Abstract: What truly drives innovation performance—and does it differ by innovation outcome? Analyzing 43 years of U.S. patent data, our study, recently published in the Journal of Management Studies, finds business units are the strongest influence across all innovation metrics, from patent output to financial value. Star inventors drive invention creation and internal value capture, while corporations matter more for turning innovation into financial gains. Drivers vary across performance types—creation vs. appropriation—offering critical insights for firms looking to tailor strategies that unlock and retain innovation value. 

How can we better understand what drives innovation? 

Strategic management seeks to understand what drives performance differences among firms. Scholars have long used different approaches to identify these drivers. One method, variance decomposition, quantifies how much specific categories—such as country, industry, firm, or business unit—contribute to performance differences. Originally applied to financial outcomes, variance decomposition has more recently been extended to study strategy-related outcomes like mergers and acquisitions. We advance this line of inquiry by applying a novel variance decomposition approach to examine innovation performance—an increasingly critical dimension in the knowledge economy. 

We analyze 43 years of U.S. patent data, from 1980 to 2022, and employ a refined method that corrects for serial correlation bias. We measure innovation using four distinct performance indicators: (1) innovation output (patent counts), (2) innovation social value (forward citations), (3) value appropriation (self-citations), and (4) financial value (stock market reaction to patent grants). 

Key insights 

Our findings yield several important insights: 

• Business unit effects dominate innovation performance across all four metrics, explaining an average of 32.6% of the variance. These effects are nearly six times greater than corporate-level effects (5.5%) and over 17 times greater than technology area effects (1.9%). Our results extend prior innovation decomposition studies that focused only on firm-level effects and underscore the critical importance of fostering innovation environments within immediate organizational units. 

• Star inventors matter significantly, contributing an average of 11.5% of the variance overall. Their impact is strongest in patent generation (17.1%) and self-citations (13.7%), suggesting they are more focused on developing interconnected inventions within their own research programs than on maximizing external recognition or financial returns. This points to an opportunity: firms could align top inventors with areas of strategic value to improve both value creation and appropriation. 

• Corporate-level effects vary meaningfully by outcome type. They are modest for patent counts and total citations (average of 1.0% and 3.0%) but considerably larger for self-citations (5.5%) and financial value (12.4%). This pattern suggests that corporations play a more central role in converting innovation into financial returns and firm-specific advantage than in simply producing innovation. 

• Technology area effects are small but consistent, accounting for about 2% of the variance. While not dominant, this effect signals that the underlying technological domain still influences patenting behavior and should remain part of strategic considerations. 

Why are these findings important?  

 We contribute to the understanding of what drives innovation performance in three key ways: 

1. We quantify and compare the impact of different innovation drivers. 
While many studies examine how specific factors affect innovation, few directly compare the relative contributions of firm-level, business-unit, technology-area, and individual effects. We show that business unit and star inventor effects far outweigh others in explaining variance, offering fresh support for the resource-based view and the growing literature on “star performers.” 

2. We differentiate the drivers of value creation vs. value appropriation. 
Our results highlight that different drivers matter for different types of innovation outcomes. Business unit activities primarily drive value creation (e.g., patent production and citations), while corporate-level structures play a more prominent role in value appropriation (e.g., self-citations and stock market reactions). This distinction offers actionable insights for managers deciding where and how to invest in innovation strategy. 

3. We introduce a methodological advancement to variance decomposition research. 
Our study is the first in the management literature to apply variance decomposition while correcting for autocorrelation using the Anderson and Hsiao (1982) approach. This refinement leads to more accurate estimates and helps future researchers explore new domains, such as individual contributions or corporate social performance, using cleaner metrics. 

Recommendations 

These findings carry important implications for theory and practice: 

• For managers, our research emphasizes the importance of investing in the inventive capacity of business units and strategically aligning top inventors with organizational goals. Capturing the value of innovation requires not just invention, but the structures to appropriate it. 

• For scholars, our work calls for more detailed modeling of the link between innovation creation and capture, and for greater focus on individual-level contributions. There’s significant promise in extending variance decomposition to other fields where individual actors may shape collective outcomes (e.g., sports, education, or healthcare). 

Finally, while our work covers a large share of U.S. patented inventions (25%), we caution against generalizing our results to firms with low levels of innovation activity. Future research might also explore the effects of country-level institutions, CEO-level strategic emphasis, or invention type (radical vs. incremental) on performance. 

In sum, our study identifies where the real drivers of innovation reside. The results consistently point to the business unit—and the inventors inside it—as the heart of innovation performance. That’s where firms must look if they want to build sustainable innovation advantages.