The technology revolution has provided unlimited possibilities for collecting and storing personal data and information. Most of this information is mined from interviews, surveys and other sources with no attention to privacy. To be successful in their marketing endeavor, companies are very proactive and can predict what a customer needs through various targeted marketing campaigns.
To many businesses, the benefits of customer profiling include:
• Personalized marketing campaigns enable them to find and keep more customers, as well as give them better service.
• Maximize revenue from each existing customer and use that marketing data to target new prospects.
• Segment their customers to ensure that they send the right message to the right audience, maximizing limited marketing dollars.
• Dynamic, real-time groups ensure that each campaign addresses key markets, and that target groups always receive the most current, accurate information.
The author recognized that privacy- preserving profiling is something that can be achieved through cryptographic means. But an efficient solution by it requires some element of trust. According to the author three distinct and conflicting privacy and security requirements that must be met in order to adequately perform privacy-preserving profiling are:
* the data must not be revealed
* the classifier must not be revealed
* the classifier must be checked for validity.
A number of classification models for privacy-preserving profiling without revealing it have been proposed. The methods used have the following properties:
* the classification result is only reveal to designated party.
* no information about the classification result is revealed to anyone else.
* rules used for classification can be checked for the presence of certain conditions without revealing the rules e.g. no race is being used.
It is important to note that privacy-preserving profiling is possible by using commutative encryption protocol. This encryption protocol can be used on data item enciphered with multiple encryption keys in any arbitrary order.
Ensuring privacy in Targeted marketing:
It is important to adequately ensure privacy while at the same time perform targeted audience marketing. We should also be mindful that customers are either static or mobile. So the best approach is to combine methods that meet the different requirements for each class. More so, when the identification of the right target audience is the essence of effective marketing. To achieve this, we can comfortably use clustering and cluster analysis.
The cluster analysis consists of 4 distinct critical implementations as illustrated by the author:
* Segmenting the market – consists for example of consumers who are homogeneous in term of the benefits they derived.
* Understanding buyer behavior – works well if consumers are group together in term of common behavior.
* Identifying new product opportunities – set the tone for a highly competitive market
* Selecting test materials- allows for testing various marketing strategies.
?Reducing data – a good representative can easily be identified in homogenized clusters.
Ensuring privacy of mobile users:
The significant use of mobile devices has raised a lot of privacy and security issues especially at it relate to commerce. Marketing on mobile devices are based on what the author called location-based services (LBS). The following are attributes of the location-based services:
* It is to a request usable
* personalized information are delivered at the point of need
* targeting of customers based on advanced knowledge such as profiles and preferences or
* perhaps using their locations
The security and privacy concerns in location based service are
- It may be possible for an adversary to physically locate a person of interest.
- Tracking of individuals is possible which can have adverse consequences.
- The profiles of the mobile customer are often retain the during marketing
- It is difficult to maintain confidentiality because the identity of the user can often be traced to the LBS.
Privacy-Preserving Data mining technique. It deals with the problem of mining data without seeing through secure computations. Two known methods are universally used to accomplished privacy-preserving data mining and these are:
- Perturbation: Individuals have access to their data and only care about the privacy of certain attributes. Although, data security have been proved not to be well established because knowing bounds of data can degrade their security.
- Cryptographic Approach: is more secure but less efficient. Mostly used in situations where there are small number of interested parties that owns large amount of data that are analyzed together.
Conclusion:
Profiling provides the basis for starting what marketers or law enforcement agencies call a "dialogue" with customers or suspects but without regard to the privacy of the customers or persons involved. But privacy-preserving profiling needs more data mining approaches in order to be trusted and to maintain confidentiality.
Reference
Acquisti, A., Gritzalis S., Lambrinoudakis C., & De Capitani di Vimercati S. (Eds.). (2008) Digital privacy: Theory, technologies, and practices. New York: Auerbach Publications.
