for a cover image. In that process, I browsed through the photos I took during that trip. Although I failed to find an adequate cover image from my library, I did find this picture of a beautiful aircraft…

It brought back memories. I actually flew to Buenos Aires (EZE) with British Airways, not Lufthansa. I took this picture shortly after arriving. My flight with BA was not smooth—the LHR-EZE flight was canceled a few days before my departure, and I was rebooked via Madrid with Iberia.

I could have flown directly from Frankfurt to Buenos Aires instead of detouring through Madrid. I could have enjoyed Lufthansa’s higher standards instead of Iberia’s offering, which did not even have an entertainment system (on a long-haul flight!). And I would have paid more or less the same price.

So why did I make that choice? My decision was driven by two factors.

First, many of my classmates were departing from London, and it seemed more fun to travel together. The BA cancellation snafu did not jeopardize that—we all ended up together from Madrid. 😊

Second, a more delicate issue: at the time of booking, Lufthansa was experiencing frequent strikes, creating uncertainty for travelers like me (2015–2016), and I could not afford to miss my South America trip. As a risk mitigation measure, I chose a competitor—even if I found the offering far less appealing in terms of price-to-quality. The risk profile simply sounded better.

When I finally landed in Buenos Aires after a rough journey, stepping out of Iberia’s overly old plane, I saw this beautiful Lufthansa aircraft. Easy to imagine what went through my mind: I could have flown on that one! And in the end, it was not even canceled. I took a picture to mark that moment of regret.

I have been working with customer data for quite some time now. And yet, I often wonder whether the patterns we try to extract from coupon data to improve customer journeys even exist in an integral form. A sort of Gödel’s Incompleteness Theorem for data—where the available information is inherently insufficient for its intended purpose.

I have seen many customer segmentations. Personas, clustering, behavioral groups… all more or less the same. “It has to be actionable,” they say, and for that, “segments must be mutually exclusive”—a given customer must fall into one box. Although not available a few years ago, today, asking a modern LLM to do this for an airline yields, in a few seconds, the same results as those expensive reports from consulting firms.

And yet, what box would I fall into? Risk aversion and the ability to absorb disruption dramatically shaped my choices—far more than traditional segmentation categories like family status, income, trip purpose, or cohort.

Take my frequent trips between Frankfurt and Cambridge (UK). The “rational” advice? Take the train from Heathrow to Cambridge—cleaner, safer, straightforward. And yet, I systematically took the National Express coach. I have done it dozens of times, and I have never regretted it. The reason? UK trains are notorious for frequent disruptions and massive delays. I have experienced this firsthand while living there. I must confess that now, living in Germany, I was not surprised to read in the Financial Times that German trains are now considered worse than UK trains (on.ft.com/4hFj02v). From my own experience, I would say they are both just as bad. But the Financial Times looked at the data—and found that German trains are even worse.

Often, the optimal choice in theory does not align with lived experience. We are often better off taking the car, despite the known theoretical advantages of the train.

My experience leads me to believe that in many organizations, data initiatives are a mix of analysis and judgment. While this is not inherently wrong, it does shape how conclusions are formed—this is simply how traditional firms use data: a mixture of expert opinion and mathematical models. BI dashboards—essentially intuitive representations of data—and, quite often, Excel sheets remain among the most useful tools for decision-making in these firms, airlines included. Despite all the hype, Big Data has not changed that.

So the real question is: how are these opinions formed?

A couple of years ago, I remember seeing posts on social media about Facebook’s engineers being skeptical about the metaverse. Apparently, those working on it did not understand it, did not use it, and did not feel curious about using it. A recipe for disaster. These same engineers loved and actively used other social media products they had built. Those products were well-designed, highly engaging, and widely adopted. The metaverse? Not so much. There are plenty of examples like this.

In Skin in the Game, Nassim Taleb illustrates a similar issue with two striking examples. First, he points out that many speakers feel uncomfortable on stage without realizing why. In reality, the harsh stage lighting affects their concentration—but since they are not lighting engineers, they attribute their discomfort simply to “being on stage,” and the issue persists. Meanwhile, lighting experts do not present on stage.

Another example comes from the redesign of Metro North trains in New York. The trains were given a sleek, modern look, complete with power outlets that hardly anyone used. But in the process, a small, practical ledge by the window—where commuters could rest their morning coffee while reading—was redesigned at an angle, making it useless. The designers, likely unfamiliar with the habits of daily commuters, had prioritized aesthetics over functionality.

Both cases highlight the same pattern: when decision-makers do not experience the consequences of their own choices, seemingly rational decisions often miss crucial real-world details.

I even remember a middle manager once asking in a meeting what an “involuntary upgrade” was—because the term had appeared on a slide. He genuinely did not know.

That moment stuck with me. It was a reminder that, as airline employees, we do not experience the product in the same way our customers do. We fly on ID tickets, booked through an internal system, often with different rules and conditions. We discuss customer pain points in meetings, but it is more like a role-playing exercise—a theoretical understanding rather than lived experience.

This is not unique to airlines. It happens in many industries. But it creates a fundamental gap. The middle manager I mentioned? He had never worked outside the airline. He had never been a real paying customer. So how could he truly understand what frequent flyers value most?

When I used to study and work in Japan (2002–2010), I traveled frequently across four continents for both business and academia. Back then, I was a real customer. I booked tickets like any other traveler, faced the same disruptions, and made choices accordingly. One of the common coffee machine discussions—since our coffee machines were shared across departments, meaning scientists, engineers, HR, IT, and international sales all talked—was about involuntary upgrades. That was the thing to get! Some frequent flyers were regularly upgraded simply because economy class was overbooked, and the airline needed to free up seats by moving loyal customers to business class. Naturally, we started looking for patterns—trying to anticipate which flights were likely to be overbooked and factoring that into our business travel plans.

I enjoyed that experience only twice—once on KIX-CDG and another time on CDG-KIX—putting me at the bottom of the leaderboard.

And then, things changed. When Air France introduced premium economy, it disrupted the unspoken upgrade game. No longer were we just moved up to business class—now, some of us were “upgraded” to premium economy instead. For international sales teams who had grown accustomed to business class, this was not a welcome change. It felt more like a downgrade masked as an upgrade.

In the end, how decisions are made—whether at the customer level, the corporate level, or even in large-scale engineering projects—often comes down to a fundamental disconnect. People who make decisions are not always the ones with skin in the game. And that gap between theoretical optimization and lived reality is precisely why so many well-intentioned decisions end up missing the mark.