When Less Feels Better: How Psychology and AI Shape Smarter Travel Decisions

March 5, 202631:04106 views

Summary

Prof. Dr. Kai Markus Müller, a professor of consumer behavior and neuroscientist at HFU Business School (Black Forest) and adviser to Rateboard, presented research on the 'Less-Is-Better Effect' (LIB) and how it diverges from AI evaluative behavior. The session opened by referencing a Chinese EEG study showing that willingness to pay for premium pricing is highest at touristic sites — higher than entertainment attractions or supermarkets — and that consumer purchase decisions involve a perceptual evaluation within 200 milliseconds followed by an emotional judgment around 600 milliseconds.

The core of the presentation was original empirical research developed with graduate student Analena Scodinski, testing 13–15 tourism-specific scenarios for the LIB effect — a cognitive bias where people prefer an objectively inferior option when evaluated in isolation versus comparison. Originally identified at the University of Chicago by Professor Christopher C, Müller confirmed the effect across diverse travel contexts with remarkably consistent results: 10 out of 13 scenarios worked 'instantly' without adjustment.

Key experiments included: (1) A wine bar scenario — a bar selling €450 high-end wine with a fuller glass received more stars than the same bar with a less-filled glass when evaluated individually, but the effect reversed when both were compared simultaneously. (2) A train journey from the Black Forest to Cologne — a 6-hour journey with no disruptions received higher willingness to pay than a 4-hour journey with frequent cancellations and delays up to 90 minutes, even though 4 hours plus 90 minutes still beats 6 hours. (3) A hotel maid scenario — a maid who performs all tasks perfectly received higher tips than a maid who does more tasks but only partially (cleaned 2 of 4 windows, disinfected 1 of 2 remote controls). (4) A breakfast scenario — a smaller, perfect breakfast received a willingness to pay around €15 versus approximately €12 for a larger breakfast with some day-old juices and unavailable egg options.

A critical finding: when the same scenarios were presented to LLMs (ChatGPT, Gemini), the AI systems consistently rated the objectively superior option higher regardless of whether options were shown individually or together — AI does not exhibit the LIB effect. However, when Müller asked ChatGPT to analyze the booking.com pricing table of the New York Hilton Midtown for behavioral economics tricks, it produced an extensive, expert-level analysis identifying anchoring, decoys, loss aversion, mental accounting, and scarcity cues — concluding 'this pricing table is not about rooms. It's about structuring regret, safety and perceived smartness using classic behavioral tools.' Müller described this AI analysis as 'at least as good as my absolute top students if not better.'

The session closed with forward-looking implications for the travel industry: as AI increasingly pre-selects and ranks options before humans see them, a fundamental tension exists between AI's rational optimization logic and human emotional, context-dependent decision-making. Travel businesses must decide whether to optimize for the AI that pre-selects or for the human who ultimately books and experiences the trip.

Key Takeaways

•The Less-Is-Better Effect is highly robust in tourism: across 13–15 original hospitality scenarios, 10 worked immediately without adjustment — a rare level of replicability in consumer behavior research where many famous effects fail to replicate.
•Humans consistently prefer a flawless but objectively inferior option over a better-but-imperfect one when evaluated in isolation. A perfect small breakfast at €15 beats a larger breakfast with some day-old juice at €12 in willingness to pay.
•The LIB effect disappears in direct comparison: when consumers see both options simultaneously, they correctly identify and prefer the economically superior choice — the effect is entirely context-dependent.
•AI models (ChatGPT, Gemini) do not exhibit the Less-Is-Better Effect. They consistently rate the objectively superior option higher regardless of whether options are shown individually or together, demonstrating stable attribute integration.
•Despite not exhibiting human biases, AI deeply understands consumer psychology. When asked to analyze the New York Hilton Midtown's booking.com pricing page, ChatGPT identified anchoring, decoys, loss aversion, mental accounting, and scarcity cues at an expert level — summarizing: 'This pricing table is not about rooms. It's about structuring regret, safety and perceived smartness.'
•Brain scan (EEG) research from China shows willingness to pay is highest at touristic sites, followed by entertainment attractions, and lowest in supermarkets — supporting the broader finding that people pay premium for experiences over goods.
•Purchase decisions occur in two neurological phases: a rapid perceptual evaluation within 200 milliseconds, followed by an emotional judgment around 600 milliseconds — both of which are shaped by context and presentation, not just price.
•The LIB effect was demonstrated across diverse populations: both students and actual travelers at Stockholm airport showed the same preference patterns, confirming the effect is pervasive and not demographic-specific.

Transcript

We're going to continue in uh about five minutes and this keynote is also I think an eye openener because um it's all about how AI is shaping travel decisions and it's about how do people actually decide because even when algorithms filter rank and pre-select offers we know that right so this is what you would like and Amazon tells you you bought that so also you would like that probably because someone else bought that um it's still in the end the human that clicks book and does the purchase. A...