Title: An empirical assessment of a univariate time series for demand planning in a demand-driven supply chain

Authors: John S. Jatta; Krishna Kumar Krishnan

Addresses: Department of Industrial and Manufacturing Engineering, Wichita State University, Wichita, KS 67260, USA ' Department of Industrial and Manufacturing Engineering, Wichita State University, Wichita, KS 67260, USA

Abstract: Many firms use customer orders time series as the basis of their forecasting and demand planning. However, there are other firms that use sales orders (shipments). Our research focused on evaluating and understanding the implications of using sales orders (shipments) to plan for a supply chain. We evaluated the structural difference between customer orders time series and sales orders time series. An experiment was conducted using a set of 48-month and a set of 576-month (long) normally distributed, randomly generated customer orders time series and shipment time series. The time series were statistically evaluated periodically by rolling the data and then comparing them using a two-sample comparison in Statgraphics Centurion XVII software. The series were then used to generate periodic forecast and their forecasts statistically tested using two-sample comparison. We found a statistically significant difference between the two series for both the 48-period time series and the extended 576-period time series. Our results show that customer orders time series is statistically different from shipment timer series due to censorship. Forecasts generated from customer orders and sales orders time series exhibit statistically significant difference. Using shipment time series to forecast and plan for a demand-driven supply chain causes a perpetual state of under-inventory.

Keywords: demand planning; forecasting; supply chain management; SCM; two sample testing; under-inventory; univariate time series; supply chain planning; sales orders; shipments; customer orders.

DOI: 10.1504/IJBFMI.2016.078607

International Journal of Business Forecasting and Marketing Intelligence, 2016 Vol.2 No.3, pp.269 - 290

Received: 31 Dec 2015
Accepted: 10 Jul 2016

Published online: 25 Aug 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article