Optimal new product launch timing in the FMCG-industry
Masterthesis by Jan Warmolt Bödeker
MSc Business Administration: Strategic Innovation Management
June 2015
Rijksuniversiteit Groningen Faculty of Economics and Business
Student no.: s1764691 j.w.bodeker.1@student.rug.nl
Supervisor: Dr. W.G. Biemans Co-assessor: Dr. K.R.E. Huizingh
ABSTRACT
In today’s product markets intensified competition pressures firms to move new products to market faster, for which optimal new product launch timing is considered a critical factor in determining ultimate product success. Empirical research on the link between launch timing and new product performance has produced mixed results. Ambiguity exists on the effects of premature versus delayed launches on new product perfor-mance. This research attempts to fill this literature gap by analyzing the implications for new product per-formance that premature and delayed product launches have, thereby taking into account the moderating ef-fect of product innovativeness. A survey was carried out that collected data on a sample of 60 new product launches in the FMCG-industry. Launch timing was assessed relative to internal firm-driven factors and ex-ternal, market-driven factors that contribute to a possible acceleration or delay of the product launch. This paper shows that delayed launch timing is a driver of decreasing new product performance. This effect is strengthened by product innovativeness, amplifying the negative effects of launch delays on new product performance for increasing levels of innovativeness. No support was found for the hypothesized relation of premature product launches on new product performance. This paper contributes to the field of new product development and innovation literature, by clarifying the relation between launch timing and new product performance from the highly competitive context of the FMCG-industry.
TABLE OF CONTENTS
1. INTRODUCTION ... 5
2. THEORY AND HYPOTHESES DEVELOPMENT ... 6
2.1 New product launch timing ... 7
2.2 New product innovativeness ... 10
3. RESEARCH METHODOLOGY ... 12
3.1 Sample and data collection ... 12
3.2 Measure development ... 13
3.3 Focal firm & launch planning process ... 14
3.4 Measures ... 15
3.4.1 Launch timing ... 15
3.4.2 New product performance ... 15
3.4.3 Product innovativeness ... 16 3.4.4 Control variables ... 17 4. RESULTS ... 17 4.1 Factor analysis ... 17 4.2 Reliability analysis ... 19 4.3 Regression analysis ... 19 5. DISCUSSION ... 23 5.1 Quantitative results ... 23
5.2 Enriching the results with qualitative insights ... 25
6. CONCLUSION ... 27
7. REFERENCES ... 28
8. APPENDICES ... 32
Appendix A – Survey ... 32
1. INTRODUCTION
The pace at which innovations are introduced is becoming more and more critical in determining successful firms (Calantone & Di Benedetto, 2012). Today’s highly competitive markets, e.g. commodity goods, hi-tech or fast moving consumer goods (FMCG), are typically characterized by large groups of suppliers attempting to address the diver-sified needs of large groups of consumers. Intensi-fied competition pressures firms to move new products to market faster, for which it is becoming more important to speed up the new product devel-opment (NPD) process (Langerak & Hultink, 2006). Shortening product life cycles render prod-ucts obsolete sooner and more often. But is this trend of shortening NPD cycles really beneficial to new product performance or does it have its limits? Product launch timing is expected to be more and more a critical success factor in determining the successful firm. Conventional literature suggests that pioneering firms launching new products ahead of competition obtain first-mover advantages potentially resulting in dominant market positions. However, the relation between new product launch timing and resulting product market performance is not as straightforward as might be suggested. The market entry has to be timed to balance the risks of premature entry against the missed opportunities of a late entry (Lilien & Yoon, 1990). Finding the op-timal time frame to launch a new product requires balancing out these risks and opportunities, espe-cially concerning the rapid pace of technological change and intensified global competition (Calan-tone & Di Benedetto, 2012).
Excellent launch timing relative to customers and competitors is considered a critical factor contrib-uting to ultimate product success (Benedetto, 1999; Su & Rao, 2011). Empirical research to date has
focused on the link between launch timing, its close concepts of development speed, cycle time and time-to-market, and new product performance (Hart & Tzokas, 2010). However, no clear consen-sus can be drawn from the mixed results with re-gards to the effect of launch timing on new product performance. Ambiguity exists with regards to es-pecially in the case of a delayed or premature launch (Langerak & Hultink, 2006; Calantone & Di Benedetto, 2012; Langerak, Hultink & Griffin, 2006). Especially in multinational firms, product launches are often delayed or accelerated, without carefully taking into account the possible negative effects on new product performance. Current litera-ture has emphasized on the potential benefits of shortening the NPD cycle to be able to quickly go-to-market with a new product. However, no re-search to date has emphasized on the effects of a premature versus a delayed launch on new product performance.
Langerak & Hultink (2006) provide an analysis on the link between development speed and new product profitability where they find that there is an optimal development speed when it comes to new product performance. This suggests that develop-ment speed should not be decreased or increased to an extreme as this becomes detrimental for new product performance. Calantone & Di Benedetto (2012) found no support of a direct relation be-tween launch timing on performance, and found only a moderating effect of launch timing on the link between a lean launch and new product per-formance.
interchangeably but imply different implications of the concept of launch timing. Namely, cycle time is fully focused on internal process time and does not bring market-related factors into the equation (Langerak, Hultink & Griffin, 2008). Furthermore, recent research has emphasized on representative product launches from different multi-firm (e.g. Langerak & Hultink, 2006) and multi-industry per-spectives (e.g. Calantone & Di Benedetto, 2012), all implying different results and giving room for selection bias, multi-industry and -firm factors that influence the analysis. Hence, a revalidation of this research is needed. Therefore, this research takes an in-depth, single industry and a holistic approach to launch timing – emphasizing on both firm-related timing dimensions and market-driven fac-tors – to rule out these possible confounding facfac-tors on the relation between launch timing and new product performance (Langerak, Hultink & Griffin, 2006; Lilien & Yoon, 1990; Langerak & Hultink, 2006). The emphasis lies on validating a possible direct relation between the effects of premature and delayed launch timing on new product perfor-mance, thereby assessing launches that range from low to high product innovativeness. Hence, the main research question is as follows: what is the effect of premature versus delayed launch timing on new product performance under different condi-tions of new product innovativeness?
