technology‑review.com

Method 09 · Assess

Patent & Publication Analysis

Scientific publications and patents are the earliest robust traces of technological activity — years ahead of products and revenues. Counting them over time, countries, institutions and topics shows where research is clustering, where research turns into property rights (a sign of maturity) and who the actors are. Bibliometrics replaces no substantive assessment, but it objectifies it: a technology with exponentially rising patent filings and a growing share of industrial applicants is in a different phase from one with stagnating university papers.

YEARS Publications Patents Industry overtakes academia → maturity signal for the radar The lag between the curves is itself a statement.
Publications run ahead, patents follow with a lag — and the moment industrial applicants overtake the universities is a maturity signal.

Procedure

  1. Fix the search strategy. Keywords, classifications (IPC/CPC for patents), period, databases. The search strategy determines the result — it must be disclosed.
  2. Pull time series. Publications and patent filings per year; compare growth rates rather than absolute numbers.
  3. Analyse actors. University or industry? Which countries, which companies? The shift from academic to industrial is a maturity signal.
  4. Citation and topic structure. Where do clusters form, which topics travel together?
  5. Combine with other methods. Bibliometric curves alone say little — together with TRL and S-curve, a lot.

Strengths

  • Data-based and reproducible — the search strategy is the evidence.
  • Very early lead, especially for publications.
  • Shows actors and regions that remain invisible in market figures.

Limits

  • Patenting behaviour is industry- and country-specific; software and services are under-represented.
  • Quantity is not quality; citation counts lag by years.
  • Publication pressure creates noise — not every curve means progress.

How we use it in our roadmaps

Bibliometrics supplies the hard time series beneath our assessments. Per bundle we track publication and patent curves of the most important sub-technologies; the transition from university to industrial filing is, for us, a signal to move a technology from "3–10 years" to "1–3 years" in the radar. Conversely, the absence of that transition despite many papers is a warning sign.

In the source apparatus of the roadmaps we state the search strategy wherever an assessment is bibliometrically supported. That makes it verifiable — and that is what distinguishes it from an opinion.

References