Science for SEO: Marketing Strategy

January 30, 2009

Heat diffusion for Social net marketing

The paper we look at here is called "Mining Social Networks Using Heat Diffusion Processes for Marketing Candidates Selection" and is by Yang, Liu and King from The Chinese University of Hong kong.

Companies have started social networks more and more for WOM promotion, increase bran awareness, attract potential clients and so on. This is openly apparent in Facebook pages, Twitter, collaborative filtering, blogs and many many others. This paper presents a model enabling marketers to make the best use of these networks using the "Heat Diffusion Process" (which is an idea borrow from the field of physics). They have 3 models and 3 algorithms to demonstrate that allow for marketing samples to be collected.

In physics the heat diffusion model states that heat flows from a position of high temperature to one of low temperature.

They show that these methods allow us to select the best marketing candidates using the clustering properties of social networks, the planning of a marketing strategy sequentially in time, and they construct a model that can diffuse negative and positive comments about products and brands to simulate the complex discussions within social networks. They want to use their work to help marketeers and companies to defend themselves against negative comments. The idea is to get a subset of individuals to adopt a new product or service based on a potential network of customers.

The heat diffusion model has a time dependant property which means that it can simulate product adoptions step by step. The selection algorithms can represent the clustering coefficient of real social networks. All users of social networks can diffuse comments that can influence other users. Based on this they say that nodes 1 and 2 represent adopters and the heat reaches nodes 3,4 and 5 as time elapses. The users in the trust circle of other users have a greater influence. Not all of the people in the trust circle however will be contacted about the heat source. Also some users are more active than others in diffusing information. they observe that bad news or negative comments diffuse much faster than other news.

Individuals are selected as seed for heat diffusion. The influence of individuals is based on the number of individuals they influence. The heat vector they construct decides on the amount of heat needed for each source. In order for them to diffuse properly they need a lot of heat. Thermal conductivity is then calculated. It sets the heat diffusion rate. The adoption threshold is then set, as if one consumers heat value is higher than others, they are likely to adopt this product.

If a user doesn't like the product, s/he is allocated negative heat as they will diffuse negative comments. At some point someone in the network will provide different information which might be positive. If that user adopts the product anyway, they will diffuse positive comments. Two defense candidates are then selected and then the negative impact is alleviated.

They conclude:

"So far, our work considers social network as a static network only, and ignores newcomers, new relationships between existing members and the growth of the network’s size. In the future, we plan to consider the evolution property of social networks, and permit our social network to grow at a certain rate"

Why should you care?

This paper shows a new and very different way of analyzing social networks. It gives you a nice opening ti discussions concerning this in a different light. The current methods used in business are not foolproof and such research shows us how simplistic they are and how they can actually be misleading.

November 19, 2008

Optimal Marketing Strategies over Social Networks

"Optimal Marketing Strategies over Social Networks" (www 2008) is a paper written by Jason Hartline (North-Western Uni), Vahab S. Mirrokni (MIT), Mukund Sundararajan (Stanford). It's interesting because it gives an idea of how businesses can use social networks in an effective way to sell their products.

I think the isse at the moment isn't selling products via SN but rather going in unintelligently with the hard sell, and the spamming. The way to sell your products and services is to find interested individuals and to approach them in a friendly, social networking way, about your stuff. Use social networking etiquette.

They looked at influence and revenue maximization. The buyers descision to buy the product is influenced by other buyers in the social network and also by the price of your product. When the buyers were completely symmetric, they could find the optimal marketing strategy in polynomial time.

They looked at approximation algorithms and used the influence-and-exploit strategy. Basically , you give the product for free to a select number of buyers, then you use a "greedy" pricing strategy for the buyers attracted by the influential individuals in their community. They developed set-function maximization techniques to locate the target buyers to influence. When other buyers are influenced from others, it's called "the externality of the transaction". When there is a positive sale, it's called a "positive externality".

We know that users with the most connections have the most influence. However the probability of people buying the product decreases as the marketing strategy progresses. This is why the ultimate method is to give the product away to start with, much like Tivo did.

To start with, you approach individuals and give the itme away, the you go on to the "exploit" stage. You visit buyers in a random sequence and offer them a "myopic price" (optimal pricing for revenue based on the influence of the initial buyers and the buyers who have already bought the item).

They used a simple dynamic programming approach to identify an optimal marketing strategy. Because it's symetric, the order in which you appraoch buyers is irrelevant. "the offered prices are a function only of the number of buyers that have accepted and the number of buyers who have not, as yet, been considered."

They found that the problem of computing the optimal strategy was NP-Hard, even when there was no uncertainty in the input parameters. Using automated things to comoute strategies also involve computing issues such as this basically, you don't have to worry too much about that right now unless this stuff tickles your fancy.

The conclusion:

If a set S of buyers have previously bought, offer the next buyer i price vi(S). Buyer is i, Vi is the value of the buyer and S is a set. Vi(S) is a non-negative number.

This price simultaneously extracts the maximum revenue possible and ensures that the buyer buys and hence exerts influence on future buyers."

There is a lot more detail in the paper and the equations are worth a thousand words as always. Take a look if you're interested in picking at it.