Network effects…

Is Justin Timberlake a Product of Cumulative Advantage?, asks Duncan Watts in the New York Times magazine. His research suggests that market success is hugely affected by a network effect, where people’s preferences are a product of the preferences of others rather than being independent.

He did an experiment in which people were asked to evaluate unknown music tracks under different conditions. Some did this without any information about their peers’ choices. The rest were told what tracks other people were downloading. In the latter circumstances, songs polarised more strongly into popular and unpopular than in the first group. More intriguingly, the second group was subdivided into eight subgroups and each subgroup had quite different favourite songs. Here’s some of Watts’ analysis of this:

The impact of a listener’s own reactions is easily overwhelmed by his or her reactions to others. The song “Lockdown,” by 52metro, for example, ranked 26th out of 48 in quality; yet it was the No. 1 song in one social-influence world, and 40th in another. Overall, a song in the Top 5 in terms of quality had only a 50 percent chance of finishing in the Top 5 of success.

In our artificial market, therefore, social influence played as large a role in determining the market share of successful songs as differences in quality. It’s a simple result to state, but it has a surprisingly deep consequence. Because the long-run success of a song depends so sensitively on the decisions of a few early-arriving individuals, whose choices are subsequently amplified and eventually locked in by the cumulative-advantage process, and because the particular individuals who play this important role are chosen randomly and may make different decisions from one moment to the next, the resulting unpredictably is inherent to the nature of the market. It cannot be eliminated either by accumulating more information — about people or songs — or by developing fancier prediction algorithms, any more than you can repeatedly roll sixes no matter how carefully you try to throw the die.

I was intrigued by Scott Karp’s interpretation of this:

All of a sudden it’s crystal clear what Web 2.0 really is — the greatest platform ever for harnessing randomly imitative social behavior. Before Web 2.0, achieving utterly arbitrary results took time and effort. Now, with platforms like Digg, we can get nowhere in a fraction of the time it used to take.

WOW — I am humbled and awestruck by the power of technology, and the power of randomly socialized human beings to snuff out each others’ critical faculties and personal tastes.

I don’t agree with that interpretation. I think it rests on a certain assumption that our intelligence is individual and not social. Karp appears to regard imitation as a mark of dumbness.

But if you think of the eight sub-communities as forms of collective intelligence, they each come up with distinctive preferences from each other: so at that level, they do not imitate each other… so I wonder if Karp would then admit that these subgroups demonstrate a kind of collective intelligence?

I also suspect there’s a distinction to be drawn between unpredictable and random but my brain hurts too much this morning to try and explore that one. If Dave Snowden‘s listening, maybe he could help?

Big Hat tip to AdPulp for spotting both items.

(There’s lots of interesting comments on Scott’s blog… pity the NYT has missed the chance to host any kind of online discussion, though it does flag an item about colourful men’s underwear)

5 thoughts on “Network effects…

  1. Johnnie Moore

    Top post.

    I too had this pointed out to me but by Gareth (https://garethkay.typepad.com) and think Duncan’s onto something really important (but not far away from what I’ve been herding on about).

    Think you’ve hit the nail on the head when you show how the journo’s ideas about copying are what let him down…a big cultural assumption here…

    M

    —–

    Cheers Mark. Yes, there’s nothing wrong with copying.. in Improv, lots of people say go for the ordinary, stop trying to be different or clever. The cumulative results of copying (which incorporates slight mutation) can be splendid. Case study: life.

    Reply
  2. Graham Hill

    Johnnie

    In Watt’s experiment, the first people to select future hit songs (in the group that knows how many times a song has been selected) do so at random.

    Through the cumulative actions of others who also select the same songs, some of these songs emerge as runaway winners whilst the rest lose out. Call it cumulative advantage or the network effect.

    The emergent nature of the winners is highly dependent upon the initial starting selections, which as we have seen, are made at random. The net result is unpredictable hits. Just like the weather.

    The initial selectors are the musical butterflys that cause the future storm of hits.

    Easy innit.

    Graham Hill

    Reply
  3. Johnnie Moore

    Johnnie,

    Great post.

    One thing that I am wary of is introducing ideas of “collective intelligence”, simply because I think that the research Watts has done is more applicable to culture markets, as opposed to information. I think collective intelligence in something like wikipedia, or software design is a different ball of was than markets that are ostensibly governed by taste.

    Culture as consensus differs from knowledge that emerges from sharing information (linux). One of the problems with Karp is that he conflates the two.

    —–

    Dino: I agree there’s a distinction between collective intelligence as manifested at wikipedia compared to culture markets. Not least, the presence of some central artefact.

    I’d also say that the networks of people influencing each other is still a kind of collective intel… albeit more diffuse. The thinking of the group is clearly being done in ways that combine thoughts/ideas.

    I suppose “intelligence” is sometimes a loaded word with an implication (by some) that its output must be of some stated “quality” to count.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.