Cumulative Advantage

Is Justin Timberlake a Product of Cumulative Advantage? An interesting series of tests where they show that commercial success is mostly a function of chance and early adoption, not quality. I wonder if the same holds true for software and websites? (Youtube?)

12 replies on “Cumulative Advantage”

I see no reason for software and websites to be immune to the effects. In fact, I bet that nowadays (rather than 5-10 years ago) the snowball builds even faster since word of mouth is so much more efficient. How long did it take YouTube to go from almost nothing to being bough by Google? Not long at all.

However, Google took a fair amount of time to gain dominance. I suspect that if something dramatically better than Google came out tomorrow, it’d take less than a year for it to find an amazingly large audience (assuming it could scale to meet the demands–music doesn’t have this problem but on-line services do).

The article actually said that the same holds especially true for anything exhibiting network effects, and in these markets it is not even news that this is so.

To take Youtube as an example: If you were looking for a video sharing site, which would you pick?
A) The quality site with excellent usability and perfect code, but just a few users (and therefore just a few videos.)
B) The not quite as good site with many users and many videos.

Unless site B is really bad, most users would pick that, thereby increasing the chance of further users also picking it.

Or look at software: An important reason so many people use Microsoft Windows is that “it is what everybody else use”. In this case it works even better. Because of the market share, it is more attractive for application developers, which means more applications, which in turn makes it more attractive for users.

And why did people pick Microsoft software in the beginning? Because it was cheaper than the alternatives, and it was considered good enough.

These are major factors in which of several new competing technologies will get adopted by a mass market:
1) Being good enough
2) Being first
3) Being cheapest

As long as all competing products are good enough, the first to market has a good chance of winning if it gains market share before the others show up. However a latecomer can win by being cheaper – especially if the first had not gained significant market share.

Its true…otherwise youtube might not have gotten so popular. There were other sites just as good (and continue to be as good, if not better being that you can get more content) e.g dailymotion

Yes – in that quality is always available to be purchased after the event. The example here might be Flickr, which was not a masterpiece of programming (I would beg to suggest…) before acquisition, certainly not significantly better than its competitors, but scoring highly on word-of-mouth and fashionability.

But what is “word-of-mouth” and what is a large marketing budget?

Then there’s snobbery, where people choose something because it is not commonly used, irrespective of intrinsic merit. William F. Buckley suggested that if his beloved peanut butter were as expensive as caviar, it would be served at Buckingham Palace. People profess to like a thing precisely because they want to be separate from the hoi polloi. Thus the cult followings of various alternative bands of marginal musicality.

At some point, things become “uncool” because of their popularity. Microsoft used to be perceived as a scrappy independent company. For all of Madonna’s success, she has had to reinvent herself myriad times to stay relevant. Just wait a few years before Google is considered as evil as Microsoft.

For all the talk of randomness in the article, one also has to wonder if there exist influential trendmakers at the key nodes of social networks who tend to put out the first influential signals on what the rest of us should like,

In the end, we are just naked apes, finding ever more complex ways to reach our ultimate bliss, the attention of our fellow apes.

I think this is just due to ‘preferential attachment’, i.e. the likelihood of a graph node getting a new link is proportional to the number of links it already has.

Comments are closed.