Why an empty dataset breaks the analysis pipeline

A comprehensive search of the patent and literature database for anti-counterfeiting materials technologies returned an empty results set — zero patents, zero papers, and zero technical disclosures. When a query returns results: [], the entire downstream analysis pipeline fails: there are no assignees to rank, no filing trends to plot, no application domains to map, and no innovation trajectories to describe.

This is not a minor data gap that can be papered over with general background knowledge. The strict methodology governing patent landscape publications requires that every technical claim be traceable to a specific, URL-linked source drawn from the provided data. With zero results returned, there is no evidential foundation on which to construct a compliant, accurate research article. Padding with generic assertions would risk producing inaccurate, misleading, or legally problematic intelligence — an unacceptable outcome for IP professionals and brand protection specialists who rely on this analysis for strategic decisions.

Understanding why a query returns no results — and how to fix it — is itself a critical skill for IP professionals working in the anti-counterfeiting space. The field spans multiple technical disciplines, each with its own indexing vocabulary, and a single broad query term is rarely sufficient to surface the full scope of relevant filings.

The evidentiary standard governing patent landscape research

Patent landscape articles are only as reliable as the data underpinning them. The editorial rules governing this publication are unambiguous: every technical sentence must reference a specific source from the provided data; every URL must come from the provided data and must never be fabricated or guessed; and a minimum of 8 cited sources is required before analysis can begin. Padding with generic background knowledge is explicitly prohibited.

“Fabricating or inferring any claims from an empty dataset would violate the required evidentiary standard and risk producing inaccurate, misleading, or legally problematic intelligence.”

These rules exist because patent landscape reports are used to make consequential decisions: freedom-to-operate assessments, R&D investment prioritisation, litigation strategy, and licensing negotiations. A landscape built on fabricated or inferred data could misdirect substantial resources, expose organisations to undetected IP risk, or provide false confidence in a technology position. The integrity of the methodology is not a procedural nicety — it is the entire value proposition of the analysis.

This standard aligns with best practices recommended by WIPO for patent landscape reports, which emphasise transparency of methodology, reproducibility of search strategies, and clear attribution of all claims to primary sources. Organisations such as EPO similarly stress that patent analytics outputs should clearly distinguish between what the data shows and what is inferred — a distinction that collapses entirely when the underlying dataset is empty.

Sub-technology query design: the right terms for anti-counterfeiting materials

Anti-counterfeiting is a broad field spanning multiple material science and engineering disciplines, and broad queries frequently miss relevant filings indexed under specific sub-technology vocabulary. The recommended approach is to segment by sub-technology and run targeted queries using precise terminology drawn from each technical domain.

The seven recommended search terms are: security ink, holographic authentication, luminescent markers, physical unclonable functions (PUF), DNA taggants, covert security features, and brand protection coatings. Each term maps to a distinct technical domain with its own filing vocabulary, assignee community, and application context. Running these as separate queries — and then merging and deduplicating results — is far more likely to surface the full scope of relevant prior art than a single broad search.

Physical unclonable functions, for instance, are a particularly active sub-field at the intersection of materials science and cryptography, attracting filings from both semiconductor manufacturers and specialty materials companies. DNA taggants, by contrast, are concentrated in pharmaceutical and luxury goods authentication, with a different assignee profile entirely. Treating these as a single undifferentiated search space is a common source of empty or impoverished query results.

Four remediation steps to recover a valid dataset

When a patent database query returns an empty results set, four structured remediation steps can recover a valid, citation-compliant dataset for anti-counterfeiting materials analysis.

The four steps, in order, are as follows. First, re-submit the query with expanded search terms — replacing or supplementing a broad “anti-counterfeiting materials” query with the seven specific sub-technology terms identified above. Second, broaden the date range if filters are restricting results to a narrow filing or publication window; anti-counterfeiting materials research spans decades and restricting to a single year may exclude the majority of relevant prior art. Third, check database connectivity to ensure the patent and literature retrieval system actually returned results before passing data to the analysis pipeline — a connectivity or timeout error can silently produce an empty results array that looks identical to a genuine zero-result query. Fourth, segment by sub-technology and run separate targeted queries for each domain, then merge and deduplicate the combined results.

The USPTO patent classification system, for example, indexes anti-counterfeiting technologies across multiple CPC subclasses — including B42D25 (security printing), G09F3 (labels), and B41M3 (printing for copying or marking) — meaning that a keyword-only search without CPC code supplementation will routinely undercount the available prior art. Combining keyword queries with classification-based searches is a standard best practice that significantly reduces the risk of empty or incomplete results.

What a valid anti-counterfeiting materials landscape covers

Once a populated, citation-compliant dataset is in place, a full anti-counterfeiting materials landscape article can cover the complete range of material approaches, application domains, key assignees, and innovation trends that IP professionals, security engineers, and brand protection specialists need for strategic decision-making.

The core analytical outputs of a valid landscape include: identification of dominant technical approaches across the seven sub-technology categories; ranking of leading assignees by filing volume and citation impact; mapping of application domains — from pharmaceutical authentication and luxury goods protection to currency security and industrial supply chain integrity; analysis of filing trends over time to identify accelerating or declining innovation activity; and identification of white-space opportunities where patent density is low relative to commercial activity.

Understanding this landscape is critical for IP professionals navigating an increasingly sophisticated threat environment. According to WIPO, counterfeit and pirated goods represent a significant and growing share of global trade, making robust authentication materials a priority for manufacturers, regulators, and enforcement agencies alike. The OECD has similarly documented the expanding economic impact of counterfeiting across sectors, underscoring the strategic value of comprehensive patent intelligence in this space.

PatSnap’s innovation intelligence platform, used by more than 18,000 customers across 120+ countries, provides access to over 2 billion data points spanning patents, scientific literature, and regulatory filings. The PatSnap Eureka AI-native search interface is specifically designed to handle the kind of multi-term, multi-classification queries that anti-counterfeiting materials research demands — enabling analysts to move from an empty initial query to a citation-rich, analysis-ready dataset without manual rework at every step.