NetLogo User Community Models
Talk of Network
by Cindy (Submitted: 11/15/2012)
## WHAT IS IT?
This model is built based on a journal article by Goldenberg et al. (2001) titled "Talk of the Network: A Complex Systems Look at the Underlying Process of Word-of-Mouth". This model is basically trying to mimic the spread of information through word-of-mouth. How a simple advertisement seen by only a few people can be effectively spread amongst the community members.
In general Word-of-Mouth (w-o-m) can be defined as the sharing of information (opinions or recommendations) through communication among people about products and services (Westbrook, 1987). In business area, a lot of companies are aiming to increase their brands knowledge through various word-of-mouth methods – offline (e.g. face-to-face) as well as online (e.g. through SNS). Of course, advertisement also plays a role in the sharing of information by being the initial contact as well as one of the influencers in getting people to know a brand (Rust and Varki, 1996). But, contrary to popular belief, in long-term, advertisement is not as powerful as w-o-m in contributing to the spread of information (Arndt, 1967; Buttle, 1998; Keller & Berry, 2003; Laborie, 2006). Not only brand knowledge, researchers also found that w-o-m influenced almost 70% of customer’s decision to buy (Balter, 2008).
For this simulation, there will be two kinds of relationship defined – strong ties and weak ties. The differentiation of the relationship is based on “the strength of weak ties” theory by Granovetter (1973). By observing the micro-level differences, a lot of differences can be observed on the macro-level phenomena. According to Granovetter, as contrary to weak ties, strong ties are characterized with a more stable, frequent and intimate interaction. Using this logic, in this simulation, strong ties are understood as those who are closely connected to the individual while weak ties are understood as those who are not directly connected to the individual.
For this simulation, members of the community will be stationary. If the collection of information coming from others with strong and weak ties is strong enough to pass their threshold, they will turn from the state of “unaware” to “informed”. Each member of the community will have its own unique threshold of getting into the informed state. And just as in real life, members will put higher value in the information spread by someone close to them than those who are not directly connected to them.
## HOW IT WORKS
As have been mentioned before, each member has its TRESHOLD. To start off, user can adjust a set of variables which will be constant throughout the simulation - STRONG-TIES, WEAK-TIES and ADVERTISING.
Besides that, each member also has its own STATE - informed "1" or not "0". With the implementation of watch-one procedure, user can observe and learn more about the detail of the process from one of the randomly selected member which is being put into the spotlight. From this member, user can track the NO. OF WEAK TIES which can tell us the number of "informed" others in the community which are not directly connected to the member. The other one is to track the NO. OF STRONG TIES which is the number of "informed" others in the surrounding which are directly connected to the member (neighbors).
The probability of a member to move from not knowing to knowing is defined using a formula stated in the journal article (Goldenberg et al., 2001):
prob = (1 - ((1 - alpha) * ((1 - beta-w) ^ j) * ((1 - beta-s) ^ m)))
The member need to have a prob >= threshold to move from uninformed to informed.
## HOW TO USE IT
As a starter, user can use the sliders provided to adjust the WEAK-TIES, STRONG-TIES and ADVERTISING.
STRONG-TIES: This is to set how strong the effect of others with a strong connection to the member knowing the information towards the probability of the member knowing the information.
After that, user can simply click on the SETUP button then the GO button to begin the simulation.
Members who have moved from "uninformed" to "informed" will be given a blue color code. This will not change until the end of the simulation.
## THINGS TO NOTICE
User can observe how the number of TOTAL INFORMED PEOPLE and TOTAL UNINFORMED PEOPLE changes throughout the simulation. A "slower" speed is recommended to fully observe the changes as this simulation is quite fast.
Besides that, user has the chance to observe how much each of the different factors affects the probability at three different times in the simulation. User can observe which factor affects the probability most at specific time as well as the trend of each factor throughout the simulation (less or more).
## THINGS TO TRY
User can try to play around with the sliders. But for this simulation though, not so much quantitative difference can be found on eye level. User may need to observe the change happening by slowing down the speed of the simulation.
User may also try out the BehaviorSpace prepared to help them to see the complete quantitative data in Microsoft Excel format. The settings are adjusted according to Goldenberg et al. (2001). User may also be interested to change the settings to match with the user's needs.
## EXTENDING THE MODEL
One of the important things to be upgraded is the count of ticks. If it's possible to make it a larger count, maybe user will be able to observe the differences in different conditions.
Apart from that, a consideration towards the content of the information may be an interesting feature to be explored.
## NETLOGO FEATURES
This simulation uses a lot of global variables to be able to show to user a lot of different data. Also observe how several variables are being used cross-procedure.
Apart from that, several variables are being used to set another value to overlap previous ones.
It will be great if fewer variables can be used to handle this simulation.
## CREDITS AND REFERENCES
This simulation as well as several of the set-ups are adapted from Goldenberg J., Libai B. & Muller E., "Talk of the Network: A Complex Systems Look at the Underlying Process of Word-of-Mouth", Marketing Letters 12:3, 211-223, 2001. Kluwer Academic Publishers, The Netherlands.
Balter, Dave (2008), The Word of Mouth Manual, Vol. 2. Boston: Bzz Pubs.
Deitz, Sarah, and Idil Cakim. “Online Influence and the Tech-fluentials,” July 2005: [URL: http://www.efluentials.com/documents/wommaconferencepaperjuly 132005.pdf].
Goldenberg J., Libai B. & Muller E. “Talk of the Network: A Complex Systems Look at the Underlying Process of Word-of-Mouth”. Marketing Letters 12:3, 211-223. 2001. Kluwer Academic Publishers, The Netherlands.
Granovetter MS. (1973). “The Strength of Weak Ties,” American Journal of Sociology, 78(May), 1360-1380.
Keller, Ed, and Jon Berry. The Influentials: One American in Ten Tells the Other Nine How to Vote, Where to Eat, and What to Buy. New York: Free Press, 2003.
Laborie, Jean-Louis. “The Theory Behind Engagement and Integration’s Early Experience Across Media.” Paper presented at ReThink: 52nd Annual Advertising Research Foundation Annual Conference and Expo, March 20-22, 2006: [URL: http://mail.thearf.org/roymorgan/ Engagement/2006.rethink.ARRThe%20Theory.pres.Laborie.pdf].
Westbrook, Robert A. (1987), “Product/Consumption-Based Affective Responses and Postpurchase Processes,” Journal of Marketing Research, 24 (August), 258–70.
(back to the NetLogo User Community Models)