What is the difference between a search report and a claim chart?

A claim chart is one section of a report: a table with each claim limitation in one column and the citation that discloses it in the other. The report is the chart plus everything needed to trust it — the priority date, what was excluded and why, where the references came from, and where the search looked. A chart without a search record cannot tell you what was missed.

What does a pinpoint citation look like?

For a US patent, column and line (col. 4, ll. 12–31) or, for post-2001 publications, paragraph number ([0042]); for a figure, the element number (Fig. 3, element 112). For a journal article, page and column or section. A citation to a reference as a whole, or to its abstract, is not a pinpoint and should not appear in a chart — the reader has to be able to find the disclosure without re-reading the document.

Should a report include references that do not read on the claims?

Yes, briefly. The near misses — references that disclose most of a claim and where exactly they fall short — are what make a §103 combination possible later, and they are what tells a litigator how crowded the field is. They belong in a short section of their own, distinct from the references relied on, with the missing limitation identified.

Why does the report need a search record?

Because a negative result is only worth something if you can see where the search looked. Classes searched, databases covered, queries run and date ranges tell the reader whether a 'nothing found' means the art does not exist or means the search did not reach it. It also lets a second searcher extend the work instead of repeating it.

Can I see a sample report?

Yes — redacted samples of the deliverable format are on the samples page. Ask any searcher for one before commissioning work; the format of the sample is the format you will receive.

Guides

What does a prior-art search report contain?

Michael W. Smith, former USPTO patent examiner · September 13, 2026

A usable prior-art search report contains six things: the priority date the search was run against and how it was established; an exclusion screen listing every reference already before the examiner; the references relied on, each mapped element-by-element to the claims with column-and-line or paragraph citations; the provenance and public-accessibility evidence for each reference, especially non-patent literature; the search record — classifications, databases, queries, and what was ruled out; and a plain conclusion, including 'nothing close' when that is the finding. A ranked list of patent numbers with relevance scores is a hit list, not a report.

A report versus a hit list

Most search products are hit lists: a ranked set of patent numbers, sometimes with a relevance score and a highlighted abstract, sometimes with a summary paragraph per reference. They are quick to produce and they look like work. What they leave to the reader is the whole job — reading each reference against each claim limitation and deciding whether it discloses it. A hit list transfers the search cost from the searcher to the attorney at three times the hourly rate.

A report is the result of that reading. It should be possible to open it, go to the chart for claim 1, and see for each limitation either a citation you can check or a stated reason none was found. The sections below are what that requires.

1. The priority date, and how it was established

Everything in an invalidity search depends on the date the art has to precede, and the date is not always the one on the front page. A patent that claims priority to a provisional gets that date only for claims the provisional actually supports; a continuation-in-part gets its parent's date only for matter in the parent. The report states the date used for each claim, and, where the chain is contestable, says so — because a broken chain moves the date forward and can bring years of additional art into play.

2. The exclusion screen

Every reference cited during prosecution, on the face of the patent and in the file wrapper, listed up front together with any argument the applicant made to overcome it. Two reasons. It tells the reader which references in the report are new to the Office — the only ones that matter for an IPR petition under §325(d). And it records where the examiner's own search went, which is the map for where the new art was found.

3. The references relied on, charted

The core of the report. For each reference relied on, a table: claim limitations down the left, the disclosing passage on the right, cited to the column and line, the paragraph, or the figure element. Where a limitation is met by a combination, the chart says which reference supplies it and the motivation to combine is stated and sourced — from the references or the art, not from the patent under attack. Where a limitation is not met, the chart says so. A chart that reads every limitation onto every reference is a chart nobody read carefully.

The chart uses the claim's own words in the left column, verbatim. Paraphrasing a limitation is how a reference that almost reads on the claim gets charted as if it did.

4. Provenance and accessibility

Each reference comes with what is needed to rely on it: the publication date and its source; for foreign art, the family member in English and whether the relied-upon passage is in the original or a translation; for non-patent literature, the evidence of public accessibility before the priority date — an archive capture, a library catalogue record, a conference programme, a distribution date. Non-patent art is where the strongest invalidity references are, and it is also the art most often excluded for want of exactly this section.

5. Near misses

A short section for references that disclose most of a claim and exactly where they stop. These are the secondary references of a future §103 ground, the evidence of a crowded field, and the honest account of how close the search came. Each entry names the missing limitation, so that a later search for that one element can be run without repeating this one.

6. The search record

Classifications searched, with the reasoning for the class set; databases and collections covered, including foreign-language sources; the queries run and their date bounds; and what was deliberately ruled out. This is the section that makes a negative result worth paying for. "Nothing found" from a search that lists its fields means the art probably does not exist in those fields; "nothing found" with no record means nothing.

7. The conclusion, in plain language

One page: what was found, how strong it is, what the best ground looks like, and what is missing. If the honest answer is that the patent looks sound over the art that could be reached, the report says that. A searcher who never reports a negative cannot be trusted on a positive, and the negative is the finding that saves the client the most money.

What is deliberately absent

Relevance scores. A number from a ranking algorithm says how similar the vocabulary is, not whether the reference discloses the limitation. It is reported when it helps a reader prioritise and never in place of a reading.
A legal opinion. A search report states what the art discloses. Whether a claim is invalid is counsel's conclusion, and a report that draws it is overreaching.
Padding. Twenty references of which three matter is a report that hides its finding. The three go in the chart; the rest, if they are near misses, go in the near-miss section; if they are not, they are left out and the search record says why.