Abstract
Information providers can significantly boost their profits by offering a "teaser" of free information before a sale, a strategy that exploits human cognitive biases. While mathematically ineffective against perfectly rational agents, this Preliminary partial Free Information Disclosure (PFID) leverages a "strategic blindness" in human buyers, who often fail to account for the seller's motives. This paper explores the efficacy of PFID through a series of behavioral experiments, demonstrating profit increases of up to 35% when selling to human subjects.
Introduction: The Value of Uncertainty
In modern economies, consumers frequently pay for information to reduce uncertainty—examples include vehicle history reports, financial forecasts, or expert consultations. Standard economic theory posits that if a seller provides partial information for free (e.g., "I'll tell you the car wasn't flooded for free, but pay for the full report"), a rational buyer should recognize the strategic intent. A rational agent understands that the seller only discloses facts that make the remaining information appear more valuable. Consequently, the agent updates their beliefs, and the seller's total profit should remain unchanged.
However, human decision-makers are "boundedly rational." This study investigates whether the PFID strategy can exploit human cognitive limitations to increase a provider's profit.
Experimental Design: "What's In The Box?"
To test the PFID strategy, researchers developed a mathematical model comparing Rational Buyers (who account for seller strategy) against Naive Buyers (who perform simple calculations based only on visible outcomes). These models were then tested against real-world human behavior using a custom-built game titled "What's In The Box?".
In each round, participants were presented with a box containing a prize of unknown value, selected from a list of possible outcomes (ranging from -50 to +50 points). Participants had three choices:
- Open the box blindly.
- Opt-out for a guaranteed zero points.
- Pay a fee to reveal the exact value before deciding whether to "take" the prize.
The study utilized 300 participants from Amazon Mechanical Turk across three treatments:
- Baseline: No free information provided.
- Explicitly Strategic: The system removes some impossible values for free, explicitly stating this was done by a self-interested agent to maximize profit.
- Non-Explicitly Strategic: The system removes values "to help the player," with no mention of strategic intent.
Results and Analysis: Strategic Blindness and Profitability
The experimental data, comprising 6,000 individual decisions, revealed several key insights into human economic behavior.
The "Strategic Blindness" Effect
The most striking finding was that human behavior remained nearly identical regardless of whether the provider was framed as "helpful" or "strategic." Participants consistently failed to ask why specific options were removed, instead taking the remaining possibilities at face value. This "strategic blindness" suggests that humans do not naturally account for the adversarial nature of information disclosure.
The Profitability of "Free"
While PFID is profit-neutral against rational agents, it proved highly effective against humans. The strategy increased the information provider's average profit by 29% to 35% compared to the baseline. Furthermore, the frequency of information purchases jumped by 15–17% when a teaser was provided.
The Teaser Trap
By eliminating the "worst-case" scenarios for free, the seller effectively "nudges" the buyer. Removing the most negative outcomes makes the remaining uncertainty feel more manageable yet still dangerous enough to warrant paying for a full reveal. While humans are generally capable of identifying high-value information, they fall back on naive reasoning as decisions become more nuanced, leading to sub-optimal purchases that favor the seller.
Discussion and Future Work
The study demonstrates that failing to think strategically is a more significant driver of seller profit than poor mathematical skills. However, the research was conducted under certain constraints, such as exogenous (fixed) pricing for the information reveal. In a real-world market, a seller would likely optimize both the content of the "teaser" and the price of the full disclosure simultaneously.
Future research will focus on integrating dynamic price-setting into the model. This will explore whether sellers can further extract profit by fine-tuning costs based on the specific information disclosed in the free teaser, potentially uncovering even deeper layers of human susceptibility to strategic information management.
Summary updated