Citation impact, blockbuster patents, self-citation dynamics, and quality metrics across leading patent holders
Ch. 10Organizational Patent Quality
While Organizational Patent Count examined the volume of corporate patenting, raw patent counts tell only part of the story. Two firms with identical output levels can differ substantially in the quality and influence of their inventions. This chapter consolidates the quality dimensions of organizational patenting: from blockbuster rates and citation impact to self-citation behavior and the temporal durability of a firm's knowledge contributions.
The analysis employs four complementary approaches: a blockbuster-versus-dud typology that classifies firms by the tails of their quality distributions; average and median forward citations to capture central tendency and skewness; self-citation rates that distinguish cumulative internal R&D from broad external knowledge sourcing; and citation half-lives that measure how quickly a firm's patents accumulate their influence.
Quality Typology
Plotting each firm's The share of an entity's patents that fall in the top 1% of 5-year forward citations within their grant-year × CPC section cohort. Rates above 1% indicate disproportionate high-impact output. against its The share of an entity's patents that receive zero forward citations within five years of grant. Higher dud rates indicate a larger proportion of patents with no measurable downstream impact. for the most recent complete decade (2010–2019) reveals distinct innovation strategies. Firms in the upper-right produce both high-impact breakthroughs and many zero-citation patents — a high-variance strategy. Firms in the lower-right achieve high blockbuster rates with few duds, the hallmark of consistent quality.
Figure 1
Amazon's 6.7% Blockbuster Rate Leads the Field, While 18 of 50 Firms Exceed a 50% Dud Rate (2010–2019)
Blockbuster rate versus dud rate for the top 50 assignees (2010–2019), with bubble size proportional to patent count and color by primary CPC section
Loading visualization…
Each bubble represents one of the top 50 assignees in the decade 2010–2019. X-axis: share of patents in the top 1% of their year x CPC section cohort. Y-axis: share of patents receiving zero 5-year forward citations. Bubble size: total patents. Color: primary CPC section.
Amazon occupies the lower-right quadrant with a blockbuster rate of 6.7% and a dud rate of 18.3%, classifying it as a consistent high-impact innovator. By contrast, several firms — predominantly those with large electronics and component portfolios — cluster in the upper-left with blockbuster rates below 0.2% and dud rates above 50%.
Select a company:
Figure 2
IBM: Forward Citation Distribution Over Time
5-year forward citation percentiles (P10-P90) for IBM patents by grant year, with company selector
Loading visualization…
5-year forward citation percentiles for IBM patents by grant year (1976–2019). Bands show P25-P75 (dark) and P10-P90 (light). Solid line = median; dashed gray = system-wide median. Only years with >=10 patents shown.
The width of the fan reveals the dispersion of quality within the firm's portfolio. A widening gap between the median and upper percentiles indicates increasing reliance on a small fraction of high-impact patents.
Figure 3
IBM: Blockbuster Rate (Top 1% Patents) and Dud Rate (Zero Citations) Over Time
Annual share of top-1% blockbuster patents and zero-citation dud patents for IBM, with company selector
Loading visualization…
Annual blockbuster rate (patents in top 1% of year x CPC section cohort, blue) and dud rate (zero 5-year forward citations, red) for IBM. Dashed line at 1% marks the expected blockbuster rate under uniform quality.
Diverging blockbuster and dud rate trajectories over time reveal shifts in a firm's innovation strategy, distinguishing periods of breakthrough-oriented R&D from phases of incremental or defensive patenting.
Forward Citations
The number of times a patent is cited by later patents. A widely used proxy for patent impact and technological importance., which measure how frequently a firm's patents are cited by subsequent inventions, provide a complementary indicator of the impact and influence of organizational innovation.
Figure 4
Microsoft Leads in Average Citations (30.7) While IBM's 5.6x Average-to-Median Ratio Reveals a Highly Skewed Portfolio
Average and median forward citations per patent for major assignees, based on patents granted through 2020
Loading visualization…
Average and median forward citations per patent for major patent holders, limited to patents granted through 2020 to allow for citation accumulation. Organizations with large average-to-median ratios exhibit skewed citation distributions indicative of a small number of highly cited patents.
The divergence between average and median citations distinguishes portfolios characterized by a few highly cited patents (high average, lower median) from those with more uniformly impactful output.
Self-Citation Patterns
Self-citation patterns reveal meaningful differences in how firms accumulate knowledge. In patent-dense fields such as semiconductors and electronics, elevated self-citation rates may reflect genuine cumulative innovation, with each patent building upon the firm's previous work.
Figure 5
Canon (47.6%), TSMC (38.4%), and Micron (25.3%) Exhibit the Highest Self-Citation Rates Among Top Assignees
Self-citation rate (fraction of backward citations to same assignee) for the 20 most-cited assignees, revealing knowledge recycling patterns.
Loading visualization…
The figure displays the fraction of all backward citations that are self-citations (citing the same assignee's earlier patents), for the 20 most-cited assignees. Canon (47.6%), TSMC (38.4%), and Micron (25.3%) exhibit the highest self-citation rates, reflecting deep cumulative R&D programs in imaging and semiconductor technologies.
Elevated self-citation rates among firms with cumulative R&D programs are consistent with long-term knowledge building on internal prior art, though strategic considerations may also contribute.
