Marketing strategies evaluation based on big data analysis: a CLUSTERING-MCDM approach

Figures & data

Figure 1. Data mining techniques (Source created by authors).

Figure 2. Different approaches to BWM (Source created by authors).

Table 1. Previous related researches.

Researcher(s)/year	Target(s)	Tool(s)	Result(s)
(Hiziroglu, 2013)	Reviewing the application of soft computing in segmentation	Soft computing	Insufficient ability of these studies togenerate knowledge
(Julio & Eugenia, 2014)	Proposing a segmentation framework to calculate the customer life time value, the current value and the client loyalty	CLV, CE, LTV calculation	Evidencing that customers grouping is a good way to identify them and offering a comfortable treatment to keep them
(Kahreh, Tive, Babania, & Hesan, 2014)	Proposing a framework to measure the value of each segment concentration on bank sector	CLV calculation	Segmenting the customers into six groups for 2005–2008
(Fahad et al., 2014)	Introducing concepts and algorithms related to clustering and providing a concise survey and comparison	SPSS, Hadoop, Distributed FS, Parallel system	Lack of a method in all big data environment
(Wang, Miao, Zhao, Jin, & Chen, 2016)	Proposing a novel Bi clustering-based market segmentation method using customer pain points	Biclustering Technique, BCBimax algorithm	Assisting the companies to take full advantages of customers’ knowledge to develop products
(Cebeci & Yildiz, 2015)	Comparing the performance of the k-means and fuzzy c-means clustering	Statistical software	Faster performance of k-means, equal clustering failure, finding no algorithm to be the best in all cases
(Sajjadi, Firuzabadi, Amiri, & Sadaghiani, 2015)	Clustering and ranking customers using RFM model and making RFMLD model	CLV, RFM modelling	Developing a new model called RFMLD applying five indexes.
(Ansari & Riasi, 2016)	Combing the genetic algorithm and fuzzy c-means method in order to cluster the customers of steel industry based on an RFML model	Genetic algorithm, fuzzy c-means, RFML modelling	Classifying the customers into two clusters
(Mohammadian & Makhani, 2016)	Reviewing on the methods to gather and analyze data to gain applicable customers insight	RFM modelling, CLV	Dividing customers into 8 categories in order to target the right customers
(Dursun & Caber, 2016)	Profiling profitable hotels applying RFM model	RFM modelling	Finding RFM as an effective method to cluster customers
(Razavi Hajiagha, Hashemi, & Amoozad Mahdiraji, 2016)	Introducing an algorithm for DEA applications non-homogenous DMUs	Fuzzy c-means clustering	Recognising the positive effect of heterogeneity on some DMUs and negative effect of it on others
(Murray, Agard, & Barajas, 2017)	Using data mining tools to recognise behavior patterns in historical noisy delivery data	Dynamic Time Wrapping, Croston Method, data mining tools	Introducing a method for market segmentation found on delivery data, calculating similarities between customers
(Hendalianpour, Razmi, & Gheitasi, 2017)	Demonstrating a way to choose the best clustering method found on density-based clustering and offering a new algorithm for both crisp and fuzzy variables	SPSS, MATLAB	Developing a new algorithm named FCR
(Wang et al., 2018)	Reviewing the latest progress in data mining	—	—
(Bashiri Mosavi & Afsar, 2018)	Analysing customer value in the TEJARAT Bank branches in Iran	FAHP, K-means clustering, random forest classification	Proposing a model-attribute with the best characteristics in accuracy, precision, recall, F-measure and class- measure error
(Vahdatzad & Khademi Zare, 2018)	Modelling and clustering the customer’s credibly behavior to optimise the allocation of financial resource and improve bank service	AHP, FNN	Achieving the 98% enhancement in prediction speed and 12% increase in accuracy
(Aryuni & Miranda, 2018)	Building a clustering model on customer profile data found on their usage of internet bank in XYZ Bank	K-Means clustering, K-Medoids clustering, RFM, AWC, DBI	Better performance of K-Medoids than K-Means based on AWC and pretty better performance of K-Means found on DBI
(Brunelli & Rezaei, 2019)	Proposing a new metric into the framework of the BWM	—	Introducing an alternative formulation of BWM applying a modified objective function
(Jintana & Mori, 2019)	Applying data analysis to better serve customers and conduct customer segmentation more effectively	RFM, k-means clustering	Clustering the customers based on RFM
(Motlagh, Berry, & Neil, 2019)	Clustering of residential customers found on their load characteristics	Load clustering	Discussing important advantages of model-based time series clustering approaches

Figure 3. Methodology process.

Figure 4. Average balances per week.

Table 2. RFM results.

Recency	Frequency	Monterey
18	8	796

Table 3. Clusters emanated from K-mean.

Cluster number	Frequency of members	Relative frequency
Cluster 1	732,615	27.79%
Cluster 2	132	0.01%
Cluster 3	687,777	26.09%
Cluster 4	141,161	5.35%
Cluster 5	262,641	9.96%
Cluster 6	812,214	30.81%

Figure 5. Frequency of members.

Table 4. Name and rank of clusters.

Cluster number	Cluster rank	Frequency of members	Cluster name
Cluster 1	4	732,615	Silver- low interaction
Cluster 2	1	132	Special
Cluster 3	6	687,777	Averted- low interaction
Cluster 4	2	141,161	Loyal
Cluster 5	3	262,641	Silver- high interaction
Cluster 6	5	812,214	Bronze

Table 5. Description of clusters name.

Cluster rank	Cluster name	Description
1	Special	Mostly companies and legal entities High account balance High interaction
2	Loyal	High recency High frequency High monetary
3	Silver- high interaction	Nearly high recency Moderate high frequency Nearly moderate monetary
4	Silver- low interaction	Similar to 3 but less interaction
5	Bronze	Less recency Less frequency Less monetary
6	Averted- low interaction	Having no interaction for a while

Table 6. Decision matrix.

Cluster number	Monetary average	Frequency average	Monetary SD	Frequency SD	Frequency of members
Cluster 1	717,490,820/35	12	5,483,337,621/14	8/53	732,615
Cluster 2	2,301,022,466,851/30	32	4,011,183,233,619/01	37/87	132
Cluster 3	530,463,451/39	18	5,162,324,110/23	21/39	687,777
Cluster 4	1,357,507,405/66	142	8,813,678,707/63	48/15	141,161
Cluster 5	1,315,814,831/33	48	7,127,540,832/04	15/26	262,641
Cluster 6	2,921,915/91	112	562,598,014/06	106/65	812,214

Table 7. Weights derived by BWM.

Criterion	Monetary average	Frequency average	Monetary SD	Frequency SD	Frequency of members
Weight	0.42	0.31	0.13	0.1	0.04

Table 8. Cluster ranks by COPRAS.

Cluster number	Absolute priority	Rank
Cluster 1	27/33	3
Cluster 2	100/00	1
Cluster 3	15/02	6
Cluster 4	31/59	2
Cluster 5	22/82	5
Cluster 6	25/87	4

Table 9. Comparison between expert opinion and COPRAS rank.

Cluster number	Cluster name	Cluster rank by experts	Cluster rank by COPRAS
Cluster 1	Silver- low interaction	4	3
Cluster 2	Special	1	1
Cluster 3	Averted- low interaction	6	6
Cluster 4	Loyal	2	2
Cluster 5	Silver- high interaction	3	5
Cluster 6	Bronze	5	4

Figure 6. Comparing ranking results.

Table 10. Digital banking strategy development.

Target	Strategic area	Strategy	Loyal customers	Silver- low interaction customers
Customer orientation and service-centrism	Countering cyber threats	Employing biological parameters for authentication	✔
		Developing multilayer authentication tools	✔	✔
	Digital banking	Extending communication channels (e.g., mobile applications, internet, etc.)	✔	✔
		Establishing digital and self-service branches	✔
		Developing boarding customer service	✔
		Expanding e-wallet service (using cellphones as bank card)	✔	✔
		Developing connection and interaction with foreign banks	✔
		Developing credit card service	✔
		Integrated management of organisational customer’s accounts	✔
		Legal customer’s payments with ID	✔
	Advanced analysis methods	Fraud detection service	✔
		Customer behavior prediction	✔	✔
Infrastructure	Cloud services	Cost reduction	✔	✔
		Concentrating on customers and becoming service-centric	✔	✔
	Concentration on innovation and competition	Participation with non-bank customers (telecommunication companies and finteches)	✔	✔
		Creating innovation lab	✔	✔

Hiziroglu, A. (2013). Expert systems with applications soft computing applications in customer segmentation: State-of-art review and critique. Expert Systems with Applications, 40(16), 6491–6507. doi:10.1016/j.eswa.2013.05.052

 Web of Science ®Google Scholar

Julio, A., & Eugenia, V. (2014). Customer segmentation model based on value generation for marketing strategies formulation. Estudios Gerenciales, 30(130), 25–30. doi:10.1016/j.estger.2014.02.005

 Google Scholar

Kahreh, M. S., Tive, M., Babania, A., & Hesan, M. (2014). Analyzing the applications of customer lifetime value (CLV) based on benefit segmentation for the banking sector. Procedia - Social and Behavioral Sciences, 109, 590–594. doi:10.1016/j.sbspro.2013.12.511

 Google Scholar

Fahad, A., Alshatri, N., Tari, Z., Alamri, A., Khalil, I., Zomaya, A. Y., … Bouras, A. (2014). A Survey of clustering algorithms for big data: Taxonomy and empirical analysis. IEEE Transactions on Emerging Topics in Computing, 2(3), 267–279. doi:10.1109/TETC.2014.2330519

 Web of Science ®Google Scholar

Wang, B., Miao, Y., Zhao, H., Jin, J., & Chen, Y. (2016). A Biclustering-based method for market segmentation using customer pain points. Engineering Applications of Artificial Intelligence, 47, 101–109. doi:10.1016/j.engappai.2015.06.005

 Web of Science ®Google Scholar

Cebeci, Z., & Yildiz, F. (2015). Comparison of K-means and fuzzy C-means algorithms on different cluster structures. Journal of Agricultural Informatics, 6(3), 13–23. doi:10.17700/jai.2015.6.3.196

 Google Scholar

Sajjadi, K., Firuzabadi, M. A. K., Amiri, M., & Sadaghiani, J. S. (2015). A developing model for clustering and ranking bank customers. International Journal of Electronic Customer Relationship Management, 9(1), 73. doi:10.1504/IJECRM.2015.070701

 Google Scholar

Ansari, A., & Riasi, A. (2016). Customer clustering using a combination of fuzzy C-means and genetic algorithms. International Journal of Business and Management, 11(7), 59–66. doi:10.5539/ijbm.v11n7p59

 Google Scholar

Mohammadian, M., & Makhani, I. (2016). RFM-based customer segmentation as an elaborative analytical tool for enriching the creation of sales and trade marketing strategies. International Academic Journal of Accounting and Financial Management, 3(6), 21–35.

 Google Scholar

Dursun, A., & Caber, M. (2016). Using data mining techniques for profiling profitable hotel customers: An application of RFM analysis. Tourism Management Perspectives, 18(1), 153–160. doi:10.1016/j.tmp.2016.03.001

 Web of Science ®Google Scholar

Razavi Hajiagha, S. H., Hashemi, S. S., & Amoozad Mahdiraji, H. (2016). Fuzzy C-means based data envelopment analysis for mitigating the impact of units’ heterogeneity. Kybernetes, 45(3), 536–551. doi:10.1108/K-07-2015-0176

 Web of Science ®Google Scholar

Murray, P. W., Agard, B., & Barajas, M. A. (2017). Market segmentation through data mining: A method to extract behaviors from a noisy data set. Computers & Industrial Engineering, 109, 233–252. doi:10.1016/j.cie.2017.04.017

 Web of Science ®Google Scholar

Hendalianpour, A., Razmi, J., & Gheitasi, M. (2017). Comparing clustering models in bank customers: Based on fuzzy relational clustering approach. Accounting, 3(2), 81–94. doi:10.5267/j.ac.2016.8.003

 Google Scholar

Wang, R., Ji, W., Liu, M., Wang, X., Weng, J., Deng, S., … Yuan, C-A. (2018). Review on mining data from multiple data sources. Pattern Recognition Letters, 109, 120–128. doi:10.1016/j.patrec.2018.01.013

 Web of Science ®Google Scholar

Bashiri Mosavi, A., & Afsar, A. (2018). Customer value analysis in banks using data mining and fuzzy analytic hierarchy processes. International Journal of Information Technology & Decision Making, 17(3), 819–840. doi:10.1142/S0219622018500104

 Web of Science ®Google Scholar

Vahdatzad, M. A., & Khademi Zare, H. (2018). An evaluation method and clustering of credibly behavior of customers using AHP and fuzzy neural networks. Journal of Intelligent & Fuzzy Systems, 35(5), 5605–5617. doi:10.3233/JIFS-16323

 Web of Science ®Google Scholar

Aryuni, M., & Miranda, E. (2018). Customer segmentation in XYZ bank using K-means and K-medoids clustering. 2018 International Conference on Information Management and Technology (ICIMTech); September 1–9.

 Google Scholar

Brunelli, M., & Rezaei, J. (2019). A multiplicative best – Worst method for multi-criteria decision making. Operations Research Letters, 47(1), 12–15. doi:10.1016/j.orl.2018.11.008

 Web of Science ®Google Scholar

Jintana, J., & Mori, T. (2019). Customer clustering for a new method of marketing strategy support within the courier business. Academia Book Chapter, 31(2), 1–19.

 Google Scholar

Motlagh, O., Berry, A., & Neil, L. O. (2019). Clustering of residential electricity customers using load time series. Applied Energy, 237, 11–24. doi:10.1016/j.apenergy.2018.12.063

 Web of Science ®Google Scholar