Forecasting new product life cycle curves: Practical approach and empirical analysis

Kejia Hu, Jason Andrew Acimovic, Francisco Erize, Douglas J. Thomas, Jan A. Van Mieghem

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We present an approach to forecast customer orders of ready-to-launch new products that are similar to past products. The approach fits product life cycle (PLC) curves to historical customer order data, clusters the curves of similar products, and uses the representative curve of the new product’s cluster to generate its forecast. We propose three families of curves to fit the PLC: bass diffusion curves, polynomial curves, and simple piecewise-linear curves (triangles and trapezoids). Using a large data set of customer orders for 4,037,826 units of 170 Dell computer products sold over three and a half years, we compare goodness of fit and complexity for these families of curves. Fourth-order polynomial curves provide the best in-sample fit with piecewise-linear curves a close second. Using a trapezoidal fit, we find that the PLCs in our data have very short maturity stages; more than 20% have no maturity stage and are best fit by a triangle. The fitted PLC curves of similar products are clustered either by known product characteristics or by data-driven clustering. Our key empirical finding is that, for our large data set, data-driven clustering of simple triangles and trapezoids, which are simple to estimate and explain, perform best for forecasting. Our conservative out-of-sample forecast evaluation, using data-driven clustering of triangles and trapezoids, results in mean absolute errors approximately 2%–3% below Dell’s forecasts. We also apply our method to a second data set of a smaller company and find consistent results.

Original languageEnglish (US)
Pages (from-to)66-85
Number of pages20
JournalManufacturing and Service Operations Management
Volume21
Issue number1
DOIs
StatePublished - Jan 1 2019

Fingerprint

Product lifecycle
Empirical analysis
Clustering
Polynomials
Maturity
New products
Forecast evaluation
Product characteristics
Out-of-sample forecasting
Small companies
Goodness of fit

All Science Journal Classification (ASJC) codes

  • Strategy and Management
  • Management Science and Operations Research

Cite this

Hu, Kejia ; Acimovic, Jason Andrew ; Erize, Francisco ; Thomas, Douglas J. ; Van Mieghem, Jan A. / Forecasting new product life cycle curves : Practical approach and empirical analysis. In: Manufacturing and Service Operations Management. 2019 ; Vol. 21, No. 1. pp. 66-85.
@article{74724f4e9db549548ff19f8d710ffd19,
title = "Forecasting new product life cycle curves: Practical approach and empirical analysis",
abstract = "We present an approach to forecast customer orders of ready-to-launch new products that are similar to past products. The approach fits product life cycle (PLC) curves to historical customer order data, clusters the curves of similar products, and uses the representative curve of the new product’s cluster to generate its forecast. We propose three families of curves to fit the PLC: bass diffusion curves, polynomial curves, and simple piecewise-linear curves (triangles and trapezoids). Using a large data set of customer orders for 4,037,826 units of 170 Dell computer products sold over three and a half years, we compare goodness of fit and complexity for these families of curves. Fourth-order polynomial curves provide the best in-sample fit with piecewise-linear curves a close second. Using a trapezoidal fit, we find that the PLCs in our data have very short maturity stages; more than 20{\%} have no maturity stage and are best fit by a triangle. The fitted PLC curves of similar products are clustered either by known product characteristics or by data-driven clustering. Our key empirical finding is that, for our large data set, data-driven clustering of simple triangles and trapezoids, which are simple to estimate and explain, perform best for forecasting. Our conservative out-of-sample forecast evaluation, using data-driven clustering of triangles and trapezoids, results in mean absolute errors approximately 2{\%}–3{\%} below Dell’s forecasts. We also apply our method to a second data set of a smaller company and find consistent results.",
author = "Kejia Hu and Acimovic, {Jason Andrew} and Francisco Erize and Thomas, {Douglas J.} and {Van Mieghem}, {Jan A.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1287/msom.2017.0691",
language = "English (US)",
volume = "21",
pages = "66--85",
journal = "Manufacturing and Service Operations Management",
issn = "1523-4614",
publisher = "INFORMS Inst.for Operations Res.and the Management Sciences",
number = "1",

}

Forecasting new product life cycle curves : Practical approach and empirical analysis. / Hu, Kejia; Acimovic, Jason Andrew; Erize, Francisco; Thomas, Douglas J.; Van Mieghem, Jan A.

In: Manufacturing and Service Operations Management, Vol. 21, No. 1, 01.01.2019, p. 66-85.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Forecasting new product life cycle curves

T2 - Practical approach and empirical analysis

AU - Hu, Kejia

AU - Acimovic, Jason Andrew

AU - Erize, Francisco

AU - Thomas, Douglas J.

AU - Van Mieghem, Jan A.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - We present an approach to forecast customer orders of ready-to-launch new products that are similar to past products. The approach fits product life cycle (PLC) curves to historical customer order data, clusters the curves of similar products, and uses the representative curve of the new product’s cluster to generate its forecast. We propose three families of curves to fit the PLC: bass diffusion curves, polynomial curves, and simple piecewise-linear curves (triangles and trapezoids). Using a large data set of customer orders for 4,037,826 units of 170 Dell computer products sold over three and a half years, we compare goodness of fit and complexity for these families of curves. Fourth-order polynomial curves provide the best in-sample fit with piecewise-linear curves a close second. Using a trapezoidal fit, we find that the PLCs in our data have very short maturity stages; more than 20% have no maturity stage and are best fit by a triangle. The fitted PLC curves of similar products are clustered either by known product characteristics or by data-driven clustering. Our key empirical finding is that, for our large data set, data-driven clustering of simple triangles and trapezoids, which are simple to estimate and explain, perform best for forecasting. Our conservative out-of-sample forecast evaluation, using data-driven clustering of triangles and trapezoids, results in mean absolute errors approximately 2%–3% below Dell’s forecasts. We also apply our method to a second data set of a smaller company and find consistent results.

AB - We present an approach to forecast customer orders of ready-to-launch new products that are similar to past products. The approach fits product life cycle (PLC) curves to historical customer order data, clusters the curves of similar products, and uses the representative curve of the new product’s cluster to generate its forecast. We propose three families of curves to fit the PLC: bass diffusion curves, polynomial curves, and simple piecewise-linear curves (triangles and trapezoids). Using a large data set of customer orders for 4,037,826 units of 170 Dell computer products sold over three and a half years, we compare goodness of fit and complexity for these families of curves. Fourth-order polynomial curves provide the best in-sample fit with piecewise-linear curves a close second. Using a trapezoidal fit, we find that the PLCs in our data have very short maturity stages; more than 20% have no maturity stage and are best fit by a triangle. The fitted PLC curves of similar products are clustered either by known product characteristics or by data-driven clustering. Our key empirical finding is that, for our large data set, data-driven clustering of simple triangles and trapezoids, which are simple to estimate and explain, perform best for forecasting. Our conservative out-of-sample forecast evaluation, using data-driven clustering of triangles and trapezoids, results in mean absolute errors approximately 2%–3% below Dell’s forecasts. We also apply our method to a second data set of a smaller company and find consistent results.

UR - http://www.scopus.com/inward/record.url?scp=85068892558&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068892558&partnerID=8YFLogxK

U2 - 10.1287/msom.2017.0691

DO - 10.1287/msom.2017.0691

M3 - Article

AN - SCOPUS:85068892558

VL - 21

SP - 66

EP - 85

JO - Manufacturing and Service Operations Management

JF - Manufacturing and Service Operations Management

SN - 1523-4614

IS - 1

ER -