On this page
A success-factor list can become misleading when managers treat its coefficients as permanent laws.
The short answer is that new-product success factors need a transfer boundary. An estimate from a published study is evidence about a relationship in a research record. It is not automatically a stable coefficient for every product, market, organization, or time period.
Evanschitzky, Eisend, Calantone, and Jiang update earlier evidence on new-product success with a meta-analysis of 233 empirical studies from 204 manuscripts. Their update covers research published from 1999 through 2011 and reports generally weaker effect sizes than the earlier meta-analysis. The study also evaluates contextual and methodological moderators. The practical lesson is not that success factors are useless. It is that a factor must travel with its context, measurement, and uncertainty.
Henard and Szymanski (2001) provide the earlier meta-analytic baseline and likewise organize success across product, strategy, process, market, and organizational dimensions. Atuahene-Gima (2005) adds a mechanism boundary inside product innovation, where exploiting existing competence can support incremental innovation while exploring competence can support radical innovation. The update is therefore read against both an earlier synthesis and a concrete capability trade-off, not as an isolated factor table.
Why is an isolated new product success factor insufficient for commercial resource commitment?
Suppose a research synthesis associates a product characteristic with success. Before using the result, four objects need to stay distinct:
| Object | Question |
|---|---|
| Factor | What product, process, market, or organizational characteristic is being studied? |
| Effect size | How large is the estimated relationship in the synthesis? |
| Moderator | Under which context or method does the relationship vary? |
| Decision | What action, if any, is justified in the current product system? |
Table 1Why is an isolated new product success factor insufficient for commercial resource commitment?
Source: Table from this essay. Sources and interpretation are given in the article.
Collapsing these objects produces familiar but fragile sentences: “This factor drives success,” “The coefficient is proven,” or “We should invest in the factor immediately.” A meta-analysis can improve the estimate while also showing why the estimate should not be detached from the studies that produced it.
| Synthetic evidence card | Earlier synthesis | Updated evidence |
|---|---|---|
| Evidence base | Earlier published studies | 233 studies from 204 manuscripts |
| Research window | Earlier period | 1999-2011 update period |
| Overall reading | Factor list appears portable | Generally weaker effect sizes |
| Boundary question | Which factor is associated with success? | Which context and method moderate the estimate? |
| Decision test | Copy the factor into a plan | Recheck measurement, setting, and expected effect |
Table 1The updated-success-factor transfer card
The comparison is qualitative and synthetic. It does not reproduce the source's effect-size values or rank individual factors.
Source: Author's synthetic framework; source claims are Evanschitzky et al. (2012), Henard and Szymanski (2001), and Atuahene-Gima (2005).
The card makes a methodological point visible. “Updated” does not mean “true forever.” It means that the evidence base, period, model, or method has changed and the relationship should be re-read.
How do empirical meta-analytic updates diminish previously published innovation drivers?
An effect can become weaker for several reasons. The new studies may measure the construct differently. Markets may have learned the practice, reducing its differentiating value. Samples may expand to settings where the relationship is less pronounced. A more suitable meta-analytic model may change the estimate. Publication and method patterns can also shape the evidence that enters the synthesis.
This article does not assign one of these explanations to every factor. Henard and Szymanski (2001) provide the earlier synthesis against which the updated evidence is read. The source reports generally weaker effect sizes than that earlier meta-analysis and evaluates contextual and methodological moderators. That combination is enough to change the managerial question. Instead of asking which factor is strongest in the abstract, ask whether the relationship is credible under the current product system and whether the cost of acting is justified by the expected decision value.
Why must product development success claims always specify market and regulatory context?
Context is not an appendix to the estimate. It can include product category, market conditions, country culture, organizational setting, development stage, performance definition, measurement procedure, sample composition, and publication period. A relationship can be positive in one boundary and weak, absent, or reversed in another.
The updated study’s attention to contextual and methodological moderators creates a useful discipline:
| Before transfer | Required check |
|---|---|
| Construct | Does the current team mean the same thing by the factor and by success? |
| Product setting | Is the product category and development stage comparable? |
| Market setting | Are competition, customer expectations, and adoption conditions comparable? |
| Method | Does the evidence use a design that can answer the current decision? |
| Time | Could the factor’s value have changed as the practice became common? |
| Cost | What would acting on the estimate require, and what is the counterfactual? |
Table 3Why must product development success claims always specify market and regulatory context?
Source: Table from this essay. Sources and interpretation are given in the article.
The checks do not invalidate the meta-analyses. Atuahene-Gima’s (2005) capability-rigidity finding is a useful reminder that the same product system can face different innovation demands depending on whether the decision is incremental or radical. The checks prevent a synthesis from being used as if it were a guarantee.
How does competitive factor diffusion erode historical product advantages?
One interpretation of weaker effects is diffusion. If many firms learn that a practice is associated with new-product success, the practice may become easier to imitate and less differentiating. That is an interpretive possibility, not a universal explanation established for every factor in the source.
The possibility changes how a team should use evidence. A factor can still be necessary, valuable, or worth improving even if it no longer differentiates firms strongly. Operational hygiene and competitive advantage are different decision categories. A weaker comparative effect does not mean that quality, customer fit, or process discipline no longer matter.
The team should therefore label the intended role of an intervention:
- remove a failure condition;
- meet a customer or regulatory requirement;
- improve a process that is below an internal standard;
- create a differentiating advantage;
- test a new hypothesis under uncertainty.
The same factor may support the first three goals without reliably delivering the fourth.
How should product leadership structure a factor transfer protocol for active pipelines?
When using an external success-factor finding, write a one-page transfer record:
- copy the source’s construct and outcome wording;
- record the evidence base and period;
- identify the moderator or boundary most likely to matter;
- state what is comparable in the current product system;
- state what is not comparable;
- define the smallest reversible test;
- choose an outcome and time horizon before the test starts;
- update the belief after the test rather than treating the source estimate as a target.
This protocol keeps the evidence useful without pretending that a meta-analysis is a roadmap for one product launch.
Which three universal product success claims must innovation teams reject?
First, do not say that an updated meta-analysis proves a factor has no value because its average effect is weaker. Average evidence and decision value answer different questions.
Second, do not say that a factor is a universal driver of new-product success. The source explicitly examines contextual and methodological moderators.
Third, do not transfer a coefficient without transferring the construct, outcome, sample, measurement, and time boundary that make the coefficient interpretable.
For adjacent decisions, compare the multi-method product forecast with the preference boundary in conjoint analysis.
Where are the empirical boundaries of new product success factor research?
Evanschitzky et al. provide the updated evidence base, the 1999-2011 research window, the generally weaker effect sizes relative to the earlier synthesis, and the role of contextual and methodological moderators. Henard and Szymanski provide the earlier meta-analytic baseline, while Atuahene-Gima provides a capability and innovation-type boundary. The transfer card and protocol are author-owned translations. They do not rank factors, assess a current innovation program, or establish a causal effect for a particular product.
References
- Evanschitzky, H., Eisend, M., Calantone, R. J., & Jiang, Y. (2012). Success factors of product innovation: An updated meta-analysis. Journal of Product Innovation Management, 29(3), 447-457. DOI
- Henard, D. H., & Szymanski, D. M. (2001). Why some new products are more successful than others. Journal of Marketing Research, 38(3), 362-375. DOI
- Atuahene-Gima, K. (2005). Resolving the capability-rigidity paradox in new product innovation. Journal of Marketing, 69(4), 61-83. DOI