The present study identifies 60 new product launches from a multinational firm active in the FMCG-industry, providing a great potential to make a valid assessment of the effects of a prema-ture versus a delayed launch on new product per-formance. A retrospective analysis will attempt to (1) validate the effects of a premature product launch versus a delayed launch on new product performance and (2) the impact of product
innova-tiveness on this relationship. The paper’s practical relevance is underlined by the managerial implica-tions that can be derived from the research. These insights are related to generating more managerial consciousness in the case of accelerating or delay-ing a launch – that management is aware of the ac-tual effects of such decisions with regards to devi-ating from the initial launch timing. Hence, the in-sights of this paper contribute to managerial deci-sion making on new product launch (planning) strategies. The academic contribution this paper makes on the field of innovation and strategic man-agement literature is that it attempts to merge the existent differing perspectives of the launch timing concept by applying a more holistic perspective of launch timing. The insights will contribute to a re-validation of the launch timing impact on new product performance that has not been fully em-phasized by empirical research to date. The re-search approach is adequate in the light of the high-ly competitive context of the FMCG-industry, as this is an industry where timing is crucial as prod-ucts quickly follow-up by competition.
This paper is structured as follows. In the next chapter relevant innovation management and mar-keting literature is reviewed, hypotheses are formed and the derived conceptual model will be presented. Then the research methodology is pre-sented, followed by a discussion of the results. The paper finalizes with a conclusion and directions for future research.
2. THEORY AND HYPOTHESES
DEVELOPMENT
hy-potheses are developed. At the end of the chapter the hypothesized relations will be graphically pre-sented in the derived conceptual model.
2.1 New product launch timing
In the decision process that evolves around a new product launch, several strategic and tactical deci-sions are made. Strategic decideci-sions entail the nature of the product and product positioning within its target market (Hultink, Hart, Robben & Griffin, 2000). Tactical launch decisions include deciding on the traditional marketing mix, i.e. decisions re-garding pricing, branding, sales-, marketing- and distribution support are made (Di Benedetto, 1999; Calantone & Di Benedetto, 2007). A new product launch strategy entails both the strategic and tacti-cal launch decisions that describe what, why, where, when and how to launch (Langerak, Hultink & Robben, 2004). In the light of the current re-search scope, product launch timing is defined as the time of the launch execution of a new product, i.e. when the new product becomes commercially available in the market place. Effective launch tim-ing implies a trade-off that balances the risks of premature entry against the missed opportunity of late entry (Lilien & Yoon, 1990). Typically the launch timing decision is out making a strategic trade-off between pioneering or pursuing a follow-er entry, and thus when to launch a new product
versus competition. An early launch ahead of com-petition could imply high performance opportuni-ties resulting from early market dominance but also brings high risks of premature entry (see also Table 1). A late launch, thereby lagging behind competi-tive product introductions, brings more market cer-tainty due to more market development, whilst also a lower performance potential is expected due to missed opportunities. In this context the adequate notion of the strategic window for launching a new product is the starting point for this research. The ‘open’ strategic window symbolizes the optimal time at which to capitalize on a market opportunity (Abell, 1978). There are only limited periods dur-ing which a fit between target market and the firm’s competency of competing in the market with a certain product is at an optimum (Abell, 1978; Calantone & Di Benedetto, 2007).
Recent literature has shown that trends such as rap-id changes in technology, globalization and intensi-fying competition have caused the strategic win-dow to narrow in many industries (thus the optimal launch time interval is becoming narrower) (Abell, 1978; Ofek & Sarvary, 2003). An extensive amount of previous research has identified the risks and opportunities that arise from an early entry as pioneer and the late entry as a follower. The key advantages and disadvantages of an early versus a late entry are summarized in table 1.
Table 1: Advantages and disadvantages of early versus late entry
Advantages Disadvantages
Early entry (pioneer) Occupy preferred market position leads to an
enhanced competitive position
High costs for technological, product and market development
Respond quickly to rapidly changing mar-kets
Risks of rival innovation driving better or more profitable innovations
and experience effects market development and no dominant design set
Relative pricing freedom as pioneer, high profit margins
High demand uncertainty
Recognition and reputation establishment creates willingness-to-pay and customer loy-alty
Risk of imitation by later entrants that are able to reap benefits sooner
Early entry negatively affects existing prod-uct sales due to cannibalizing effects Sources: Ali, Krapfel, LaBahn, 1995; Urban, Carter, Gas-kin & Mucha, 1986; Cooper &
Klein-schmidt, 1994; Nie-drich & Swain, 2004; Langerak & Hultink, 2006; Zhang & Markman, 1998
Late entry (follower) Followers are most successful when superior
products are developed (supported by strong promotional spending and aggressive pric-ing)
Requires substantial marketing investments to gain market share
More time invested in product development and optimization
Missed market opportunities imply a lower performance potential due to possible prod-uct obsolescence and opportunities taken by competition
High costs of entry due to entry barriers by established players
Sources: Urban, Carter, Gas-kin & Mucha, 1986; Cooper & Klein-schmidt, 1994; Nie-drich & Swain, 2004; Langerak & Hultink, 2006; Zhang & Markman, 1998; Lilien & Yoon, 1990; Cordero, 1991
As table 1 shows there are both potential risks and benefits attached to the timing aspect of new prod-uct launches. There is a stream of literature that assumes an early launch to be beneficial for new product performance, as speed of launch and de-creasing cycle times enhances competitive ad-vantage by beating competition in being first to market (Lambkin, 1988; Cooper & Kleinschmidt, 1994). Also, being able to launch products early in the market implies a quick and flexible respond to rapidly changing markets and shortening product lifecycles (Langerak & Hultink, 2006). Being the first in the market implies a higher potential for
the risk of failure is high because the potential de-mand is not known with certainty. Uncertainty and high investments in market testing and develop-ment have an adverse effect on the possible high performance that can be achieved whilst being the first in the market that enables to set your own price and win an early loyal segment that possibly results in a dominant market position (Langerak & Hultink, 2006).
On the other hand, the aforementioned pioneering risks are substantially lower or even not present in the case of a follower entry strategy due to a devel-oped market and established technology. Also, the development costs of entry may be lower since the innovator has created the primary demand and the basic product design already exists in the market (Urban et al, 1986). On the other hand, costs of entry may be high due to established entry barriers by competitors (Cooper & Kleinschmidt, 1994). The risks taken are substantially lower, however also leveraged on the performance potential that might be generally lower for a late entrant in the market. Barriers of entry due to established tech-nology by competition or product obsolescence play an important role. This causes a later product entry to potentially even fail as the changeover is not embraced by the market. Next to this initial market opportunities are already reaped by compe-tition, by creating their own loyal consumer seg-ments thereby achieving an established and possi-bly dominant market share (Lilien & Yoon, 1990). Hence, a late entrant generally faces a lower per-formance potential for its new product due to an established market demand taken by competition and therefore high investments need to be done in marketing the new product to win consumers from competition (Urban & Hauser, 1993).
product launch not only is impacted by the (mini-mum) cycle time of the NPD process, also external factors may play a role as competitive pressures urge firms to pioneer with new product forms or follow-up with improved products. As is the case in industries, such as the FMCG, where under the effects of intensified competition product life cy-cles have shortened. With regards to launch timing, there is a narrow strategic window of opportunity for new products to succeed in the market that symbolizes the product-market combination that needs to fit.