Marketing Automation: Customer Profiling. Retrieved March 8, 2011 from http://www.netsuite.com/portal/industries/wd/marketing_cus_profiling.shtml
Smirnov-M H. (2007) Data Mining and Marketing. Retrieved March 7, 2011 from http://www.estard.com/data_mining_marketing/data_mining_campaign.asp
• Personalized marketing campaigns enable them to find and keep more customers, as well as give them better service.
• Maximize revenue from each existing customer and use that marketing data to target new prospects.
• Segment their customers to ensure that they send the right message to the right audience, maximizing limited marketing dollars.
• Dynamic, real-time groups ensure that each campaign addresses key markets, and that target groups always receive the most current, accurate information.
The author recognized that privacy- preserving profiling is something that can be achieved through cryptographic means. But an efficient solution by it requires some element of trust. According to the author three distinct and conflicting privacy and security requirements that must be met in order to adequately perform privacy-preserving profiling are:
* the data must not be revealed
* the classifier must not be revealed
* the classifier must be checked for validity.
A number of classification models for privacy-preserving profiling without revealing it have been proposed. The methods used have the following properties:
* the classification result is only reveal to designated party.
* no information about the classification result is revealed to anyone else.
* rules used for classification can be checked for the presence of certain conditions without revealing the rules e.g. no race is being used.
It is important to note that privacy-preserving profiling is possible by using commutative encryption protocol. This encryption protocol can be used on data item enciphered with multiple encryption keys in any arbitrary order.
Ensuring privacy in Targeted marketing:
It is important to adequately ensure privacy while at the same time perform targeted audience marketing. We should also be mindful that customers are either static or mobile. So the best approach is to combine methods that meet the different requirements for each class. More so, when the identification of the right target audience is the essence of effective marketing. To achieve this, we can comfortably use clustering and cluster analysis.
The cluster analysis consists of 4 distinct critical implementations as illustrated by the author:
* Segmenting the market – consists for example of consumers who are homogeneous in term of the benefits they derived.
* Understanding buyer behavior – works well if consumers are group together in term of common behavior.
* Identifying new product opportunities – set the tone for a highly competitive market
* Selecting test materials- allows for testing various marketing strategies.
?Reducing data – a good representative can easily be identified in homogenized clusters.
Ensuring privacy of mobile users:
The significant use of mobile devices has raised a lot of privacy and security issues especially at it relate to commerce. Marketing on mobile devices are based on what the author called location-based services (LBS). The following are attributes of the location-based services:
* It is to a request usable
* personalized information are delivered at the point of need
* targeting of customers based on advanced knowledge such as profiles and preferences or
* perhaps using their locations
The security and privacy concerns in location based service are
- It may be possible for an adversary to physically locate a person of interest.
- Tracking of individuals is possible which can have adverse consequences.
- The profiles of the mobile customer are often retain the during marketing
- It is difficult to maintain confidentiality because the identity of the user can often be traced to the LBS.
Privacy-Preserving Data mining technique. It deals with the problem of mining data without seeing through secure computations. Two known methods are universally used to accomplished privacy-preserving data mining and these are:
- Perturbation: Individuals have access to their data and only care about the privacy of certain attributes. Although, data security have been proved not to be well established because knowing bounds of data can degrade their security.
- Cryptographic Approach: is more secure but less efficient. Mostly used in situations where there are small number of interested parties that owns large amount of data that are analyzed together.
Conclusion:
Profiling provides the basis for starting what marketers or law enforcement agencies call a "dialogue" with customers or suspects but without regard to the privacy of the customers or persons involved. But privacy-preserving profiling needs more data mining approaches in order to be trusted and to maintain confidentiality.
Reference
Acquisti, A., Gritzalis S., Lambrinoudakis C., & De Capitani di Vimercati S. (Eds.). (2008) Digital privacy: Theory, technologies, and practices. New York: Auerbach Publications.
Marketing Automation: Customer Profiling. Retrieved March 8, 2011 from http://www.netsuite.com/portal/industries/wd/marketing_cus_profiling.shtml
Smirnov-M H. (2007) Data Mining and Marketing. Retrieved March 7, 2011 from http://www.estard.com/data_mining_marketing/data_mining_campaign.asp
No comments:
Post a Comment