Analysis of Customer Characteristics in Online Food Delivery System

This report examines the significance of customer characteristics within the context of online food delivery, specifically using data from Receipo Spain. The analysis identifies key factors contributing to order cancellations, including address errors, courier issues, duplicate jobs, and end-customer satisfaction. The methodology involves using a large dataset with approximately fifteen thousand records, employing R code generation and histogram generation techniques to find the various characteristics of customers. The study categorizes various issues, such as address errors due to incorrect entries, courier issues arising from high order volumes, and courier requests from difficult locations. Duplicate jobs, end-customer satisfaction, job expiration, and late courier issues are also explored. Histograms are used to represent the significance of customer characteristics. The report also differentiates between good and bad customers based on their behavior and order patterns. The analysis offers a detailed understanding of customer behavior in the online food delivery ecosystem, enabling businesses to improve service and customer satisfaction.

Significance characteristics of customers
Introduction
The creation of online food ordering giving various good things to the customers who
don't have time to order the food directly and who don't need to travel and order the food by
going to the restaurant directly. The online food delivery system also offers the things like offers,
discounts and free delivery, etc. to fulfill the customer’s satisfaction. They also offers the other
things like order line without any cash (cash on delivery), customers receives the delivery within
the few minutes of time. If this these things are not properly maintained it leads to cancelling of
an order by the customer (Kieling et al., 2011). In this topic we are going to see about the
customer’s response in the online food delivery management system and the significance
characteristics of customers, who orders the food through online from the Receipo Spain
restaurant.
Literature review
‘Ayazlar and Artuğer’, Carried a research on the characteristics of customers. It is
identified using different ways according to the number of occurrences by the customers. The
good and bad customer’s specification is also identified based on the characteristics of
customers. According to the author, the analysis of the large datasets involves many
complications. The present method was not adequate for the efficient analysis of the datasets
(Ayazlar and Artuğer, 2015). It facilitates the better efficient analysis reports. Also, it provides
the better support for the decision making process.
Methodology
For finding the significance characteristics of customers there are some methods,
researches and techniques are used. Using the given receipo Spain data various things are
identified in order to find the characteristics of customers. The given the data set for the Receipo
Spain is too large and contains nearly fifteen thousand records. So, we need to use different
technologies like R code generation, histogram generation for finding the various characteristics.
The other methodologies are also used for finding the rich and poor customers and also for
finding good and bad customers.

Significance characteristics of customers
The significance of the customer’s characteristics also represented in terms of histograms.
They are:
 address error,
 courier issue,
 courier request,
 duplicate job,
 end customer satisfaction,
 job expired,
 late courier,
 long time wait at do,
 long time wait at PU,
 no courier available,
 package damaged,
 package incorrect,
 package not ready,
 PU closed,
 wrong transport type,
 And other reasons.
Address error:
The address error may happen by means of human errors like the customer misspelled
their addresses or the retailer who takes the order can type the addresses wrongly or the voice of
the customer is not clear through the communication media at the time there is a possibility of
writing the address may incorrect at that point everything will be wrong and causes this kind of
errors (Inman, Winer and Ferraro, 2009). There are thirty orders are canceled due to address
error. It is an error or fault happens during the orders placed by the customer through the
communications media like telephone or internet. It happens because of the minute error
happened during the placement of the order by the retailer.so these kinds of error lead to address
error. There is a possibility of canceling the orders happens due to these kinds of errors.

Figure 1 address error
The above histogram represents the address error occurs during the delivery of orders for the
specific order.
Courier issue:

The courier issue is one of the reasons for cancellation of orders. The possibility of these
errors are very high and hence cancellation behind these types of error also high (Motulsky,
2014). It is happened because of the huge amount of orders. For example, there are 14365 orders
are initiated for the Receipo Spain client in the given data set. At that time most of the orders are
needs to temporarily stopped, for other orders to complete. At that moment the courier issue and
its related issues have occurred. This issue also occurs during the careless mistake done by the
delivery agent like it is not delivered properly to client, most of the time the delivery agent forget
to deliver the product in the proper zip code due to a large number of orders placed by the
particular delivery agent and these are the reasons which lead to courier issue.
Figure 2 courier issue
The above histogram represents the courier issue in the given data.
Courier request:

Courier request is one of the reasons for cancellation of orders. For example, the
address request given by the customer is too long or it is a remote area. Hence the process is
declined by the client by giving different reasons. At that time customers mentality is giving bad
reviews about the client (Pejović et al., 2014). And also this kind of issues happens because of
the request given the customer has already canceled many orders previously at these cases client
giving different reasons and canceling the particular order. Because almost all the time that
particular client cancels the order and those things are already noted in the database and those
kinds of customers are treated as bad customers from the retailer’s point of view. At these cases,
the courier request will not be accepted by the client. So these kinds of orders also treated as
courier request. The impact of the bad customers is not that much even though the total number
of time wasted for the particular category is keep on increasing. And also some of the customers
are giving some silly reasons and hence they are also added to the category of bad customers.
The action taken against them is not high but in terms, small actions like their third orders are not
taken. Because the analyzers already checked two times and hence they are obeying and
cooperated well with the shop.
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
8012
8013
8014
8015
8017
8018
8020
8022
8024
8025
8026
8027
8028
8029
8036
8037
0
500
1000
1500
2000
2500
3000
3500
Zipcodes vs Cancelled/Finished rate
cancelled finished
Zipcodes
no.of orders
The above histogram explains about relationship between zip code vs. cancelled and
finished rate for people of different area and their orders. Where the total number of cancelled
and finished orders happened by courier request.

