1
Real Estate Customer Relationship Management
using Data Mining Techniques
Tianya Hou and Andy K.D. WONG (852) 27667805
[email protected] and [email protected]
Department of Building and Real Estate, The Hong Kong Polytechnic University
Hung Hom, Kowloon, Hong Kong Abstract
The development of real estate market is undergoing three stages: from product-oriented to market-oriented, and finally to customer-oriented. Incorporation of customer relationship management (CRM) system into real estate enterprises would help explain and predict the behaviors of real estate decision-makers, which would benefit enterprises in fighting for market shares and winning customers in the fierce competition. However the large amount of the complicated data produced from real estate CRM system has already outpaced the digestion of human beings, and the traditional statistics methods cannot fulfill the requirements of the analysis. To tackle this problem, data mining techniques are proposed to integrate into real estate CRM system. Data mining techniques would help real estate enterprises identify the taste and the preference of real estate purchasers. According to that information, strategies about the land location, the apartment type, and the marketing tools will be set up. This paper firstly presents a review of some CRM concepts relevant to real estate and the corresponding data mining applications in this area, and suggests how these can expand real estate study. Secondly, it provides an example: the customer housing demand analysis using data mining technique. Factors which most affect customer housing demand have been identified. Some implications about developing strategies and market investigation for the real estate firms are inferred, in order to improve the sales turnover rates and reduce the market cost.
Key words: Real Estate Market, Customer Relationship Management (CRM), Data Mining
1. Introduction
The development of real estate market is undergoing three stages: from product-oriented to market-oriented, and finally to customer-oriented (叶开). With the size of the housing, real estate, and mortgage-related industry expanding and the high market competition, it is critical that traders use the best possible information available to improve backend operations and sharpen their customer offerings in real estate business. So the incorporation of customer relationship management (CRM) system into real estate enterprises must be enforced.
2
communications, tracking sales, and developing marketing strategies, where generates a huge amount of data to outpace the digestion of human beings. Data alone is meaningless to most people. But when it’s used to create interesting products for consumers and agents, it can add value to a business. Data mining can enable home buyers and sellers to access more home information that previously was either unavailable or locked up in real estate business. Effective and comprehensive market information can enable the stakeholders stand out from the competition and give the consumers what they are looking for.
The rest of this paper is organized as follows. In Section 2, this paper firstly presents a review of some CRM concepts relevant to real estate and the corresponding data mining applications in this area, and suggests how these can expand real estate study. In Section 3, it provides an example: the customer housing demand analysis using data mining technique. Factors which most affect customer housing demand have been identified. Some implications about developing strategies and market investigation for the real estate firms are inferred, in order to improve the sales turnover rates. Finally, the prospect about application of data mining techniques in real estate CRM are made in Section 4.
2. Customer Relationship Management
Some real estate organizations are still asking “why CRM?”. Unlike other industries (pharmaceuticals and financial services, for example) where sales support and relationship management are ubiquitous, the value proposition for CRM systems in real estate is not well understood. This is especially odd since real estate is traditionally relationship and sales driven, built on commission incentives, having complex referral models and long sales cycles. The negotiations are complex and resulting transactions are often unique. It operates in a high competitive environment where retention of tenants and successful sales is directly correlated to asset value and rate of return.
With an abundance of projects to invest in, buyers have vast choices to make their real estate investment decisions. For the developers, the main challenge is to identify the market segment through the marketing campaigns with sustainable follow-up efforts to build strong pipeline and convert the footfalls into customers. Customer servicing and satisfaction can be a key differentiator for the real estate developers, where strategic technologies like customer relationship management (CRM) can play an important role in this.
CRM is a business approach that integrates people, processes and technology to maximize the relations of an organization with all types of customers (Chen and Popovich). It entails acquiring and developing knowledge about one’s customers and using this information across the various touch points to balance revenue and profits with maximum customer satisfaction. Leveraging data from CRM and external list data, target campaigns based on numerous customer attributes can be conducted. CRM helps in improving segmentation scalability for targeting millions of prospects. This also helps in enhancing marketing productivity, increasing conversion rates and
3 incidentally reduces campaign execution costs.
Because of the rapid process of information technology, the amount of information stored in real estate CRM is rapidly increasing. These huge databases are often represented by high-dimensional feature vectors. Finding the valuable information hidden in those databases and identifying appropriate models is a difficult task. All the problems can be solved by data mining.
3. Data Mining and its application in real estate industry.
Data mining is the process of exploration and analysis, by automatic or semi-automatic means, of large quantities of data in order to discover meaningful patterns and rules. It combines techniques from machine learning, pattern recognition, statistics, database theory, and visualization to extract concepts, concept interrelations, and interesting patterns automatically from large corporate databases. Two primary functions of data mining are: prediction, which involves finding unknown values/relationships/patterns from known values; and description, which provides interpretation of a large database (Guo.).
The application of data mining technology has become more widely in real estate industry. The typical uses of data mining technology in real estate industry are shown in Table 1.
Table1 The application of data mining technology in real estate industry Application
type
Application purpose Application methods
Authors Property
assessment
Evaluate the house for many factors such as the
residential location, transportation,
surroundings, public institutions, and so on.
