White space analysis: finding gaps in a patent field

White space is a part of a patent field where filings are thinner than the activity around it would predict. Finding it means measuring filing density by classification code, technology cluster or combination of concepts, and comparing what you find with what you would expect. Deciding whether a gap is worth filing into means ruling out the ordinary reasons an area is quiet, then testing the gap with a targeted search before anyone starts drafting.

What white space is, and what it is not

White space is relative. A cell with 40 patent families is white space if each of its neighbors holds 400; the same 40 would be crowded in a field of 60. The comparison is the finding, not the raw count.

Two things white space is not:

• An absence of prior art. Papers, product documentation, standards and old expired patents can all sit in an area with few recent filings. A gap in filings is not a gap in disclosure.
• Freedom to operate. Broad claims held by a company in a dense neighboring area can still read on a product placed in the gap. Filing density says nothing about claim scope.

How density is measured

By classification code

Count families per Cooperative Patent Classification (CPC) group and compare sibling groups. The CPC, launched jointly by the EPO and the USPTO and in use since January 2013, gives a shared vocabulary across offices. Its weakness is that codes follow examiner practice: a new technology can sit scattered across older codes before it gets codes of its own.

By cluster

Text-based clustering groups documents by what they describe rather than how they were classified. Gaps show up as small clusters next to large ones, or as empty ground between two clusters that a product would naturally connect. The patent landscape reading guide covers how to check whether a cluster holds together before trusting its size.

By concept matrix

Pick two axes that matter to the business, such as application against technical approach, and count families in each cell. This is the format most people mean by a white space map, and it makes expected density easy to estimate.

Expected against actual

A simple expectation for any cell: row total multiplied by column total, divided by the grand total. It assumes rows and columns are independent, which is crude, but it turns "this cell looks small" into a number. Take a hypothetical battery thermal-management field of 2,000 families:

Charging control with phase-change material would hold about 69 families under the simple model (500 × 275 ÷ 2,000). It holds 5. That is a candidate gap and nothing more yet. There may be a plain technical reason for it: phase-change materials are passive, and charging control is an active function. Better expectations also use the cell's neighbors and the field's growth over time, but the logic is the same.

Why a quiet area can be a dead end

Each of these produces a low count without any opportunity behind it.

1. The field is immature. The components that would make the combination work do not exist yet. Filings may come later, or never.
2. It was tried and abandoned. Engineers attempted it and it failed. Expired patents, abandoned applications and papers reporting negative results are the evidence.
3. The subject matter is hard to patent. Under 35 U.S.C. § 101 and Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014), a claim directed to an abstract idea or another judicial exception is eligible only if it adds an inventive concept that transforms it into patent-eligible subject matter. Areas where most claims would fail that two-step test stay thin because applicants expect rejection, not because nobody works there.
4. Companies keep it secret. Manufacturing process details and tuning parameters that cannot be seen in a finished product are often protected as trade secrets. Activity can be high while filings stay low.
5. Classification artifacts. The filings exist but sit somewhere else: under a broader code, under a code introduced recently, or in vocabulary the query did not use.
6. The filings are not public yet. Applications publish about 18 months after their priority date, so the newest work in any area is invisible for at least that long. Why recent patent filing trends look like a decline explains how to correct for it.
7. No commercial reason. The combination works but nobody would pay for it.

The first five are the reasons to test for. The sixth is a correction to apply. The seventh is a business question, not a search question.

Testing a candidate gap, step by step

1. Write the gap as a short claim-like statement. List the elements a claim covering the gap would need. Vague gaps cannot be tested.
2. Search again with different words. Use synonyms, older terminology and the neighboring classification codes. A gap that disappears under a second vocabulary was a query artifact.
3. Search non-patent literature. Papers, conference proceedings, theses, standards and product documentation. Disclosure outside patents still counts as prior art against a later filing.
4. Check recent and pending filings. Allow for the publication lag, and look at neighboring families with pending continuations. A continuation can still claim into the gap if the parent's disclosure supports it, as described in Patent families: continuations, divisionals and CIPs.
5. Read the dead ends. Find abandoned applications and expired patents in or near the gap and work out why they stopped.
6. Test eligibility. Ask whether the likely claim would be directed to an abstract idea and, if so, what inventive concept it would add.
7. Look for commercial signals. Products, standards work and neighboring filings by companies that sell in the area.
8. Read the dense neighbors' claims. A filing gap is not freedom to use the space.
9. Decide. File, publish defensively, monitor or drop. Record the reason, because the same gap will come up again.

Steps 2 to 5 usually settle most candidates. A gap that survives all nine is rare and worth the drafting cost.

Where White Space Analysis fits

One Transform's White Space Analysis maps filing density across a domain and highlights areas where activity is unexpectedly thin, with context for each: the adjacent clusters, the density you would expect and the filings actually found. That covers the measurement. For the testing steps, a Similar Patents search started from a plain-language description of the gap finds the closest existing documents, and a Freedom to Operate search checks whether in-force claims next door could read on a product placed in it. Every finding links to its sources.

Next step

Take one gap you already suspect in your field, write it as a claim-like statement, and run a White Space Analysis on the surrounding domain to see whether the expected density supports it. Start at onetransform.ai.

This article is general information about patent practice, not legal advice. For advice on a specific matter, consult a registered patent attorney or agent.