Research until now has emphasized cycle time an important aspect in timing new product launches as it decides the minimum timeframe needed for a product to be ready for launch. However this does not take into account market factors that pressure the firm to either accelerate or slow down the NPD process, as a new product is best launched within its strategic windows. Ideally, a new product launch is well timed if it is timed according to competition, customers, distribution channel and business-unit goals (Calantone & Di Benedetto, 2012). If there exists an optimum cycle time, I as-sume an optimal launch time to exist. The strategic trade-off that has to be made whilst deciding on the optimal launch date means balancing out the aforementioned risks of premature entry against the missed opportunities of a late entry. Concerning the different pioneering and following advantages and disadvantages resulting from the literature, several assumptions are made.
1. Advancing or accelerating a product launch implies an overall increasing performance po-tential due to early market development and the potential to lock-in consumers and achieve dominant market positions.
2. Accelerating a product launch implies higher risks due to i.e. market uncertainty, stressed resources that could result in production errors and higher costs of market development as a pioneering product implies high investments in building awareness and consumer learning. 3. Delaying or postponing a product launch is
assumed to be detrimental for new product performance, as established players in the market have preoccupied preferred market po-sitions and high investments need to be done to overcome entry barriers and marketing the new product to win market share.
Hence, in either case a premature launch or a de-layed launch is on average assumed to be negative-ly impacting new product performance. Any launch that differs from the launch time optimum will therefore have lower resulting product performance versus the optimum leading to the following hy-potheses:
Hypothesis 1a: Early launch timing negatively im-pacts new product performance.
Hypothesis 1b: Late launch timing negatively im-pacts new product performance.
2.2 New product innovativeness
In this paper I consider innovativeness not as an independent variable that directly impacts new product performance, but rather as a variable that moderates the relationship between launch timing and new product performance. Modeling innova-tiveness as a moderator has more managerial mean-ing because the framework illustrates link between both constructs under different conditions of inno-vativeness (Lee & O’Connor, 2003). The extent to which the new product is innovative to the firm affects launch timing, as highly innovative prod-ucts take longer to develop and therefore extends the cycle time and thus the earliest possible launch date at which the product can be launched (Griffin, 2002). On the other hand new product performance is influenced by the degree to which the new prod-uct is innovative relative to the current market standard (Robinson, 1990). Empirical researches have shown that more innovative new products are associated with slower development speeds and higher new product performance (Langerak & Hultink, 2006). Furthermore, whilst rushing the development of a new product it might not be in-novative enough for the market, hence the strategic window will still be closed on launch that will be detrimental for performance (Ali, 2000). On the other hand, taking too long and thereby launching a product that is less innovative will cause the strate-gic window to be closed, hence the opportunity to maximize profit is gone. This underlines the im-portance of the moderating role that new product innovativeness has on the relation between compet-itive new product launch timing and new product performance. For a comprehensive view on the
im-pact of product innovativeness on this relationship, both product newness to the firm and to the market will be considered.
The more innovative the new product project, the more incompatible the development process is with the existing firm base of knowledge, skills and pro-cesses that require changes to come up with an in-novative product (Ali, 2000). Next to that, taking to little time to develop an innovative product implies risks of customer inexperience with the product and an underdeveloped market. On the other hand, taking too much time for developing a less innova-tive product, cause firms to face an already closed strategic window. Customers could already have switched to another supplier that delivers more novative products as competitors already have in-troduced similar innovations in their product as-sortment. Thus, consumers are already exposed to existing products and they are unlikely to postpone their purchase decision to await only minor im-provements (Langerak & Hultink, 2006). In sum, the more innovative the new product, the more likely the ideal launch date will be later in time for achieving optimal new product performance. The interaction effect of innovativeness is hypothesized as follows:
Hypothesis 2a: More innovative products strength-en the negative relation of early launch timing on new product performance.
Hypothesis 2b: More innovative products strength-en the negative relation of late launch timing on new product performance.
Figure 1: Conceptual model
3. RESEARCH METHODOLOGY
In this chapter the data collection methods will be presented, by describing the research sample and respondents, method of data collection, measure development and a brief background is sketched of the NPD planning process of the focal firm in this research.
3.1 Sample and data collection
This paper used a retrospective analysis of a sam-ple of 60 new product launch projects from a mul-tinational firm active in the fast moving consumer goods (FMCG) industry. This industry is character-ized by short product life cycles, fierce competi-tion, rapidly evolving technologies and demanding consumer needs (Lilien & Yoon, 1990). For the scope of this research, the consumer goods industry has adequate dynamics as new products are often introduced and are quickly followed-up by compet-itors. Hence, timing is a ‘critical success factor’ in the industry, as mentioned by interviewees at the focal firm. The data was collected in a single indus-try and one focal firm as this rules out potential inter-industry effects that may have detracted from most previous new product performance studies
(Cooper & Kleinschmidt, 1994). The focal firm markets products across different product catego-ries, such as cosmetics, healthcare products, laun-dry detergents, skin care products and household goods. The sample of new product launches was collected across these categories using several cri-teria for data collection. Projects had to be repre-sentative for the brand and product category and launched across France, Belgium and The Nether-lands over the past 3 years. Every product launch in one of these countries was considered a separate NPD project because of different market dynamics. To adequately measure initial product performance, the selected projects had to be commercially avail-able for at least six months. New product launches were selected for analysis if they were accredited a new EAN code, implying a ‘hard product changeo-ver’ and supported with a launch plan consisting of a communication and media plot. The ‘hard prod-uct changeover’ implies physical prodprod-uct changes including formula upgrades, new packaging and entirely new product forms.
carry single brand responsibility that entails brand-ing, new product launch planning and execution, consumer research and strategic market planning. They are fully knowledgeable on the required tim-ings of the NPD projects, product innovativeness and resulting commercial performance. Each brand manager was contacted per e-mail and telephone to fill out a presurvey for project data collection. The presurvey consisted of (1) sharing representative new product launches for their brand, and addition-ally (2) qualitative data on project description, mo-tivation of the timing decision and an indication of initial new product performance in the first year of launch and product innovativeness on a scale from 1-10.