Regression Hedonic analysis Artificial neural network Leontion. et al. Carlos. Association analysis based on customer segmentation Analyze customers’ characteristics for different customer segmentation, based on those results predict the customers’ purchase intention. Decision tree Association analysis 吉同路等 北京大视野社 会经济调查有 限公司 Customer loyalty analysis
Judge customers’ loyalty base on their attributes such as residential zone, age, income, gender, education and so on Decision tree 李莉 Residential mortgage default analysis
Provide some advices to the risk management department on how to prevent the mortgage default based on different
Artificial neural network Clustering 岑希 Cunha. and Agard.
4 risk class Housing trading and matching system
Based on housing trading and matching system, customers can obtain the desired type of housing information as soon as possible Nearest neighbour algorithm Chang and Chu Building and apartment layout analysis
Find out the desired building from a large amount buildings in different condition Clustering Decision tree Bayesian analysis Artificial neural network 王洋 卫易辰
4.Case study: customer demand analysis of high-rise housing. 4.1 Data
In order to understand the consumer’s decision when choosing high-rise houses, and which of the factors individual customers prefer in combination with one another. A questionnaire survey was conducted, and each consumer was asked which factors they valued most when buying a high-rise house, and each of them can only choose 9 factors. As a result, a data set that list the factors preferred by 1079 customers was produced. The possible factors affecting the sale of high-rise houses and their abbreviations are represented as shown in Table 2.
Table 2 The factors affecting the sale of high-rise houses
Factors Abbreviations Regional plate RP Convenient transportation CT Supporting facilities SF Appreciation potential AP Education facilities EF Large district LD Low price LP Free area FA Apartment layout AL Brand developers BD Landscape vision LV Property management PM Hospital H School S Kindergarten K Bank B
5
Supermarket S
Convenient store CS
Restaurant R
Public garden PG
Chess and card room CCR
Senior citizen activity centre SCAC Children recreation facilities CRF
Sports ground SG
Open air theater OAT
Swimming pool SP
Club C
There are 2 columns in the data sets, shown in Table 3. Table 3 The final data style
Name Model Role Measurement level Description
RS ID Nominal Respondents serials
AF Target Nominal Affecting factors
The data set has over 3000 rows. Each row of the data set represents a customer-factors combination. Therefore, a single customer can have multiple rows in the data set, each row representing one of the factors he or she think very important for value the houses.
4.2 Association Analysis
Association analysis is conducted to do customer’s high-rise housing demand analysis. It will produce thousands of rules. Because of the limitation of space, not all the rules are included in our studies. Some interesting association rules are shown in Table 4. Table 4 Some interesting association rules
Association Rules Support (%) Confidence (%) Lift
CS&FPM→LP 4.36 21.56 1.73
SCAC&CCR→H&CT 5.01 58.70 1.64
CS&S→LP 4.17 20.27 1.63
SP&SG→AP 4.73 31.48 1.60
SCAC&CT→FPM&H 6.68 19.83 1.56
The table contains information for each rule. Consider the rule A→B: Support of A→B is the probability that a customer prefers both A and B. Confidence of A→B is the probability that a customer prefer B given that the customer prefer A. lift of A→B is a measure of strength of the association. If the Lift =2 for the rule A→B, then a customer preferring A is twice as likely to prefer B than a customer chosen at random. Take the rule CS&FPM→LP for example, 4.36% of customers prefer the house which is nearby convenient stores and farm production markets with the low price. 21.56% of customers like the house near to convenient stores and farm production markets care more about the price. The lift of the relationship CS&FPM→LP is 1.73.
6
Therefore, if taking low price as a selling point, the customers like convenient stores and farm production markets are more likely to pay money to buy the house than the customers chosen at random. Other association rules can be explained in a similar way.
Large amount of data already outpace the digestion of the human being, In this case study, over 3000 rows with 28 variables are included. It is impossible to extract some useful information from that data just by experts’ experiences and statistics methods. Association analysis can assist us to solve that problem, broaden the analysts’ thinking mode, and produce some strategies for reference.
5.Conclusion
With the rapidly growing data size in real estate industry when conducting customer relationship management, real estate enterprises strongly need a powerful data analysis tools to transfer those “huge data” into “valuable knowledge”. The traditional statistics methods can not fulfill this requirement yet. Data mining techniques can compensate the limitations of statistics. Data mining can assistant real estate enterprises to identify the taste and the preference of real estate purchasers, to broaden the analysts’ thinking mode, and to provide some strategies for reference at the same time.
References
叶开。房地产企业的客户细分战略。
Chen I.J. and Popovich K.. Understanding customer relationship management (CRM): People, process and technology, Business Process Management Journal, Vol. 9 Iss: 5, pp.672 – 688, 2003.
Guo L.J.. Applying Data Mining Techniques in Property/Casualty Insurance. Chapter 8 of Industry Applications of Data Mining.
Leontion. et al.. Need a Home? Start the Data Mining!- A Data Mining Application in Weka on Real Estate Market.
Carlos.. A comparison of data mining methods for mass real estate appraisal. 吉同路等。住宅与房地产电子政务中数据挖掘的应用研究。
北京大视野社会经济调查有限公司·房地产行业研究咨询部,房地产行业客户 信息的数据挖掘
李莉。基于决策树数据挖掘技术在地产营销中的应用。 岑希。数据挖掘技术在房地产行业的应用综述 。
Cunha. and Agard.. Business process reengineering with data mining in real estate credict attribution: a case study.
Chang. and Chu.. Applying Data Mining and XML Technology to Build a Web-Based House Trading and Matching System.
王洋。基于数据挖掘聚类分析的房地产市场信息处理技术的应用 卫易辰。数据挖掘技术在房地产户型选择分析中的应用