How to Design Mystery Products That Work
Elgayar, Guhl, Stich & Spann: "Designed Uncertainty in Mystery Products" CRC Discussion Paper No. 565
Pop Mart, the Chinese blind-box company, more than doubled its revenue to RMB 13 billion in 2024. Stitch Fix ships “mystery” clothing bundles to 2.5 million active clients. In food retail, Too Good To Go connects over 100 million registered users with roughly 175,000 business partners through surprise food bags. Mystery products, offerings that deliberately conceal key attributes until after purchase, are no longer a niche curiosity. They represent a massive and rapidly growing market across industries. Yet while some mystery-based offerings scale quickly, others quickly lose momentum. But how should firms design the uncertainty at the heart of these products?
A new CRC TRR 190 Discussion Paper by Alaa Elgayar (HU-Berlin), Daniel Guhl (HU-Berlin), Lucas Stich (University of Würzburg), and Martin Spann (LMU Munich) tackles exactly this question. In “Designed Uncertainty in Mystery Products”, the authors show that the success of mystery products depends not on how much uncertainty firms create, but on how carefully they design it. The authors focus on two design levers that firms can control.
Two levers firms can pull
The first lever is what goes into the mix: the composition of possible outcomes a consumer might receive. Are the potential products similar in perceived value, or is there a clear “best” and “worst” option? The second lever is how much consumers know about the odds. Are the probabilities of each outcome disclosed (a 50/50 chance, say), or are they left vague?
These two decisions, the authors argue, are very relevant: how consumers evaluate products, what they are willing to pay, and which brands win or lose at the market level.
How they tested it
The authors ran two incentive-aligned experiments. In both, participants faced the same core decision: choose one of two fully specified options, choose a mystery option that would turn out to be one of them, or walk away. The composition of the two fully specified options changed across choice tasks, sometimes the options were close in value, sometimes far apart. Half of the participants were told the mystery option would resolve with equal probability; the rest were simply told it could be either, with no probabilities specified.
In the first study, 48 lab participants chose between monetary options with known values and a mystery option whose value was hidden; a clean setting with no product features that could affect decisions based on personal taste.
The second study was a realistic shopping experiment with over 1,000 participants on Prolific. Participants chose between jeans described by brand (Levi’s, Tommy Hilfiger, Calvin Klein, and Wrangler) as well as fit, color, and price. The mystery product hid only the brand, everything else (fit, color, and price) remained visible.
What they found
The mix matters more than the mystery. Consumers don’t react to uncertainty in general, they react to what they might get. When the possible outcomes are similar in value (say, Tommy Hilfiger and Calvin Klein), mystery jeans can actually command a premium over the average of the two brands. But when the set includes a dominant brand like Levi’s alongside a weaker one, consumers discount the mystery product below average, as if they expect to end up with the less desirable option.
Hiding the odds hurts, but only sometimes. Ambiguity (not disclosing probabilities) substantially lowers willingness-to-pay when the possible outcomes differ a lot in value. But when outcomes are close in value, the penalty largely vanishes. In other words, whether to disclose probabilities is not a one-size-fits-all decision, it depends on what’s in the mix.
Under ambiguity, consumers shift their attention. When probabilities are hidden, consumers do not stop choosing the mystery option. Instead, they change how they evaluate it. They place less weight on brand and “mystery” cues, and more weight on tangible, observable features like fit and color. Ambiguity doesn’t make people stop thinking; it changes what they think about.
At the market level, mystery products shift profits. In market simulations under price competition, introducing mystery jeans shifts demand and profits toward participating brands, especially weaker ones, which see substantial gains. Dominant brands gain little from participating directly but may join defensively to prevent erosion of their market position. Non-participating brands consistently lose. Mystery products also turn out to be among the most price-elastic options in the market, making them powerful tools for demand expansion rather than margin extraction.
The bottom line
Mystery products work best when the downside risk is constrained. Firms should compose outcome sets from alternatives that are similar in perceived value, avoid bundling a clearly dominant option with a clearly inferior one, and think carefully about whether to disclose probabilities based on how different the possible outcomes are. The takeaway: it’s not about how much you hide, it’s about how well you design what’s hidden.
Link (pdf): Designed Uncertainty in Mystery Products