After a list of projects was collected, a survey was sent out to all respondents willing to cooperate. To increase the response rates, several reminders were sent out over e-mail and telephone meetings were scheduled with the brand managers to collect their input. Also an initiative for them was given via al-lotting two cheques that could be spent on compa-ny products.
These efforts yielded a sample of 62 new product launches spread across different product categories (i.e. laundry detergents, diapers, shampoos, razors, dental products, perfumes, etc. etc.). Data was col-lected through the responses of 25 managers, a re-sponse rate of 73.5% of the total sample of ABMs that were initially contacted for the presurvey. 62 new product launch projects were collected in total, of which 2 were rejected for analysis, due to 1 in-complete response and 1 case was withdrawn from the analysis as it was a clear outlier in the dataset. The sample contained product launches of which 34.4% (21) launched in France, 16.4% (10) launched in Belgium and 49.2% (30) launched in The Netherlands. Out of the remaining sample of
60 new product launches, 13 were launched on the perceived optimal launch date (rated ‘4’ on the tim-ing scale), 8 launched too early (rated ‘< 4’) and 39 cases were launched too late (rated ‘> 4’). A di-chotomous measurement item for product success or failure was also included. This yielded 77% of the product launches to be rated a financial perfor-mance success and 23% as failures. Furthermore, resulting from a categorizing scale of rating inno-vativeness ranging low to high, 16% of the launch-es were categorized as low innovative, the majority of 52.5% of the launches were rated moderately innovative and 31.1% of the cases were rated as highly innovative. Lastly, 50.0% of the new prod-ucts were launched following competition, and 50.0% of the products were launched first in the market (as measured in a dichotomous scale for pioneer or follower product launch).
3.2 Measure development
and characteristics of each product category. Cor-recting data for these different parameters would hinder retrieving significant results from the re-gression analysis. On top, the sample size is too small to separately assess every product category as a market in itself ruling out these disturbing ef-fects. Instead, collecting subjective data on launch timing, new product performance and innovative-ness, facilitates comparison across product catego-ries and their subsequent different product markets. Using perceptual data to measure these constructs is common in innovation research to enable cross-market and -category comparisons (Atuahene-Gima & Ko, 2001). A recognized problem of using subjective data is respondent bias, but this is be-lieved to be less problematic for more concrete items such as financial product (Blindenbach-Driessen, Van Dalen & Van den Ende, 2010). Multiple measurement items were generated using the literature and interviews with academics and practioners at the focal firm. Multi-item scales were adapted and derived from literature to meas-ure each construct as well as the control variables. The use of multi-item scales enhances reliability and validity and is effective for the domain of new product performance (Henard & Szymanski, 2001) and launch timing (Calantone & Di Benedetto, 2007). The survey was developed and then pretest-ed in face-to-face interviews with two academics and three brand managers at the focal firm, whom were asked to evaluate whether these measures were appropriately measuring the constructs and if there were any questions unclear, difficult to inter-pret or to respond to (Langerak & Hultink, 2006). The pretest ensured that all questions were clear and that the scale items adequately represented the desired constructs (Calantone & Di Benedetto, 2007). Problematic items were then reviewed or
eliminated from the survey and new ones were added. The items were then reviewed for another time with the participants, and after no further comments the survey was ready for distribution.
3.3 Focal firm & launch planning process Procter & Gamble Europe is organized through a strategic head office in Geneva which is the Global Business Unit (GBU) Europe. The European mar-ket is split up in several Sales & Marmar-ket Organiza-tions (SMOs) that are operational and executional units serving several consumer markets reporting to the GBU in Geneva. For example, the SMO in the scope of this research, consists of France, Belgium, Netherlands and Luxembourg (FBNL) and serves as one operational unit within the P&G company structure. The NPD process at P&G takes place as a responsibility of the GBU. R&D continuously works on innovating the current product portfolio and P&G works mostly through a ‘consumer push’ approach, in which P&G is considered a pioneer with innovations that are pushed in the market, ra-ther than a ‘pull’ approach.
SMO or GBU experiences internal or external pressures that cause the SMO to fail to execute on the intended launch date. Internal effects could be a delay in plant production capacity for the new product to be launched full-scale, external market-related effects could also pressure the launch to be rescheduled. Possible pressures could be to align launch timing with fixed dates on which key cus-tomers change their shelve assortment.
3.4 Measures
3.4.1 Launch timing
To come up with a holistic assessment of the launch timing construct, both emphasizing on firm-related timing dimensions and more external, mar-ket-driven factors, a multi-item scale was adapted from Calantone & Di Benedetto (2007; 2012). The measurement items assess the extent to which the timing of the product launch was optimal with re-gards to a range of stakeholders and internal objec-tives. Appropriateness of the timing decision was measured using 7 item Likert scales (1 = too early, 4 = optimal, 7 = too late). Timing was measured by the following items:
1. overall product launch timing, 2. consumer read-iness, 3. direct competition, 4. major customers, 5. top management beliefs, 6. customer shelf muta-tion, 7. competitive advantage, 8. business unit ob-jectives and 9. timeliness of channel/trade promo-tion.
In addition to above items this paper attempted to generate some qualitative insights on top of the quantitative analysis of the data. In the survey an open question was included that asks for a motiva-tion why the respondent thought the launch was suboptimal and why it should been launched earlier or later then the actual exercised date. This
con-tributes to generating also a qualitative judgment of premature versus delayed launch timing.
3.4.2 New product performance
Perceptual data was collected to assess the new product performance of the projects in the research sample. The initial new product performance was measured in the first year of launch. Namely, sev-eral environmental and market variables which are beyond the control of managers, influence long run market performance. In that case it would be diffi-cult to isolate the effect of launch timing on new product performance (Ali, Krapfel & LaBahn, 1995). Also, as brand managers were asked to pro-vide their opinion on relatively detailed matters, product launches that were done more than five years ago was not suitable concerning their memory.