Citation Half-Lives
The rate at which a firm's patents accumulate citations varies considerably. The The time it takes for a patent or portfolio to accumulate half of its total forward citations. Shorter half-lives indicate more immediately impactful inventions; longer half-lives suggest foundational work. — the time required to accumulate 50% of total citations — distinguishes firms whose patents have immediate versus foundational impact.
Figure 6
Citation Half-Lives Range From 6.3 Years (Huawei) to 14.3 Years (US Air Force)
Years until a firm's patents accumulate 50% of total forward citations, using patents 15+ years old to ensure complete citation windows.
Loading visualization…
Years until a firm's patents accumulate 50% of their total forward citations. Only patents 15 or more years old are included to ensure a complete citation window. Pharmaceutical and chemical firms demonstrate longer half-lives, whereas electronics and IT firms exhibit shorter ones.
Pharmaceutical and chemical firms tend to exhibit longer citation half-lives, reflecting the gradual accumulation of citations in science-intensive fields. Electronics and IT firms demonstrate shorter half-lives, with citations peaking shortly after grant.
Quality Metrics Across Leading Organizations
Beyond citation impact and blockbuster rates, patent quality can be assessed through claims complexity, technological scope, and examination timelines across leading organizations.
Figure 7
IBM and Samsung Led in Average Forward Citations Through the 1990s, but Convergence Narrowed the Gap by 2015
Average forward citations per patent per year for the top 5 assignees by total patent count
Loading visualization…
Average 5-year forward citations per patent for the five most prolific assignees, plotted annually from 1976 to the most recent complete citation window. Declining averages across all firms reflect system-wide citation dilution as the patent population expands.
The long-run convergence in forward citations suggests that scale alone does not sustain citation advantage — as portfolios grow, average impact per patent tends to regress toward the system-wide mean.
Figure 8
Claim Counts Rose Across All Top Filers, With Samsung and Canon Converging on IBM by the 2010s
Average number of claims per patent per year for the top 5 assignees by total patent count
Loading visualization…
Average number of claims per patent for the five most prolific assignees, plotted annually. Rising claim counts reflect a system-wide trend toward broader patent drafting, with notable firm-level variation.
The universal rise in claim counts indicates broader patent drafting strategies across all leading firms, likely associated with changes in prosecution practice and competitive pressure to secure wider protection.
Figure 9
Patent Originality Scores Rose Sharply After 1980 and Stabilized Near 0.5 for Most Leading Firms
Average originality score (diversity of backward citation sources) per year for the top 5 assignees
Loading visualization…
Average originality score for the five most prolific assignees, plotted annually. Originality measures the diversity of technology classes cited by each patent's backward references, with higher values indicating broader knowledge sourcing.
The convergence of originality scores near 0.5 across diverse firms suggests a common shift toward drawing on broader prior art bases, regardless of a firm's primary technology domain.
Figure 10
Grant Lag Doubled From ~600 Days to Over 1,200 Days Before Declining After USPTO Reforms
Average days from filing to grant per year for the top 5 assignees by total patent count
Loading visualization…
Average grant lag (days from application filing to patent grant) for the five most prolific assignees, plotted annually. The rise and subsequent decline reflect USPTO examination backlog dynamics and procedural reforms.
The parallel trajectories across firms indicate that grant lag is associated primarily with USPTO capacity and policy rather than firm-specific prosecution strategies, though firms filing in certain technology-dense classifications historically experienced slightly longer prosecution times.
Non-Patent Literature and Foreign Citation Patterns
Two additional dimensions of patent quality illuminate the knowledge base underlying corporate innovation: non-patent literature (NPL) citations, which capture links to scientific research, and foreign prior art share, which reveals the degree to which a firm's patents draw on international knowledge sources.
Figure 11
Apple Leads NPL Citation Intensity at 35.9 Average NPL Citations Per Patent
Top 20 firms ranked by average non-patent literature citations per patent, revealing science intensity of corporate R&D
Loading visualization…
Firms ranked by average non-patent literature (NPL) citations per patent. Apple (35.9), Microsoft Technology Licensing (21.5), and Semiconductor Energy Laboratory (21.2) lead, reflecting deep connections between corporate R&D and the scientific literature.
High NPL citation intensity distinguishes firms whose R&D is closely connected to the scientific frontier from those pursuing more incremental, application-driven innovation.
Figure 12
Huawei's Prior Art Is 78.3% Foreign — Nearly 7x IBM's 11.4% Foreign Share
Top 20 firms ranked by share of backward citations to foreign (non-US) patents, revealing reliance on international knowledge sources
Loading visualization…
Firms ranked by the share of their backward patent citations directed to foreign (non-US) patents. Huawei (78.3%), Toyota (63.2%), and LG Electronics (61.0%) lead, consistent with their non-US headquarter locations and reliance on international prior art.
The foreign citation share strongly correlates with the assignee's headquarter location, but also reveals the degree to which firms draw on a global versus domestic knowledge base.
The quality analysis reveals that large patent portfolios mask substantial heterogeneity in impact. Blockbuster rates, forward citation distributions, self-citation patterns, and citation half-lives each expose a different dimension of organizational innovation strategy. The next chapter, Patent Portfolio, examines how firms distribute their innovation across technology domains and whether portfolio diversification strategies correlate with sustained leadership.
Data coverage: January 1976 through September 2025. All 2025 figures reflect partial-year data.