Rapid improvements with no commercial production: How do the improvements occur?by Jeffrey L. Funk, Christopher L. Magee

Research Policy


How do I provide leukapheresis products? Blood center experience and evidence for process improvement

Yelena Ginzburg, Debra Kessler, Manlio Narici, Melinda Caltabiano, Mark Rebosa, Donna Strauss, Beth Shaz

When and how much do average forecasts improve predictive accuracy

David Whitlark, Michael Geurts

An Improved Commercial Fluorine Cell

C. A. Powell, R. E. Cable, W. K. Henderson

Do Introduced Grasses Improve Forage Production on the Northern Mixed Prairie

Walter D. Willms, Toby Entz, Ryan Beck, Xiying Hao


Please cite occur? Re

ARTICLE IN PRESSG ModelRESPOL-3079; No. of Pages 12

Research Policy xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Research Policy jo ur nal ho me page: www.elsev ier .com/ locate / respol

Rapid i uc improv

Jeffrey L. a National Univ Engin b Engineering S enue, a r t i c l

Article history:

Received 16 A

Received in re

Accepted 4 No

Available onlin







R&D techn ercia ulat e. Th anism s cha to e al tim chan 1. Introduction

Rapid im nologies en and large im tant issues how do the a technolog on the facto duction. In costs fall as tory more e better at ta 1936; Argo and Thomp et al., 2000) 1994), and

Some sc proxy for e cess design (Lieberman and Lieberm ∗ Correspon

E-mail add learning by experience suggests that all of the improvements in performance and cost for a technology can be considered endogenous http://dx.doi.o 0048-7333/© this article in press as: Funk, J.L., Magee, C.L., Rapid improvements with no commercial production: How do the improvements s. Policy (2014), http://dx.doi.org/10.1016/j.respol.2014.11.005 provements in the cost and performance of new techable technological discontinuities (Christensen, 1997) provements in productivity (Solow, 1957), two imporwithin the fields of management and economics. But se improvements occur during what Dosi (1982) calls ical trajectory? For cost, most of the literature focuses ry floor and links cost reductions with cumulative prowhat has been termed learning by doing (Arrow, 1962), firms learn to produce a single design in a single facfficiently and thus with lower costs. Workers become sks and firms introduce better work flows (Wright, te and Epple, 1990; Adler and Clark, 1991; Thornton son, 2001), better process control (Argote, 1999; Lapre , and automated manufacturing equipment (Utterback, promote organizational learning (Benkard, 2000). holars consider cumulative production as a general ffort and thus the driver of new product and pros and thus improvements in performance and cost , 1984; Dutton and Thomas, 1984; Balasubramanian an, 2010). This formulation which is sometimes called ding author. Tel.: +65 98560674. ress: etmfjl@nus.edu.sg (J.L. Funk). to a model linking cumulative production to the improvements (Dutton and Thomas, 1984; Ayres, 1992; Weiss et al., 2010; Nagy et al., 2013) where the relative contribution of factory floor activities and new product and process designs are unclear. On the other hand, a few scholars have questioned the importance of cumulative production and demand and the possibility that R&D effort or time may be a better independent variable (Koh and Magee, 2006, 2008;

Nemet, 2009; Nordhaus, 2009; Thompson, 2012; Funk, 2013a,b).

This paper attempts to better understand the impact of product and process design changes vs. factory floor activities on cost and performance by detailed analysis of improvements in cost and performance in a novel empirical domain. It focuses on new technologies that have experienced rapid improvements in cost and performance before commercial production has been started and it examines the specific mechanisms that enable these improvements to occur. An analysis of these mechanisms enables us to identify more specific modes of learning and to extend models of learning (Argote and Epple, 1990) into the pre-commercialization phase that some define as invention (Arthur, 2007). Our analysis suggests that key aspects of this learning include creating new materials, improving processes, and reducing scale and that this learning occurs in laboratories.

A second contribution of the paper is for theories of invention. Building from others (Fleming, 2001; Fleming and Sorenson, rg/10.1016/j.respol.2014.11.005 2014 Elsevier B.V. All rights reserved.mprovements with no commercial prod ements occur?

Funka,∗, Christopher L. Mageeb ersity of Singapore, Division of Engineering and Technology Management (EA-5-34), 9 ystems Division, Massachusetts Institute of Technology, N52-395, 77 Massachusetts Av e i n f o pril 2014 vised form 27 October 2014 vember 2014 e xxx a b s t r a c t

This paper empirically examines 13 ments occurred even while no comm reductions through increases in cum improvements from a new perspectiv production cases arises through mech anisms – materials creation, proces improvements to occur and use them can also apply during post-commerci ative contributions of these three me technologies.tion: How do the eering Drive 1, Singapore 117576, Singapore

Cambridge, MA, United States ologies in which significant cost and performance improvel production occurred. Since the literature emphasizes cost ive production, this paper explores cost and performance e results demonstrate that learning in these pre-commercial s utilized in deliberate R&D efforts. We identity three mechnges, and reductions in feature scale – that enable these xtend models of learning and invention. These mechanisms e periods and further research is needed to quantify the relisms and those of production-based learning in a variety of © 2014 Elsevier B.V. All rights reserved.

Please cite ith occur? Re

ARTICLE IN PRESSG ModelRESPOL-3079; No. of Pages 12 2 J.L. Funk, C.L. Magee / Research Policy xxx (2014) xxx–xxx 2001; Arthur, 2007), the three mechanisms for the improvements in cost and performance suggest that product and process design concepts are improved over time in a recursive process during a transition from invention to commercialization. We view this transit which the t ing set of ap begins.

This pap ing learning and analyz improveme marized. Th of 13 diffe the relatio levels of co of the tech presented. anisms mi commercia theories of policy. 2. Literatu

Since th costs in 19 cumulative straight line

As analyzed doing (Arro tion for imp work on lea cific factori changes me tory produc to technolo new factori mance, albe (Ayres, 199 ships (Thor semiconduc been analyz across sign factories.