In this paper, new product performance implies the commercial performance of the launched product in the first year of its commercial availability. Commercial performance can be measured accord-ing to two broad and distinct measure categories of new product performance, namely operational per-formance (1) and product perper-formance (2). The first relates to meeting project goals such as adher-ence to schedule, budget and quality requirements, whereas the latter covers the commercial (financial and market) performance for the new product (Blindenbach-Driessen et al, 2010). Multiple new product performance measures are used in this study, due to the multidimensionality of the con-struct (Cooper & Kleinschmidt, 1994). Six measures for new product performance are adapted from Calantone & Di Benedetto (2007) that cover the latter measure category of product perfor-mance.
Kleinschmidt, 1987; Hultink and Robben, 1995), therefore performance was measured on 3 items, namely volumes shipped, revenues and achieved market share, measured relative to business unit objectives and business unit’s other new product launches. Adapted from Calantone & Di Benedetto (2007) is a seven-item scale of performance using Likert scales ranging from 1 (far below objective) to 10 (far exceeded objective). Additionally a di-chotomous success measure was added, adopted from Cooper & Kleinschmidt (1994). These items are all measured on the NPD project-level. The fol-lowing items were included:
1. Overall success of this market entry from profitability standpoint;
2. Performance relative to business unit objec-tives – in terms of a. volume (shipments), b. revenue / sales and c. market share.
3. Performance relative to the business unit’s other new product launches was also measured in terms of a. volume (shipments), b. revenue / sales and c. market share.
Lastly, the dichotomous success measurement item asks to classify a launch as a success or failure concerning the initial product performance.
3.4.3 Product innovativeness
Booz, Allen & Hamilton (1982) proposed six new product typologies that have been widely used to categorize various degrees of innovativeness along the two dimensions. This typology classifies new products by the level of newness in relation to both the market and the firm. In this paper I adapt the self-typing approach as used by Langerak & Hultink (2006) to measure innovativeness, as well as a continuous scale adapted from Ali, Krapfel & LaBahn (1995) in order to analyze the moderating effect of innovativeness. Two different scales of innovativeness were used to perform the
hierar-chical regression analysis. The categorized varia-bles1 were thereby used a. as dummy variables and thereby checking the significance under three (low, moderate, high) conditions of innovativeness and b. the continuous scale of innovativeness was (finally) used for a more granular interaction analysis. I use three categories of product innovativeness applied by Kleinschmidt & Cooper (1991) that were adapted from the original typologies from Booz et al (1982). The self-typing approach was used for several reasons. First, this approach is straightforward and logically appealing. Second, measuring product innovativeness objectively is virtually impossible. Third, this approach has prov-en to be reliable and valid in related research evprov-en with a limited number of variables (James & Hat-ten, 1995; Langerak & Hultink, 2006). The re-spondents were asked to assign the new product to one of the three categories: (1) highly innovative (i.e. new-to-the-world products) – (2) moderately innovative (i.e. new product lines to the firm, not new to the market) – (3) low innovative (i.e. modi-fications or revisions of existing products). Re-spondents were asked to assign the new product to one of these three categories.
Four additional measurement items of product in-novativeness were adopted from Ali et al (1995) to also develop a continuous scale of product innova-tiveness. Three items were measured on a 7-item Likert scale ranging from low to high and the fourth item was measured dichotomously (first in the market yes / no).
Categorizing the product launch into three ranges of innovativeness:
1The categorizing measurements did not provide
1. Highly innovative: new-to-the-world products and innovative new product lines to the firm 2. Moderately innovative: new product lines to
the firm, but the product is not new to market, and new products in an existing product line 3. Low innovative: modifications of existing
products, revisions of existing products, repo-sitioning of existing products and cost im-provements
Measure items of product innovativeness (continu-ous scale): Ali et al, 1995:
1. Innovativeness of the product to the market 2. Unique features for customers
3. Product technology is new to the customer 4. Market pioneering (first into the market with
type of product (yes/no))
The questions used in the survey to assess new product performance and product innovativeness can be found in Appendix I.
3.4.4 Control variables
A wide number of studies focused on the determi-nants of new product performance. We therefore need to control for other effects on new product performance by adding control variables into the analysis (Montoya-Weiss, 1994). In this paper three market environmental variables are used to control for other effects on new product perfor-mance. These three variables are irrelevant for the scope of this study that emphasizes on the relation between launch timing and new product perfor-mance. However, previous studies have shown that these environmental variables may influence new product performance (Lee & O’Connor, 2003). Hence, these variables are controlled in the regres-sion analysis.
The three control variables were adapted from Ja-worski & Kohli (1993) and all items were meas-ured on a 5-point Likert scale. Market turbulence
items assessed the extent to which the composition and preference of customers tended to change over time (e.g. ‘Our customers tend to look for new product all the time.’). The items measuring com-petitive intensity assessed the behavior, resources and ability of competitors to differentiate (e.g. ‘One hears of a new competitive move almost eve-ry day.’) and the items measuring assessed the ex-tent to which technology in an industry changed rapidly (e.g. ‘It is very difficult to forecast where the technology in our industry will be in the next 2 to 3 years.’).
1. Market turbulence 2. Competitive intensity 3. Technology turbulence
4. RESULTS
In this chapter the results are presented of the anal-ysis. First, the factor analysis, reliability analysis and normality checks are discussed. Then the chap-ter presents the regression results.
4.1 Factor analysis
0.001). In the factor analysis three constructs were retained. The three constructs all exceeded eigen-values of 2.0, thereby exceeding Kaiser’s criterion of >1.0, and explained 83.891% of the variance. A scree plot justified retaining three factors as these were plotted before the point of inflexion. The PCA was carried out using an oblique rotation (Di-rect Oblimin) in SPSS v20. This rotation was pre-ferred as it provided the clearest loadings on each of the constructs without problematic cross-loadings and cross-loadings on the wrong construct. An oblique rotation method is also preferred when items are expected to be somewhat related (Tabachnick & Fidell, 2007), e.g. as is the case with innovativeness and new product performance that were found to be directly correlated (Calantone & Di Benedetto, 2007). As selection criterion for
item loadings a covariance of > 0.40 was applied, as suggested by Stevens (2002).
Table 2 shows the factor loadings after rotation and deletion of two items, intended to measure launch timing;
− TIM5_2, measuring the relative timing to con-sumer needs & behavior;
− TIM5_6, measuring timing relative to the dis-tribution channel.
These items were withdrawn from the analysis as they were cross-loading on two different constructs by more than 0.40. The resulting pattern matrix (table 2) shows that component 1 represents new product performance (measured by all 7 items), component 2 innovativeness (all 3 items) and com-ponent 3 launch timing (remaining 7 items).
Table 2: Pattern matrix
Component Extraction method:
Principal Component Analysis.
