Thursday, March 14, 2019

Self organization in Digital Business Ecosystems - Long tail distributions, fluctuations



Extremely fragmented imbalances in the order of the systems participating in the ecosystems can result in long tail distributions. 

The evolving YouTube long tail distributions tell the story of both desirable and not so desirable consequences of the patterns observed in the long tail. 

New entrants can get significant attention (distribution of the pattern near the long tail), more popular stars get even more popular (rich get richer) as a result of network effects of people’s (fans) decision making. It is easy to observe while listening to music or watching an entertainment channel, YouTube directs you towards more of the kind, getting you more. 

Initial popularity does not mean the popularity remains effective. Favorite musicians or movie stars (with their popularity at their peak)  usually remain in people’s list at most a decade (just my observation on how tastes change with more exposure) - that too sharing and enjoying their most popular hits. Meanwhile, someone else takes the spotlight and the show goes on. There are always stars waiting to be recognized among the niche players. This tendency leads to fragmentation over time leading to long tail distribution patterns.  

In politics, unforeseen driving forces towards extremism and radicalization can happen. The Atlantic article on YouTube extremism highlights the autoplay feature of YouTube going a step above recommendations - towards promoting inflammatory content. 

In firm ecosystem distributions - long tail fluctuations can result in catastrophes, and volatility increases - leading to undesirable results. 

Take the popular Amazon’s role in retail. It is clearly spreading its wings cashing on the core digital distribution frameworks it has been operating the book selling business for the convenience of the customer. While entering retail continued to look like a phenomenon expected of an digital venture, establishing stores on ground does not sound the same. This aspect is leading to fragmentation in the long tail of retail stores on ground. Similarly delivering fresh groceries and cooked food is very different from delivering other retail items. This is another case of fragmentation in the long tail distribution patterns for order. 

The logic behind the long tail distributions:

Long tail distributions can be a result of manifestation of power law. It requires verification because not all long tail distributions represent power laws.

Questions arise on the models behind the long tail resulting in power law. In the case of YouTube example discussed above, the preferential attachment model is in play. 

The sand pile and forest fire models account for the volatility and catastrophes that can be caused by Power law distributions due to massive expansions of businesses. 


What does the understanding of models to do with self organization for ecosystems? 

When an entire forest will catch fire on lightening is determined by a critical point or threshold parameters. This point is difficult to predict.  However, the parameters causing the criticality can be understood to plan better. For example, it is a fact that the density of trees in a forest determine the critical state to self organize - by burning down. But business ecosystems can do better. They can determine these critical points caused by interdependencies and interconnections and avoid reaching those thresholds causing failures and catastrophes. 


About order and disorder, entropy: 

More disorder or fragmentation in the long tail is growing uncertainty. It is a measure showing maximum entropy. Grasping the models behind lead to better choices and justification of the order that prevails reaching equilibrium or the disorder that leads to even more fragmentation. 

Conclusion:

Getting customers of retail or consumers of content to long tail sounds desirable for both the businesses and customers, but the consequences are not so desirable for politics, super stars or retail stores alike. Inflammatory content can have adverse affects shifting the order and the health of the society. Too much customer convenience centered, widely expanding business can be creating the ground for catastrophic results for the industry. 

Controlling and handling fluctuations balancing from disorder and chaotic changes is not easy. Perhaps statutory warnings as in the case of cigarette smoking is injurious to health, the consumer of inflammatory content must be warned about where his desires might lead to. Consumers can re-think their options for gratification and encourage on ground local markets. 


References:




Books: 
Linked by Barabasi
Chapter The Seventh Link: Rich get richer 
The Model thinker By Scott E Page - chapters 6 and 12
Chapter 6: Power law distributions, Long tails 
Chapter 12: Entropy: modeling uncertainty 

Networks, Crowds, and Markets: Reasoning about a highly connected world
 by David Easley and Jon Kleinberg     
Chapter 18: