Study On Credit Card Fraud And Fraud Resolution At Visa Inc.

Objectives of the Study

Visa Inc. is one of the most popular firms that allow transaction of small business. This gives them a platform to transact for businesses, customers and other financial institutions. It allows them to do transaction for enterprises, in doing so they face challenges such as credit card fraud.

The online fraud is an issue of concern that requires to be addressed. Visa Inc seeks knowledge on its security measures and customers concern to credit card fraud. In order to improve customers’ loyalty and trust the company needs to understand the source of online fraud and how to counter it.

1. To determine number of fraud experienced in different genders.
2. To determine if age affects credit card’s fraud
3. To determine average time to resolve a credit’s card fraud.
4. To determine number of occurrence of a card fraud.
5. To determine occurrence of online and offline fraud.

Data on customer experience on personal fraud was collected. The sampling technique used to sample from the population was simple random sampling. A total of 2000 customers were selected. In the selected sample which consisted of 2000 customers, 420 responded. The questionnaires were sent via email and mobile phone. The respondents’ ethic was upheld since the personal information was not detailed. The research made use of survey method to obtain descriptive and analytical data, (Perry & Perry, 2014).

In the study the following hypothesis were tested

Do the number of credit card fraud experienced across gender

H0:  Credit card fraud is affected by gender

H1: Credit card fraud is not affected by gender 

Customers fraud resolution team experience is satisfying

Hypothesis;

H0: Customers Fraud Resolution team is satisfying

H1: Customer fraud resolution team is not satisfying

Do credit card fraud differ across age groups

H0:  Card fraud differs across age groups.

H1:  Card fraud does not differ across age groups.

Is the average time used in resolving a credit card fraud equal to 12 hours

Hypothesis;

H0: The mean time used in resolving a card fraud is 12 hours

H1: The mean time used in resolving a card fraud is not 12 hours

How likely is an online or offline fraud to occur

Hypothesis;

H0: an online or offline card fraud occur in 12hours

H1: an online or offline card fraud doesn’t occur in 12 hours

Do response time, level of communication and level of advice affects customer satisfaction on fraud resolution

Hypothesis;

H0: Response time, level of communication and advice affects customer satisfaction

H1: Response time, level of communication and advice does not affect customer satisfaction 

The study used analysis of variance (ANOVA) and test of hypothesis in data analysis. The p-value in the analysis of variance and comparing it with the level of significance helped in rejecting the null hypothesis or failing to reject the null hypothesis. If the p-value calculated was less than the level of significance we reject null hypothesis and p-value calculated is greater than the level of significance we thus fail to reject the null hypothesis, (Neuman, 2004).

Research Design

The pie chart above shows customers who experienced fraud in debit and credit for past 12 months. Most of the customers which is 66% of the 420 respondents experienced credit, debit or EFTPOS were female and the rest were female. 

The frequency table below indicates the number of customer who experience of offline fraud. 142 of the respondents didn’t experience any form of offline card fraud. 112 of the 420 respondents strongly agree that the offline card fraud was existing. The respondent who declared that they had no knowledge and experience of offline card fraud was 141 out of the 420 respondents. These indicates that a total of 283 of the 420 respondents have no experience of offline fraud these number is above average.

offline card fraud	frequency
0	142
1	11
2	2
3	2
4	1
5	1
6	1
7	1
8	2
9	1
10	112
9999	141

The table below indicates the number of online card frauds was experienced. In these case most of the respondents experienced online card fraud but different levels. Only 17 of the respondents who didn’t experience online fraud, those who experienced 1 to 12 online card fraud ranged from 10 to 30. 141 respondents didn’t have the understanding of if they experienced online card fraud. This indicates that most of the card frauds occur when the card is online and only on few occasions when the card is offline.   

online fraud	Freq
0	17
1	30
2	19
3	15
4	17
5	20
6	19
7	13
8	18
9	26
10	17
11	29
12	26
9999	141

A pie chart of gender of the respondents  

57% of the respondents were females while 43% of the respondents were males. Out of the 420 respondents 237 of them were females while 182 of them were males. These indicate the majority of the respondents were female.

Inferential statistics is used to make conclusion about a population using a sample, (Desaro, 2011).

H0:  Credit card fraud is affected by gender

H1: Credit card fraud is not affected by gender

SUMMARY
Groups	Count	Sum	Average	Variance
Males	183	240	1.311475	0.215637
Females	237	322	1.35865	0.230995
ANOVA
Source of Variation	SS	df	MS	F	P-value	F crit
Between Groups	0.229807	1	0.229807	1.024515	0.312036	3.863801
Within Groups	93.76067	418	0.224308
Total	93.99048	419

The number of males in the sample selected was183 with average of 1.3 which means that they at one point or another experienced credit, debit fraud.  The number of females in the sample was 237 and an average 1.3 which is next to 2 thus experienced some kind of fraud. The p- value is 0.31 which is greater than 0.05 and thus the null hypothesis is not rejected and conclude that number of the card fraud was experienced across gender. Both gender experienced fraud.

The total number of respondents was 420. Most of them were 237 were female and 183 were females. The average means of both the males and females’ is1.3 which is approximately 1 thus both genders experienced card fraud. 

Hypothesis

H0:  Card fraud differs across age groups.

SUMMARY
Groups	Count	Sum	Average	Variance
25 and below	100	133	1.33	0.223333
26-35 years	108	150	1.388889	0.239875
36-45 years	112	144	1.285714	0.20592
46-55 years	64	86	1.34375	0.229167
56 & above	36	49	1.361111	0.237302
ANOVA
Source of Variation	SS	df	MS	F	P-value	F crit
Between Groups	0.613611	4	0.153403	0.681777	0.604895	2.393438
Within Groups	93.37687	415	0.225004
Total	93.99048	419

The analysis of variance of the age groups and experience of fraud, the p-value 0.604895 which is greater than level of significance 0.05 we fail to reject the null hypothesis and conclude that there isn’t a significant difference across the age groups regarding card fraud. Card fraud is experienced across the age groups and no age group has a higher risk compared to the other. The card fraud doesn’t target the aged as one may make a prediction but done across the age groups.

The age group have gender experienced fraud and thus on average attaining a mean of 1.3 which is less than 2 thus most of the clients experience fraud. 

H0: The mean time used in resolving a card fraud is 12 hours

Most of the customers experience online fraud and as a result they tend seek solution. Customers tend to review the average time lost in resolving these type of fraud.

SUMMARY
Groups	Count	Sum	Average	Variance
Column 1	278	278	1	0
Column 2	278	3795	13.65108	308.4302
ANOVA
Source of Variation	SS	df	MS	F	P-value	F crit
Between Groups	22246.92	1	22246.92	144.2591	1.07E-29	3.858298
Within Groups	85435.15	554	154.2151
Total	107682.1	555

The p-value is less than significant level which is 0.05 thus the null hypothesis is not rejected that the mean time used for a card fraud to be resolved is 12 hours. Customers who experience card fraud tend to find time and seek solution for their card fraud. The average time spent to resolve any kind of a fraud doesn’t prevent persons from seeking solutions.

The average number of hours taken by the card fraud team is 13 hour these is greater than

The value in the hypotheses which is 12 hours thus we reject the null hypothesis.

H0: an online or offline card fraud occur in 12hours

H1: an online or offline card fraud doesn’t occur in 12 hours

SUMMARY
Groups	Count	Sum	Average	Variance
Column 1	278	278	1	0
Column 2	278	1743	6.269784	15.40349
ANOVA
Source of Variation	SS	df	MS	F	P-value	F crit
Between Groups	3860.117	1	3860.117	501.2004	1.51E-79	3.858298
Within Groups	4266.766	554	7.701744
Total	8126.883	555

ANOVA above online card fraud

The hypothesis was testing the occurrence of online card fraud within intervals of 12 hours, the significant level 0.05 is greater than probability value and therefore the null hypothesis is not rejected and conclude that an online fraud can occur less than or more than 12 hours. How often online card fraud can be controlled through administering new security terms to avoid them from occurring.   

The null hypothesis tests that online frauds occur in the intervals of 12 hour but the average time in the summary is 6 hour and conclude that the card fraud happens is statistically significance different from 12 hours.

Do response time, level of communication and level of advice affects customer satisfaction on fraud resolution 

Statistical Analysis

SUMMARY
Groups	Count	Sum	Average	Variance
Column 1	420	1648	3.92381	9.865303
Column 2	420	1786	4.252381	12.27983
ANOVA
Source of Variation	SS	Df	MS	F	P-value	F crit
Between Groups	22.67143	1	22.67143	2.047532	0.152826	3.852579
Within Groups	9278.81	838	11.07257
Total	9301.481	839

The p-value is 0.15 greater than the level of significance (0.05) thus null hypothesis that the mean satisfaction score from the groups of level of advice, level of communication and response time are same and conclude that they are statistically different. Most of the customers found that the time taken to respond on card fraud did influence their rating. Customer choose to spend time on the card resolution team as much as could provided they solved their problem and gave advice on overcome future cases of card fraud.

The mean value of rating the response time is 4 which are below average value which is 5 thus the response time didn’t affect.

SUMMARY
Groups	Count	Sum	Average	Variance
Column 1	420	1648	3.92381	9.865303
Column 2	420	1995	4.75	15.05668
ANOVA
Source of Variation	SS	df	MS	F	P-value	F crit
Between Groups	143.344	1	143.344	11.50342	0.000727	3.852579
Within Groups	10442.31	838	12.46099
Total	10585.66	839

 The probability value is 0.00072 that is less than significance level (0.05) thus the null hypothesis is not rejected that Do response time, level of communication and level of advice affects customer satisfaction on fraud resolution and conclude that satisfaction rating was greatly influence by the level of advice given to a customer. The rating was highly directed by the advice of the team since that was the only channel the customer would prevent future cases of card fraud.

The average value of the level of advice which is 4.7 and significantly 5 thus we conclude that the rating is above average and for these case null hypothesis is rejected that and conclusion is that the level of advice influenced customer satisfaction rating.