Rotation method: Direct oblimin with Kaiser Normalizationa. a. Rotation con-verged in 6 itera-tions. Component 1: New product performance Component 2: Prod-uct innovativeness (continuous scale) Component 3: New product launch tim-ing
Item 1 2 3
Relative to category objectives, how successful was this product launch in terms of shipments?
0.984
How successful was this product launch from an overall stand-point?
0.950
Relative to category objectives, how successful was this product launch in terms of NOS (sales)?
0.933
Relative to other new product launches in the category, how successful was this product launch in terms of shipments?
0.888
Relative to category objectives, how successful was this product launch in terms of value market share?
0.877
Relative to other new product launches in the category, how successful was this product launch in terms of NOS (sales)?
0.873
Relative to other new product launches in the category, how successful was this product launch in terms of value market share?
0.871
The innovativeness of the product to the market is -0.882 The product technology is new to the consumers -0.879 The product has unique features for consumers -0.795
The overall timing of the product launch was 0.813 Relative to our planned launch date, the timing of this launch
was
0.723
From the point of view of our major customers, the timing of this launch was
Relative to the customer shelf mutation, the timing of this launch was
0.666
With regards to the achievement of a competitive advantage with this product, the timing of this launch was
0.641
Top management believed the timing of this launch was 0.556 Relative to our direct competition, the timing of this launch was 0.421
4.2 Reliability analysis
After a consistent factor pattern was found, the re-sulting scales for each factor were tested on con-struct reliability. All concon-structs showed high meas-urement reliabilities, with Cronbach’s Alphas α > 0.80 that exceed the reliability criterion of α > 0.70 (Kline, 1999) (table 3). Reliability for the control variables was also assessed. Applying the reliabil-ity criterion of α > 0.70, one item measuring mar-ket turbulence (MTB12_3) was deleted. Also, two items measuring competitive intensity (CIN13_6) and technological turbulence (TTB14_5) had to be rescaled as they were reverse phrased. This resulted in Cronbach’s Alpha’s of α > 0.70 for competitive intensity and technological turbulence. Market tur-bulence was slightly below at α = 0.653, but still acceptable (see also table 3).
Table 3: Reliability statistics
Construct Cronbach’s Alpha
No of items New product performance 0.973 7 New product launch timing 0.780 7 Product innovativeness 0.903 3 Control variables
Market turbulence 0.653 3 Competitive intensity 0.841 6 Technological turbulence 0.811 5
This studies’ respondents were managers that were personally responsible for the product launch and the efforts for new product performance, one must
check for biases that arise from single informant and single instrument, as the rating of launch tim-ing, innovativeness and performance could be un-der- or overstated that results in common method bias. Common method bias is problematic as measurement errors of the independent variables are correlated significantly with those of the de-pendent variables. As it is a potential threat for this research and the interpretation of its outcomes I tested for this. A Harmon’s single factor test was carried out using confirmatory factor analysis in SPSS v20 (Harman, 1976; Calantone & Di Bene-detto, 2012). This test did indeed shows a common method bias can be assumed, as the explained vari-ance of a single factor accounted for the majority of the variance in the model (67.98%). This is pos-sibly due to a lack of precautions taken for prevent-ing such measurement errors; as items in the survey were not randomized, scales were not reverse scored nor were irrelevant ‘filler’ scales included – all methods that could account for preventing common method bias to arise in the dataset. This makes the survey items susceptible to response styles resulting into common method biases (Lin-dell & Whitney, 2001).
4.3 Regression analysis
4.762, far exceeding the acceptable range of -1.96 < x < 1.96 (Aiken & West, 1996) and heterogene-ous variance as shown in a Shapiro-Wilk test (sig-nificant at p < 0.01) implying a non-normal distri-bution and therefore not complying with the as-sumptions of linear regression. To correct for these variables, the data set was checked for possible outliers. Indeed one data point was identified an outlier by calculating Z-scores (standardized val-ues) for the launch timing variable. Namely, any standardized value exceeding |3.29| can be identi-fied as an outlier (Tabachnick & Fidell, 2001). This response was then deleted from the data-set and once again the assumptions were re-tested, result-ing in acceptable measures of skewness and kurto-sis for all constructs within the range of |1.96| (see table 4). The Shapiro-Wilk test of normality showed significant for the launch timing variable implying a non-normal distribution, at D(60) = 0.944, p < 0.05. New product performance showed non-significant, implying a normal distribution, at D(60) = 0.980, p > 0.05. Furthermore to check for
the homogeneity of variances in our dataset, a Levene’s test shows insignificant for all variables, where new product performance eat F(2, 57) = 1.948, p > 0.05 and launch timing F(2, 57) = 0.708, p > 0.05, implying a homogenous distribu-tion of the variables. Lastly, all variables were plot-ted in histograms and Q-Q plots, showing approx-imately normal distributions (appendix 2).
Furthermore, by loading a macro into SPSS derived from Garcia-Granero (2002) effects of possible heteroscedasticity were also tested using a Breusch-Pagan test and Koenker test. Both tests were insignificant at p > 0.05, thereby not rejecting the null-hypothesis of homoscedasticity. Thus, het-eroscedasticity was not found to be an issue in our dataset.
Concluding from the above, we assume the dataset to be approximately normally distributed and pro-ceed with regression, however the results must be interpreted with the limitation in mind that the da-taset is not perfectly normally distributed.
Table 4: Normality statistics
Value Launch timing New product performance Innovativeness N 60 60 60 Skewness -.557 -.261 -.317 Std. Error of Skewness .309 .309 .309 Z-value -1.803 -.845 -1.026 Kurtosis 1.810 -.533 -.950 Std. Error of Kurtosis .608 .608 .608 Z-value 2.977 -.878 -1.563
To test for hypothesis 1, stating early (1a) and late (1b) launch timing both negatively impact new product performance, a hierarchical linear regres-sion analysis was carried out. This relation is
ing the launch was timed too early – to ‘4’ imply-ing an optimal launch timimply-ing – and then ‘7’ implied the launch was timed too late. Hence, this scale was not ordinal and therefore not meeting Likert scale conditions so that an interval scale could be assumed for regression. Hence, any deviation from optimal, rated ‘4’, implied a negative effect so the scale was not ranging in an ordinal manner. In sum: − A rating of < 4 implied the launch timing was
perceived to be too late.
− A rating of > 4 implied the launch timing was perceived to be too early.
