The snowflake effect


Image from

I like Erik Duval’s concept of the snowflake effect[1] in terms of user interactions with new technologies. He defines this as follows:

In the same way that all snowflakes in a snowstorm are unique, each user has her specific characteristics, restrictions and interests. That is why we speak of a “snowflake effect”, to indicate that, more and more, the aforementioned facilities will be relied upon to realize far-reaching forms of personalization and “mass customization”. This effect will be realized through a hybrid approach with push and pull techniques, in which information is actively requested or searched by the user, but also more and more subtly integrated in his work and learning environment. In this way, a learning environment can be created that is geared to the individual needs of the teacher or student.

He goes on to as the question: What could, a “snowflaked” learning environment look like?

I think there are a lot of synergies here with the way in which I have been using Gibson’s notion of affordances.(Gibson 1979) I argue that technologies have a set of characteristics or affordances, that will mean different things to different people, i.e. each user comes to a particular technology with a unique set of characteristics, these include their personal preferences for how they want to interact with technologies, their skills level, their context, etc.  


Gibson, J. J. (1979). The ecological approach to visual perception. Hillsdale, New Jersey, Lawrence Erlbaum Associated.




[1] See

 for more on this.

19 Responses to “The snowflake effect”

  1. Phil Greaney Says:

    Hmmm… I’ve always been a little bit suspicious of complete individual relativism (and of the ‘You are a beautiful snowflake’ metaphor come to that). I think that there is the potential for users to use technologies in ways that are unexpected and *that* should be considered when facilitating learning and so on. However, there is - as in all disciplines, I think - the propensity for intersubjective agreement, where (in this case) many users use the technology in more or less the same way, based upon convention of the tool’s use; shared experiences; and so on. As such, I’m not sure that this precise metaphor, extreme as it is (having a meaningful ‘uniqueness’ is a grand claim), is helpful as it stands. If there’s another way to suggest the necessity for those facilitating learning to cater for individualistic learning paths, influences of experiences and so on, then I’m completely for that - I think it’s a very good idea. Bu we’re not snowflakes: we often share a lot of ideas, preconceptions, experiences that colour our learning. To forget this is as big an error as it is to suggest we’re completely unique (in a meaningful sense: we are of course unique fundamentally).

  2. Gráinne Says:

    Thanks Phil good points and I partially agree with you - particularly that we tend to have conventional ways of using things and shared experiences - nonetheless I do think we each have our own personal preferences and styles as well and hence think that the concept has some merit for that reason.

  3. Mary Says:

    The technology-rich literacy environment enables each individual to personalize his or her learning. Indeed, an individual’s ability to access the environment and to participate with confidence and competence is influenced by his or her personal preferences, cultural background, and situation. The metaphor of the snowflake, as used by Grainne in this entry resonates for me.

  4. Gráinne Says:

    Yep I agree Mary and the issue of what kinds of digital literacies individuals have, along with their personal preferences for communicating and sharing with others is key.

  5. Erik Duval Says:

    Thanks, Grainne for picking up on our Snowflake Effect work!

    @Phil, it’s true that people are also very similar to one another in many respects, but actually … so are snowflakes! And I would argue that we _are_ all unique.

    In fact, we did some work on the ’snowflake number’, a metric for defining how many items you need to consider before one becomes unique. For instance, there will be other people who read the last book I read (Tonio - see if you understand Dutch). And there may be other people who have read the last two books I read. How many books would you have to go back before I would be unique in my reading? (Our paper at expands on this idea…)

    Anyway, the basic point is that, if you consider enough characteristics, then all learners are unique. In fact, I am not who I was yesterday, so we are all unique at each specific instance. It seems odd to me that we pay more attention to this fact with respect to food (few restaurants serve one dish only), clothes (uniforms are a bit passe, and even those who wear uniforms personalize them), etc, than we do in learning?

    Hope this helps to clarify…

  6. Phil Greaney Says:

    As long as you don’t think that I ignore or refute the idea that individuals have their own backgrounds, experiences and so on, and that they bring them to their learning. Of course, I’m not suggesting otherwise. Rather, I’m suggesting that there is a wonderful set of shared experiences that transcend the notion of the individual, our collective knowledge and so on that we can draw on in order to facilitate learning. We have our own, deeply personal set of experiences with the English language, for example, but we can still often agree a great deal on what a sentence means, including this one. I have to say I find the idea of a ’snowflake’ a little sentimental and nebulous - but perhaps I’ve watched ‘Fight Club’ too often :-)

  7. Gráinne Says:

    LOL Phil! Each to his own ;-)

  8. Gráinne Says:

    Yes would agree Erik we are each unique but also we are connected to others through shared experiences and socio-cultural context. Interesting mix! Anyway that is enough serious conversation I am off to the pub! :-)

  9. Mary Says:

    Grainne, I agree that in many cases an individual’s willingness to share and communicate within an online learning environment are shaped by the individual’s preferences, but I also feel that an individual’s willingness and ability to share may be influenced by situational factors other than learner’s preferences. I agree with Phil, too, when he writes that the set of shared experiences transcend the individual and the situation. A random, but related connection to the metaphor… In Alaska, where there is a lot of snow, people have many generated many words for snow and they describe snow in more sophisticated and complex ways than people do in Florida or Spain, for example…. Those who interact within technology-rich and literacy-rich environments may also generate language for and gain a more sophisticated understanding of how to use the affordances of the online learning environment and to share and communicate. This gets back to your final connection with Erik’s post… “We are each unique but also we are connected through shared experiences and socio-cultural context.” … and the network….

