CLO Business Decision Making Project, Part 2
QNT/275

April 11th 2016

Knowing that is statistics data is everything. So the collection of data is very important, it comes in handy when making a conclusion, doing calculations such as range, median and modes. So the types of descriptive data that could be used to describe and summarize this particular data set. When considering the problem previously address both types of descriptive data can be used. For instance, with the My voice of the customer survey based upon how often a customer has to call back if that is ranked in the 30 percentile meaning that at least 30 percent of all the customers that called on one given day called back within 48 hours we are able to calculate the range of callbacks therefore we are able to calculate the average amount of callbacks on a given day. However, the most effective method that would be used of course would be finding the Median, Mean and Mode so that we are able to find the average reason for customers calling back within forty-eight hours and also the average My voice of the customer score coming from these particular customers. By these customer being routed to a special team of operations to assist this can also help to determine the feedback given by the representatives to help better assist the customers.
Know because inferential statistics come from the sample to the population and determine the possible characteristics of said population based upon the sample. Inferential Statistics based upon the problem previously described of the Augusta Site reaching the number one spot in the nation will allow me to make my findings more generalized. There are a few inferential statistics that can play a big role in determining if the data set information will be beneficial to the company. One of them being the contingency tables because if we can determine if scores are based on gender such as a woman calling back will generate a lower score than if a male calling back, can definitely assist with know what group to target. Another will be Bi-variate Regression because is the amount of time a customer service agent spends in training and the information gained during said training for instance during training if one agent received 50 calls in one session and another 25 calls within that same session will the outcome of their scores be different? Are the hours spent on the phone positively related to their voice of the customer survey’s?
With this particular problem using probability analysis may be somewhat effective because we are talking about using data that is random and in one set of data the group random surveys that the customer is receiving and by knowing previous years losses and what the scores where during that time we can predict what events are happening that could be driving the company’s scores higher or lower and may be able to ramp up during those times to help bring in more higher scores and reduce the amount of low scores. Therefor we should be able to figure out the probability of the outcome the best way that we can and then invest the time to move forward and attack the scores head on.
Trend analysis will be our biggest ally in the statistical game of becoming the number one site in the nation and bringing home that wonderful trophy to Augusta, GA. Because by being able to collect the data and really spotting that pattern or trended behavior not only in our customers but in our agents by listening to a large portion of calls and finding the trended behavior that the agent is displaying and working towards fixing that behavior will almost guarantee a quick and permanent fix. Think about how many agents are driving the lower scores if we could hand pick those individuals who score low month to month and hone in on what is driving the low scores by listening to all the calls an agent receives a 4 or below and then identifying the behavior that tends to be a trend from call to call, have that agent work with someone to fix said behavior we can develop a power house of individuals that are satisfying their customers getting higher my voice of the customers survey scores and less customers calling back we can be sure to take over the number on spot without a doubt. The purpose of the trend analysis is to predict future events and in this situation we are using past events or past scoring to help improve future scoring of the representatives, using the trend analysis will be a much greater benefit to the company because with the thousands of calls taken per month per agent this particular analysis can help us get to the root cause a lot quicker and can bring about some quicker results in a shorter time frame.

The role that linear regression for trend setting might play in this particular problem would be helping with forecasting and reducing errors. Because as mentioned the biggest issue is the voice of the customer survey and if something goes wrong when the customer calls in it is more likely that the customer will provide the representative with a much lower score than if everything was done correctly and information provided was accurate. So by reducing error we would be able to surely increase the customers my voice survey scores, because we are able to observe the data set of y and X values, which will allow us to because more efficient and effective for our customers.
The role of time series in this particular problem can benefit from year to year because it’s used over a continuous time interval and since time series is used in things like the Dow Jones Industrial average I am sure it can benefit the company because in the end of things it is all based on the average at the end of the day what is the average percent to goal to beat the number on team in the nation. What will it take to reach 125% to goal monthly? How often are our customers completing survey’s each quarter? How often are we recognizing and identifying trended behavior’s each quarter? How often are customer’s calling back each quarter? All of these would be time series questions because we are seeking out the average based on a specific period of time.