So to test for the effects of early versus late timing and the hypothesized negative effects on new product performance, two new variables were cal-culated. First, all cases scoring < 4.0 on the launch timing scale were converted into a new variable representing all cases that were perceived to be launched too early, by subtracting the mean (TIMavg) of the timing variable;
𝐸𝑎𝑟𝑙𝑦 𝑡𝑖𝑚𝑖𝑛𝑔 = 𝑖𝑓 𝑇𝐼𝑀4.0 − 𝑇𝐼𝑀!"# !"# < 4.0
Second, all cases scoring ≥ 4.0 were selected and also converted into another variable:
𝐿𝑎𝑡𝑒 𝑡𝑖𝑚𝑖𝑛𝑔 = 𝑖𝑓 𝑇𝐼𝑀𝑇𝐼𝑀!"#− 4.0 !"# ≥ 4.0
Hence, the data-set was split up representing one split of launches that were perceived to be launched too early, and the second split represented all launches that were perceived to be launched too late. A hierarchical linear regression analysis was carried out in an attempt to identify the effects of a launch that deviated from the optimum to test for H1a and H1b. The variables were mean-centered to overcome possible multicollinearity issues (Aiken & West, 1996). Indeed, the VIF-factors showed multicollinearity was not an issue (ranging from
1.000 in model 1 to 1.980 in model 2, acceptable as VIF should not exceed 10.0).
To test for H1a, the proposed negative effect of early launch timing on new product performance, two regression models were tested. The first model tested for the linear timing effect and the second model tested for the timing and the interaction ef-fect of innovativeness (both models accounted for the possible effects of our control variables). The relationship is supported if the linear term for early launch timing is negative and significant. The re-sults indicate however that this relation was not significant (at p > 0.05), therefore no support is found for a negative relation between early launch timing and new product performance. Hence, H1a is rejected. The second model accounted for the interaction effect of product innovativeness (H2a) on the relation and again the model was found in-significant (p > 0.05), thereby not supporting a strengthening effect of product innovativeness and rejecting H2a. The control variables did not show a significant effect on new product performance in this model.
in-significant (at p > 0.05) – thereby initially reject-ing H2b. Again, the control variables did not show a significant impact in both models.
Finally I tested if the insignificant interaction ef-fects of product innovativeness from model 2 that lead to rejection of hypothesis 2b, would be im-proved in the case of analyzing a subsample of new product successes. Using the dichotomous success measure was rated in our survey, all cases that were rated a new product success were selected by ap-plying a dummy variable (n=41). Interestingly, the regression showed a significantly increased
ex-plained variance of a total of 42.3% (R2=0.423). The interaction effect of late launch timing and in-novativeness on relation between late launch tim-ing and new product performance was also found significant (at p < 0.01) showing an increased negative linear coefficient of B = -1.699 (sig at p < 0.01). These results provide partial support for H2b as the interaction effect of innovativeness on the late launch timing and new product performance relation is only supported for new product success-es. The regression results are summarized in tables 5 and 6 below.
Table 5: Model summary
a. Timing early: not sig. (ANOVA p > 0.05)
b. Timing late: sig. for model 1 (ANOVA p < 0.05)
c. Timing late for new product success: sig. for model 2 (ANOVA p < 0.01)
Significant regression results are reported in table 6
Model R R2 Adjusted R2 Std. error of the estimate
Timing earlya
1 Timing early, Control variables
.882 .778 .483 1.34708
2 Timing early, Control variables and Innova-tiveness interaction
.973 .947 .628 1.14274
Timing lateb
1 Timing late, Control variables
.298 .089* .071 1.97172
2 Timing late, Control variables and Innova-tiveness interaction
.318 .101 .025 2.01982
Timing late for New
product successc 1 Timing late, Control
variables
.256 .065 -.038 1.50686
2 Timing late, Control variables and Innova-tiveness interaction
Table 6: Regression results
*sig. (one-tailed) at p < 0.05
** sig. (two-tailed) at p < 0.01
Model B Std. Error Beta
Timing late Model 1 Constant 6.594 1.531 Timing late -1.409* .789 -.262 CV Market turbulence .173 .451 .065 CV Competitive intensity -.312 .404 -.144 CV Technological turbulence .120 .565 .040
Timing late for New product success
Model 2 Constant 7.170 1.060 Timing late .614 .667 .135 CV Market turbulence .404 .324 .224 CV Competitive intensity .670 .332 .422 CV Technological turbulence -1.240 .460 -.583 Innovativeness .521** .165 .451 Timing late * Innovativeness -1.699** .578 -.448
5. DISCUSSION
This chapter explains the paper’s findings in rela-tion to other studies and interprets the results. First, the quantitative results will focus on the extent to which the hypotheses are significantly supported from the analysis. Second, the chapter will add qualitative insights to the findings of this paper, that will briefly discuss both firm- and market-driven factors of the product launches that were found to be launched premature or delayed.
5.1 Quantitative results
The results show insufficient support for the nega-tive effect of early or premature launch timing on
perfor-mance, especially in cases of higher new product innovativeness (Savin & Terwiesch, 2005).
The negative effect of late or delayed launch tim-ing on new product performance as reflected by H1b was significantly supported. First, a direct negative relation was found between delayed launch timing and new product performance. In other words, delayed product launches – i.e. launches that are perceived to be launched too late relative to market-related factors (e.g. customers and competition) as well as internal objectives (e.g. planned launch date) – are found to be negatively impacting the resulting (initial) new product per-formance. These findings are supported in recent empirical research (Langerak & Hultink, 2008; Song et al, 1999; Kalyanaram & Krishnan, 1997). A launch time delay or late launch gives the new product a shorter growth period and thereby the net sales are expected to be smaller based on a forth-coming competitor’s product introduction that will follow-up soon (Rosenthal, 1992). Kalyanaram & Krishnan (1997) suggest that a delay in new prod-uct launch time causes the prodprod-uct to be launched into a later competitive stage, thereby missing (pi-oneering) opportunities for market penetration and early consumption peaks. In this later stage, it also becomes likely that competition has already intro-duced similar product types, that even further de-creases the effect due to diminished innovativeness (Ali et al, 1995); in support of our finding that in-novativeness even further decreases the perfor-mance effects of delayed launches. A more recent research suggests life-cycle sales for products de-creases due to launch delays, however at a dimin-ishing rate (Savin & Terwiesch, 2005; Pichler & Smith, 2003).