  10. Phil Greaney Says:

    Thanks Erik - we must have written more or less simultaneously, since at the time of writing your post wasn’t present. I’ve read your links (the first in translation) - thanks for these and your comments. You write, in ‘The Snowflake Number’ abstract:

    The better a web application adapts to these characteristics and interests [of the human individual], the more relevant and useful it is.

    The thing is - I agree with you, or want to agree. But I can’t commit fully. I *believe* in personalisation, in the growing attention paid to accommodating difference amongst learners. But I wonder - do we really want a web application, or suite of applications, that are fundamentally different, used within the context of a course of study? How do those facilitating learning, including fellow learners, support the potentially infinite number of configurations that increasingly complex web applications might provide? How do we invite learners to collaborate in online environments where fundamental differences and incompatibilities in their choice and setup of software exist?

    One answer would be that the web applications really don’t differ too much difference to be a problem. You can change the colour of menus, customise the ‘look and feel’ without altering basic functionality. But then, that is hardly evidence of web applications reflecting the kinds of innate individuality that the snowflake metaphor implies. To extend the idea, it suggests that *no* web application or course of study where a qualification is offered can ever be entirely personalised or reflect the individual identity. (This raises the question, of course, of the value of qualifications and the role of ‘courses of study’; but that’s for a different conversation.)

    I remember, when I first began to help design courses, and assess the feedback on courses from learners, being surprised at how many learners preferred the reduced complexity of the virtual learning environment tools when compared to freely available, widely-used online tools. As a result, I would question your assumption above, or any like it. Again - this is counter to my understanding and assumptions - I thought learners might want a tool that offered greater levels of functionality (within which personalisation was included).

    Forgive me if I sound like a Luddite; or if I appear to reject the notion of learner’s individual approach; or if I appear to seem to be advocating a ‘one size fits all’ approach. I intend none of these. I merely suggest that such fundamental assumptions as these be questioned and considered before we forge ahead.

    Grainne - I hoped you enjoyed the pub!

    Mary - thanks for that interesting example, and in keeping within the theme, too!

  11. Phil Greaney Says:

    Me again. I shall trouble your blog no more, Grainne, save to *read* the comments here. I’ve been reading this:

    “Standardized Uniqueness: Oxymoron or Vision of the Future?”

    Ok - I get it. I like it, too. I even feel a little bit warmer towards the ’snowflake’ metaphor, now I understand that standards underpin it (we are all unique but share certain similarities which makes connections possible).

    I wish you all the best with your project, Erik - I’ll keep track of its developments online.

  12. Gráinne Says:

    Excellent Phil Erik sure loves his standards! ;-)

  13. Erik Duval Says:

    Quick reaction to Phil: personalization is not an excuse for increased complexity!

    A trivial example: twitter is extremely simple. Yet, my twitter feed is probably quite unique - nobody else gets the same mix…

    In any case: thanks for the feedback. Much appreciated!

  14. Gráinne Says:

    Good come back Erik as I would expect ;-)

  15. Phil Greaney Says:

    Erik - you know, if you specify an interest in a subject (as you might in an educational course) - say, educational technology, history, English, etc - then I bet your Twitter stream is not as unique as you might think in this respect. Others interested in same subject will share a surprisingly high degree of agreement on who to follow I should think. Just as there’s a ‘canon’ of texts, there is now a ‘canon’ of people to follow, (especially following Twitter’s ‘Who to follow’ function.)

    Conversely, whilst Twitter is technically simple tool, there are a variety of ways to access it - web-based, mobile, apps and a host of clients, not withstanding the interactions with other online services. This is potentially problematic for those offering technical support. I say ‘ostensibly’ because many still find Twitter in any form quite complex, not least in its cultural terms, given its specific conventions. I think that often has to do with its basis in human interaction, which is far from simple :-)

    Again, all the best with your project

  16. Martyn Cooper Says:

    Interesting discussion but no one has picked up on Grainne’s attempt to link the Snowflake Affect to Gibson’s affordances. I think it is flawed (sorry Grainne). I haven’t read Gibson for over 20 years but will press on without checking my sources in my normal “paragraph heading” knowledge way. I recall that he expresses affordances as a property of the object not the user other than it is dependent of common expectations/experiences across users. Using a simple architectural example: we expect to pull on a door handle but push a door plate. The affordance as to whether to push or to pull is ascribed to the door design. However if an extreme “snowflake” came along that had say electro-magnets for hands they might find it easier to pull on the plate. However I would argue that does not change the affordance of a door with a plate because that results from a common expectation of most users not the exception. Even the guy with the magnetic hands would learn that most doors with plates need be pushed.

    Thus if you want to design for a high diversity of users (snowflakes) you need to deliberately reduce the level of affordances in your design so that different users can discover their own way of using it without it being dictated to them by the designer. I think this is the antithesis of what Gibson argued.

    Of course as in most, if not all, things in life, a balance is required. If we are concerned with utility and usability we need to exploit affordances to promote making the utility obvious and the learning of its use quick and easy but to cater for diversity of users’ expectations and objectives we need to allow personalisation. However is this balance best achieved by a masterpiece of universal design or a framework that allows for diverse “widgets”, say, to meet the diversity of user needs.

    Anyway that waffle contains the one point I was trying to make. To my view, affordances negatively correlate with degree of personalisation.

  17. Gráinne Says:

    See I can do controversial! ;-) I partly agree with your points but still think there is a connection between the two concepts. I am a great fan as you know of Gibson’s concepts applied to how different individuals interact with technology. We are all unique hence the snowflake idea but as Phil says we also conform to certain patterns of behaviour based on culture context and discipline.

  18. Martyn Cooper Says:

    To my view deliberate affordances only exist because of shared cultural conditioning!

    BTW - sure you can be more controversial than that! ;)

  19. Gráinne Says:

    Stop taunting me Cooper!!

Leave a Reply