The Serendipity Society - Artificial Serendipity

Overcoming the Paradox of Artificial Serendipity

Intended, Afforded and Experienced Serendipity

This post is based on

Smets, A. (2025). Intended, afforded, and experienced serendipity: overcoming the paradox of artificial serendipity. Ethics and Information Technology, 27(3), 33. https://link.springer.com/article/10.1007/s10676-025-09841-6

Serendipity is all around in digital systems. From music streaming services and public service algorithms to news recommenders, social media feeds, and academic panels; the term keeps showing up. It beholds surprise, diversity, discovery, or delight. Designers want to build it, researchers want to measure it, and public discourse is increasingly calling for it. But what we rarely admit: everyone seems to mean something different.

In this Ethics and Information Technology article, I propose a call for clarity. If we want serendipity to be more than a feel-good feature or a vague value, we need to ask sharper questions. What kind of serendipity are we talking about? For whom is it serendipitous? And why do we care about it?

Rather than treating serendipity as a single, fixed concept, I thus propose breaking it down into three interconnected dimensions:

Intentions: the motivations or objectives behind why designers, developers, or institutions aim to foster serendipity in digital systems.
Affordances: the structures of a system that shape how serendipity can emerge, such as interface elements, content organization, and navigational pathways.
Experiences: the user’s situated and subjective interpretation of an encounter as serendipitous, influenced by personal context, timing, and expectations.

Smets, A. (2025). Intended, Afforded, and Experienced Serendipity.

Why? Serendipity is rarely 'just' that

We often take it for granted, but why design for serendipity in the first place? Serendipity is typically framed as something we all seem to want more of in (digital) environments, but the reasons behind designing for it vary widely. In my research, I identified four distinct intentions for designing for serendipity:

Serendipity as ideal: the designer values the valuable outcome recognized by the user. This reflects an altruistic intent, designing with the goal to enrich individual users’ knowledge, creativity, or perspective (e.g., in libraries or education).
Serendipity as common good: the designer is motivated not by the individual benefit, but by the societal impact resulting from many individuals experiencing serendipity—for instance, to support democratic reflection or cultural cohesion.
Serendipity as mediator: the designer values the consequence of the user experiencing serendipity, such as increased satisfaction or commercial return. The serendipitous experience is instrumental in achieving a separate objective.
Serendipity as feature: the designer deems serendipity essential for the functioning of the environment: it is part of the core design proposition. The user seeks out the environment because of the promise of experiencing serendipity (e.g., co-working spaces, personalized discovery tools).

These intents do need to be made explicit. If we don’t clearly articulate why we are designing for serendipity, then efforts to build systems or evaluate their impact risk becoming directionless. After all, how can we meaningfully design for or evaluate serendipity if we don’t know what purpose it’s meant to serve? Making these design intents explicit is essential, not just for transparency, but to ensure that the systems we build are actually aligned with the kinds of outcomes we care about. Moreover, it also allows us to build cumulative knowledge about what serendipity means and how it functions in different contexts.

How? Serendipity beyond the algorithm

How can you even design serendipity, that's a paradox, right? Indeed, you cannot design serendipity itself, but you can design for the conditions that make it more likely to occur. In the context of online services, the design interventions that shape serendipity don’t just relate to algorithms, it’s shaped just as much by the interface, the content structure, and the pathways users could follow. In the context of recommender systems, we proposed a feature repository that identifies which elements like content metadata, navigation structure, and interface layout can make serendipitous encounters more likely.

What? Serendipity is a user experience

Ultimately, serendipity is a user experience. We explored how users recognize and make sense of serendipitous encounters. We found that experienced serendipity exists in many flavors. And importantly: what feels serendipitous to one person might feel irrelevant or disruptive to another. System designers need to account for this variability; not just in how serendipity is experienced, but also in how systems shape the potential for such experiences. Designing for serendipity, if done carelessly, can lead to unintended consequences: users may be distracted, overwhelmed, or nudged away from their goals. Thinking about how to design for serendipity means reflecting not only on your intentions, but also on how these might translate into meaningful, and contextually appropriate, user experiences.

Why this matters

I hope this proposal helps pracitioners and academics to actually work with serendipity. If you're designing for serendipity, reflect first on your intentions: What kind of encounters do you want to make possible, and why? Then consider what affordance features your system needs to support those goals, and how this might actually be experienced by users, both in intended ways and through unintended consequences.

If you're studying or measuring serendipity, avoid overgeneralizations. Be clear about which dimension you're investigating and what they mean: intentions, affordances, or experiences. This kind of differentiation helps prevent conceptual ambiguity and keeps us focused on the purposes, structures, or experiences. Concretely, here are three questions to ask:

What is the objective of the design for serendipity, and for whom? Clarify whose goals are being prioritized, and what kinds of value you hope to generate through serendipitous encounters.
What specifies experienced serendipity in this context, and what are potential negative outcomes? Make sure your definition is context-aware and includes possible downsides, like distraction or misalignment with user goals.
How can we put this into practice? What specific design interventions can you do, and how will you evaluate if they succeed? Make sure to consider how intentions, affordances, and experiences continuously shape and constrain one another. Understanding serendipity means attending to this dynamic interplay.

To make sense of serendipity, and to meaningfully engage with it, I believe we need to ask these questions. I'm curious to hear your answers.

Author bio

Annelien Smets is a Research Professor at the Department of Communication Studies at Vrije Universiteit Brussel, Belgium. Her research focuses on serendipity in recommender systems, and the value of serendipity in digital platforms.