Showing posts with label models. Show all posts
Showing posts with label models. Show all posts

Wednesday, January 08, 2014

ADV vs WOM: Models for marketers




Word of mouth, or viva voce, is the passing of information from person to person by oral communication, which could be as simple as telling someone the time of day. Interpersonal influence and word-of-mouth (WOM) are ranked as the most important information source when a consumer is making a purchase decision. These influences are very important, especially in industries whose intangible products are difficult to evaluate prior to their consumption. Furthermore, as WOM becomes digital in the 21st century, the large-scale, anonymous nature of the Web induces new ways of capturing, analyzing, interpreting, and managing the influence that one consumer may have on another.

Let's now assume that we are managers who wish to understand the possible effectiveness of traditional advertising (ADV) vs WOM for promoting a product or service targeting a specific target segment.By comparing ADV and WOM we can better understand the relative effect of these two methods as well as their effectiveness, which is usually higher for WOM, at least in the short term. For ADV effectiveness, a more intense advertising effort is most of the times needed to reach the same level of informed people than for WOM. WOM is considered to be growing in importance due to the increase in digital social networking. This digital version of WOM is considered to provide a wealth of new possibilities to reach market segments that would have been otherwise hard to access.

ADV



Most of the models used to explain the effects of ADcampaigns have attempted to provide analytical mathematical expressions for the phenomenon (Bass, 2007). The intensity and duration of the campaign memory effects and number of broadcasted messages are considered as parameters. ADcan be described as follows: a message is broadcast to a target population of individuals. The targets receive the message and form an opinion: they can either accept the information content or ignore it. A number of models have been proposed to explain or simulate the process of opinion formation in a group of individuals. Most of these have been devised by using simplified versions of well known physical models used to describe magnetic properties in a material. The analogy is due to the similarity of the situations: a magnetic material can be seen, in a rough approximation, as an ensemble of elementary magnets, each one having one of the two possible magnetizationsalso called spins: positive or negative. The numbeand the distribution of the spins is influenced by the temperature T of the material and can be influenced by the strength of an external magnetic field H (Feynman1963).

Among the many possible models the ‘Magnetic Eden Model’ (MEM) is a leader, originally proposed by Eden,1961. Lets assume that a certain number N of elements in a material can be magnetized by assuming a spin ±1. A randomly chosen element starts the process. At each point in time all the neighbors (elements directly connected to a magnetized one) can be magnetized. The global energy E of a configuration of spins is given by:


where Si = ±1 indicates the orientation of the spin for the magnetized element,  J > 0 is coupling constant (representing the strength of the interaction between neighbors), H >0 is the external magnetic field. 

The probability for a new spin to be added to the set of  magnetized elements is actually proportional to exp(−DE/T), where DE is the energy change involved and T is the absolute temperature of the material. During each step, alperimeter elements are considered and the probabilities of adding a new spin (either up or down) to each site must be evaluated. This can done by employing a Monte Carlo simulation method: all probabilities are first computed and normalized then the new element and the orientation of the new spin determined. The process ends when all the elements of the material have been magnetized. A characteristic of this model is that once a spin is set, it is frozen’ and cannot change.

Applying this model to an advertising process, it is relatively easy to interpret the whole process as a single campaign -at the end a certain number of target individuals will have absorbed’ the message- the magnetic field H is a measure of the intensity of the campaign, and the temperature T is a measure of the social cohesion of the target group (or responsiveness to advertising messages), a spin S = +1 means an individual has accepted the advertising message, S = -1 means that the message was ignored (or refused). Once the campaign is ended, the opinions of the target populatioare set and we may assume that they remain stable until the next campaign.

WOM


WOM resembles to an epidemiological infection process.More specifically, a population of actors is susceptible to an advertising infection’. 

During the unfolding of the process, a number of actors become infected while some recover and become immune by forgetting the information received.  The mathematical models describing an epidemiological infection process are well known (Bailey, 1975). WOM was studied by using an epidemiological diffusion model. This was built by considering the infection cycle in an individual(actor). The actor is first considered susceptible (S), i.e. able and ready to receive a piece of information. Then, if reached by a message it becomes ‘infected’ (I) and is considered as such for a certain period of time. Finally, the individual can recover (R) by forgetting what he received or even be considered as (D) or "dead" as regards obedience to the advertising message.

The mathematical treatment is well known and consists of a system of differential equations that can be solved and produce curves describing the results of the infection. These are mostly s-shaped curves belonging to the family of logistic curves. Traditionally a perfect mixing is assumed: all individuals are equally able to infect all others, the contacts between them are ignored or,better, considered as having a random distribution. When dealing with WOM, however, the diffusion process develops by using the relationships among the social network formed by the individuals as infection channels.


Conclusion

A Nielsen Online Global Consumers Research three years ago found out that:
  1. 14% ONLY of consumers trust ads 
  2. 56% of consumers avoid buying products from companies who advertise too much
  3. 65% of consumers feel they are constantly bombarded with too much advertising
  4. 78% of consumers trust recommendations from other consumers
That means, traditional advertising is almost becoming an antiquated way of marketing. But don’t despair; social media marketing is here to help make some distinctions. We are in the 21st century and marketing is evolving with relative flexibility.

"The only thing we know about the future is that it will be different."
Peter Drucker