Figure 3 courier request
The above diagram represents the courier request issues in the customer characteristics.
Duplicate job:
The duplicate job is another reason for the cancellation of orders. It may happen when the
delivery agent uses the third person for delivering the product or agent gives his work to some
other person due to his unavailability. At that time the particular delivery person’s job is not a
valid one, he may be working as the part-time worker and his experience in this field also very
less so he can’t fulfill the customer needs. Hence these kinds of orders are canceled. And also the
orders canceled for these kinds of reasons are very high (Zhu, 2017). Customers can easily be
identified as the difference between these people by analyzing their behavior and the way they
approach. This may cause cancellation of an order.

Figure 4 duplicate job
The above histogram represents the duplicate and various transport types used for
duplicate jobs. Totally there are forty seven duplicate job entries found in the data. Among them
twenty six orders are used through bikes. Hence majority of duplicate jobs are done through
bikes.
End customer satisfaction:
End customer satisfaction is another reason because of their lack of experience and lack
of maturity. The delivery agents are the one who acts as an interface between the customer and
retailer. So they have huge responsibilities while delivering an order. Even though the
customer’s behavior is not good they have to accept the behavior of the customer for achieving

the end customer satisfaction (G. Black, H. Vincent and J. Skinner, 2014). It is also called as the
end customer cancellation. The representation of rich customer and the poor customer can be
calculated with the help of zip codes in the data. The zip code da in the data are analyzed using
the software and then it is processed and needs to enter the postal zip codes then it will give the
histogram or graph in a pictorial representation that gives a clear idea about the customers status
that is who are all rich consumers and who are all poor consumers exactly using the zip codes. It
is calculated by means of the majority of the consumers in the given area. The number of users
living in the first range is very high when compared to the second range. So the customers who
are all repeatedly bought and given the good review about the restaurant from both the ranges are
called as good customers and customers who did not give the proper address and not responds
when the time of delivery and who simply orders and those orders are separated and the persons
who ordered such a product are treated as bad customers (Correia et al., 2018).
8001 8003 8009 8010 8015 8018
0
0.5
1
1.5
2
2.5
zipcode vs end customer cancellation
zipcodes
no. of cancellation
The above histogram explains the cancellation orders of end customer. The number of
cancellations occurred in the zip codes. There are two zip codes which has more end customer
cancellations.

Figure 5 end customer satisfaction
The above histogram represents the end customer cancellation in the form of graphical
representation.

Job expired:
Most of the delivery agent’s jobs expired already even though they are working
and they are used by the retailer for delivery because their experience in this field is high. These
kinds of reasons have happened in very fewer cases, for example, the delivery agent failed to
deliver the orders because their job is expired. It also happens when the task was given for the
particular agent valid for only order but due to huge of orders the orders may increase at that
time the agent failed to deliver the order for various reasons like the long distance, customer not
available etc. It is also one of the reasons for cancellation. It happens in worst case scenarios
(Kouda, 2017). It mainly happens when the maximum numbers hours spent for only one order
due to long waiting time in delivery of the order, at this cases other orders get the delay or can’t
able to deliver at the given time.
Figure 6 job expired

The above histogram represents the graphical representation of job expired for the delivery
agents.
Late courier:
This is one of the reasons for canceling and it happens when there is a delay in the
delivery of the products. Hence customers waiting time also increases which leads to
cancellation. The major reasons for late courier are the agent who delivers the orders may handle
one or more orders. So the time taken to the previous order may also high. Or the chance of late
courier happens because of the reasons like huge orders, wrong transport type used, issues in the
transport type, long waiting time taken for delivery to previous customers etc. for example the
delivery agent takes one or more orders may cause late in the courier (Correia et al., 2018).
Figure 7 late courier
The above histogram represents the streets those are the reason for late courier.

8009 8007 8029 8015 8013 8014 8010 8007
0
50
100
150
200
250
300
350
400
121
78
127
344
160
92
311
78
1
cancelled orders for late courier
zipcodes
orders
Long time wait at delivery of order:
There is some certain average of waiting time are maintained during the delivery of the
product. If that average waiting time exceeds then the delivery agents also have the rights to
abort the order (Sun, 2014). The customers are not available at the time of delivery, change in the
delivery address once after the order confirmation, not paying or cash not Hence the cancellation
occurs for the mentioned reasons. According to the retailer, the delivery of food products are
completely considered as the business and also the food products need to deliver within the time
unless it will get wasted. So the distance was the main parameter in this case.