SUMMARY
Groups	Count	Sum	Average	Variance
Column 1	420	1648	3.92381	9.865303
Column 2	420	1738	4.138095	12.31501
ANOVA
Source of Variation	SS	df	MS	F
Between Groups	9.642857	1	9.642857	0.869497
Within Groups	9293.552	838	11.09016
Total	9303.195	839

The probability value 0.35 which is greater than significance level of 0.05 and the null hypothesis is rejected. Thus the level of communication is significance in customer’s rating of satisfaction score of assistance given in card fraud resolving.  The customer satisfaction score when asked about communication they didn’t experience enough communication through the credit fraud team.

The average value for level of communication is 3.9 hence below 5 thus level of communication on the credit card fraud didn’t influence the customer satisfaction rating.

 The level of significance used in the study is 95% giving alpha as 0.05. The Visa Inc show that they hypothesis tested can be used in inference the future of the company.

In the test of the hypothesis if the credit card fraud is across the gender then we conclude that it happens across the gender. Both males and females are exposed to fraud and for this case they happen to be careful with their credit cards.  The study has also identified that most of the respondents were female as compared to males. Females were more willing to respond to the questionnaires than the men. The response of the questionnaire was directly proportional to the percentage of gender, males and females.

Results, and Statistical and Non-statistical Interpretation

The credit card fraud across the age group it is both the young and the aged have experienced card fraud and for that case doesn’t dependent on the age. One from logics may tend to predict that since people in the working class have money in their cards then they are higher targets compared to the students and children and thus less exposed to the credit card fraud. Those conducting the fraud have need for even the small monies. That refers even to the savings made by parents for their children and youths.

The time taken by the resolution team is less than the 12 hour that has been listed by the management. Most of the customers who seek resolution team tend to get these services faster than expected. This is due to the fact that similar challenges occur to customers and thus becomes easy for the resolution team to solve this case.

Time taken for a credit card fraud to occur is less than 12 hours. Most of the customers didn’t have an understanding that these kind of fraud happens very shortly. Most of the frauds were identified to happen during the online of the card. The card needed to be online so that fraud to be transacted. When the card was offline the fraud didn’t happen thus the fraud team ensured they identified when the card was online.

The customer satisfaction rating of the support team was influenced by different issues. The response time taken did not by any chance influence the rating of the customer satisfaction team. Mostly the customers didn’t consider that to be important. Most of the customer had created time to seek a solution and due to the fact that most problems were solved in less than the depicted time thus failed to rate using time.

Customer satisfaction was done using the level of advice. The advice that the customers’ were given seems was the most important they required. This was due to the fact that it would prevent future cases of credit card fraud. Customers’ satisfaction rating was with how much appropriate this advice was to the problem. The communication by the customer service didn’t influence customers in any way. Most of the customers were satisfied by the services they were offered by the response time.

The customer satisfaction should improve their means of communication that is via the phone and also do be made on the desk. The time within which these services are offered should be reduced. These would improve the rating of customer satisfaction. 

• The credit card fraud team should consider having a fraud proof. These helps when the customer inputs billing number and then one should be given proof that he/she is the owner of the credit card that is in use. Most of the credit cards one finds that they are directly to transaction even without any further confirmation.
• The credit card fraud resolution team should input a log of credit card numbers. Each of the customers should input a log that is specific to them and no other person. These log should be limited by number of times such that when these times are excided the card get blocked and can’t function thus to seek resolution team.
• The customer satisfaction should include that IP address and credit card address match up to ensure that they don’t mix up in the process of using the credit card. Most cards are used in locations that are not specific to them. Thus if the credit card is used in the wrong location, then it’s identified as fraud.
• The customers should be advised on the credit card fraud. These should be done by watching for suspicious email accounts. Incase customer identifies such email he/she should report the case to the customer resolution team to allow them do research on them and tackle them.   

References

Alderman A. & Salem B. (2010). Survey design. Plastic Reconstruction Surgery, vol 4, pg 9.

Dean, J., (2014). Big Data, Data Mining, and Machine Learning: Value Creation for Business Leaders and Practitioners. John Wiley & Sons.

Desaro S. (2011). A Students guide to conceptual side of inferential statistics.

Kothari, C. (2004). Research Methodology Methods and Techniques New Age International. (P) Limited, Publishers :New Delhi 

Lazear, E. (2005) Entrepreneurship: Journal of Labor Economics , vol. 23, pg. 649-680. 

Neuman, W. L. (2014). Social Research Methods: Qualitative and Quantitative Approaches, 7th Edition. Pearson Education Limited: UK.

Perry, J. & Perry, E. (2014). Contemporary Society: An Introduction to Social Science, 12th Edition. Pearson Education, Inc.: Singapore

Tashakkori & C. Teddlie (2003). Handbook of mixed methods in social & research. Thousands Oaks, CA: Sage.

Visa Inc. (2017). Visa Website [online]https://www.visa.com

Whitley E, Ball J. (2002). Statistics review 1: Presenting and summarizing data. Crit Care. 

Zikmund, W., Babin, B., Carr, J., & Griffin, M. (2013). Business research methods (9th ed.): Cengage Learning