Innovativeness was found to be strengthening the negative relationship between delayed launch
tim-ing and new product performance. However in con-trast with these findings, Savin and Terwiesch (2005) suggest that, for higher values of innova-tiveness, it might also be even beneficial for a firm to delay a launch so potential cost reductions can be achieved. Empirical literature shows more inno-vative new products are linked with slower devel-opment speeds (Ali et al, 1995) and higher new product performance (Robinson, 1990). However, this research shows that taking more time to devel-op a more innovative product has its limits, as our results show a launch delay implies a negative link with new product performance. Moreover, for higher levels of product innovativeness, generally products that take longer to develop, holds launch timing becomes more crucial as a delay implies a more negative tilted downward slope versus new product performance.
ap-proach of launch timing. Also an external perspec-tive of launch timing, taking into account competi-tion, direct customers, consumer readiness etc. brings revalidated results on the relation of launch timing and new product performance. The contri-bution that is delivered to academic literature is that delayed launch timing is negatively impacting new product performance, and is found to be more decreasing for higher levels of innovativeness in cases of new product success. The academic impli-cations of these results are that this paper specifi-cally targets the effects of launch timing on new product performance from both an internal firm and market-driven perspective, thereby contrib-uting to the existing research base on launch tim-ing, NPD cycle time and other related concepts. From these results, also several managerial impli-cations can be drawn. Namely, an interesting con-tradiction is found. Increasing levels of product innovativeness was found to lengthen cycle time at an increasing rate (Ali et al, 1995). This research shows that for higher innovativeness, a launch time delay negatively impact new product performance more than for lower levels of innovativeness, in the case of high performing new products. Thus, for those launches that are most promising for firms, generally high performing products and highly in-novative products, it is crucial for firms to get the launch timing right. Especially in the light of high-ly innovative products that generalhigh-ly take long cy-cle times to develop and are likely to be delayed (Cordeor, 1991). Hence, in the NPD process man-agers should take this into consideration whilst de-ciding on the launch time.
5.2 Enriching the results with qualitative in-sights
well. As large visibility needs to be ensured on launch, winning support of key retailers is crucial in terms of trade acceptance and new product pro-motion to drive penetration and initial product sales. Lastly, seasonality is another driver that was mentioned to be a market-driver of deviation in launch timing. For example, December gift month or summer sales peaks imply crucial performance implications for several categories, e.g. perfumes and beauty products. Most drivers were market-driven, however a couple of internal- or firm-driven pressures of rescheduling launch timing were also given. The focal firm is a multinational company and therefore operating in many regions.
This brings complexity in terms of new product launches, possibly resulting in launch delays. For example, in a couple of cases a launch was delayed due to inconclusive global product positioning strategy or insufficient plant capacity to be able to supply a sufficient product stream to the target re-gion. The aforementioned drivers are summarized in table 7. In sum, these drivers emphasize the im-portance of the holistic measurement approach, taking into account internal and market-driven per-ceptions of launch timing. As shown from these drivers, the main reason for our focal firm to devi-ate from the planned launch ddevi-ate is mostly exter-nally or market-driven.
Table 7: Key drivers of premature versus delayed launches
Reasons for premature launches Reasons for delayed launches Lock correct shelving at customers to receive right
visibil-ity
No earlier customer support for launch was achieved driv-en by the innovation, that made launching impossible Low priority in-store due to other projects that were
run-ning, later timing been better
International launch date (start of shipment) delayed NL launch
Early versus fixed shelf mutation at customer Late versus fixed shelf mutation at customer Synchronization with big customer promotion Missed seasonality (see other column also) Seasonality influence decides consumption peak (launch
had to be rushed to make December sales -/ summer sales season)
Competition anticipated due to launches abroad that were done earlier, hence missed opportunities in local market
Highly innovative product, consumers were not ready for it (did not understand the product)
Technical issues at plant caused a lack of sufficient pro-duction capacity, hence supply was hampered
Rushing to be pioneer with a typical product Competition already had similar products in market, det-rimental for perceived innovativeness
Switching barrier established by strong innovation from competition
Slow distribution build up
Budget insufficiency, launch postponed to new fiscal year Development of product and positioning strategy that needed to be aligned with the global corporate strategy and top management took to much time
6. CONCLUSION
Launch timing is supported to be a driver of new product performance. This research shows that de-layed launch timing has negative effects on new product performance. The negative effect on new product performance is mainly driven by missed market opportunities. Competition might already has introduced a similar innovation in advance of your new product launch, consumers might have switched to this innovation or established technol-ogies might have created entry barriers for a later entry. These effects are contributing to the fact that a later product introduction is not beneficial to a new product launch in terms of performance. Alt-hough the market risks are limited due to more de-mand certainty, this does not add up to the increas-ing downsides of a late product launch.
In addition, innovativeness plays an amplifying role in this relation, where at an increasing degree of innovativeness the negative effects of delayed launches on new product performance are in-creased. Unfortunately, the current paper remains inconclusive on the hypothesized negative effect that a premature launch has on new product per-formance.
The data shows that most launches that were se-lected for this research on the basis of a physical product change and supported by a media and communication plot, were actually more often de-layed than accelerated. Deviation from the initial launch timing was mainly driven by market-related pressures. These pressures include customer as-sortment changes that happen generally only twice a year for large customers and insufficient custom-er support was gencustom-erated for the innovation. The latter point is interesting from a managerial point of view, especially in the light of the negative moder-ating impact of innovativeness. Customer readiness
is key in launching an innovative new product on to the market. All financials apart, the customer has to support the innovation in other to get it on the shelves. Hence, the more innovative a product (and bigger the impact for new product performance on delays), the more essential it gets to get your cus-tomer on board of your launch plan and align the launch timings with your customer.
Limitations & future research implications
The limitations of this research have implications for future research to direct. Five general issues that were found to be limiting the interpretability of this paper’s results are suggested.
First, a key issue in the generalizability of results in this paper is the common method bias that was found. The use of a single respondent approach, a lack of secondary sources and the non-randomized survey, lacking of reversed questions and dummy questions imply room was given for the common method bias to arise, that lead to measurement er-rors. In this light, a future study should attempt to revalidate the results in this paper in a single indus-try, also directing for a bigger sample size that lev-erages the effects of the common method bias (Calantone & Di Benedetto, 2007). The generaliza-bility of this research is limited due to the single industry approach. However, a single industry ap-proach enables this research paper to draw a clear conclusion on the effect of a delayed launch on new product performance, that would be more problematic in a multi-industry approach due to having to correct data for unobserved industry ef-fects.
