Moving Forward with the On Demand Real-time Enterprise

Chapter 7. The case study 283

7.1.2 The data flow

Figure 7-1 depicts the initial data flow through a subset of some of the sub

systems that exists in ValueTrend - our fictitious retail company. The numbering

in the Figure 7-1 relates to the case study described below.

Figure 7-1 The ValueTrend original data flow

Case study

The following is a brief description of the events and activities involved in the

case study:

1. A purchase order for a ValueTrend hair-dryer is placed by the central

purchasing department.

2. The order for the hair-dryer is fulfilled by the supplier and delivered to the

ValueTrend central distribution center.

3. The hair-dryer is then distributed to the ValueTrend retail store.

4. The Inventory system is updated to reflect this event.

5. The hair-dryer is now available to satisfy incoming sales orders from the

ValueTrend sales channels.

6. Time passes and orders are processed. This should typically result in

products being delivered and the result a happy customer.

7. ValueTrend receives calls from a number of unhappy customers complaining

about the hair-dryer. They all indicate a belief that it is defective.

Purchase

Order

Inventory

CRM

Supply

Management

Distribution

Point of

Sales

Marketing

External

Suppliers

6

54

3

1

2

Client

Client

7

8

9

Business

User

12

EDW

ODS

284 Moving Forward with the On Demand Real-time Enterprise

8. All returns, and a code indicating the reason for the return, are registered in

the CRM system. Records of the returns are created in the CRM system, and

are then used as measures for determining customer satisfaction. The reason

codes have been implemented to handled all direct customer dialogs. All

product return records are given a status code reflecting the seriousness of

the defect.

9. The return reason information is collected and transferred to the ODS by a

weekly batch ETL process.

10.The returns information is, at a scheduled point-in-time extracted,

transformed, and loaded into the ValueTrend EDW.

11.The central returns department receives the actual hair-dryers back from the

stores, and, with their BI capability, discover that a high percentage of the

sold hair-dryers are being returned.

12.An investigation is launched by quality control. Sales statistics are produced

by taking information from EDW and the ODS.

13.It is decided by ValueTrend to recall the hair-dryer. Through a press release,

and stores notice boards, customers are informed that they can return the

ValueTrend hair-dryer and get a full refund.

During the time span from when customers first began returning the hair-dryers,

to the point-in-time when the decision is made to remove them from the store

shelves and the eBusiness catalog, many days has passed. And in that time, a

large number of hair-dryers had been sold.

Because of the long period of time taken by ValueTrend to respond to the

problem, the company was exposed to a high risk of loosing credibility and

thereby also potential loss of revenue on similar products.

It was determined that an improved process to handle these types of issues

needed to be designed and deployed - quickly.

Improving the data flow

The objective of the improved process is to achieve the data flow through the

processes with much lower latency. This is a much discussed issue in many

companies today. How can access to real-time data about the processes be

made available and result in improving the processes and minimizing their cost?

In this case study, we show a simple automated closed-loop process. It

illustrates how a number of technologies, and approaches, can be used, along

with process changes that enable better integration with those technologies, to

enable a real-time enterprise environment. With the real-time data, companies

can be more proactive. That is, trends and issues can be discovered and

resolved more quickly, thus avoiding, or at least minimizing, their impact. This