Product mapping patent infringement analysis is the comparison that decides every infringement question, and it is mechanical once you commit to doing it one element at a time.
The discipline is the method. Reading a claim and forming an overall impression produces wrong answers in both directions — people conclude they infringe because the products are similar, or conclude they do not because theirs has extra features.
One row per element removes that entirely. Either the element is present or it is not.
And the credibility comes from evidence, not argument. A chart citing a manual page carries weight; a chart asserting a conclusion does not.
The method
| Step | Output |
|---|---|
| 1. Confirm the patent is in force | Go or stop |
| 2. Select the claims to map | Usually the independent claims |
| 3. Parse the claim into numbered elements | A checklist |
| 4. Gather public evidence about the product | A source pack |
| 5. Map one element per row, with a citation | The chart |
| 6. Grade each element | Present / absent / arguable |
| 7. Check the prosecution history on arguable rows | Estoppel position |
| 8. State what could not be verified | An honest chart |
Step one first, always. Nearly three in five US utility patents lapse before term, and mapping against an expired patent is wasted work — see the patent survival curve.
Parsing the claim
Claims have a structure, and breaking it correctly is most of the work.
Claim 1. A fluid control assembly comprising: a housing defining an inlet and an outlet; a poppet disposed within the housing and biased toward the inlet by a spring; and a bypass channel extending between the inlet and a region downstream of the poppet.
| # | Element |
|---|---|
| 1 | A housing |
| 2 | The housing defines an inlet and an outlet |
| 3 | A poppet within the housing |
| 4 | The poppet is biased toward the inlet by a spring |
| 5 | A bypass channel |
| 6 | Extending between the inlet and a region downstream of the poppet |
Split at every limitation, not every clause. "Biased toward the inlet by a spring" contains two requirements — the direction and the mechanism — and both must be present.
Note the transition word. "Comprising" is open, so extra features in the product are irrelevant. "Consisting of" would be closed and extras could avoid the claim.
Gathering evidence
| Source | Weight | Availability |
|---|---|---|
| User and service manuals | High | Usually public |
| Specification sheets | High | Public |
| Regulatory filings | High | Often public |
| Teardown photographs | High | Purchase and disassemble |
| Published source code | High | Where applicable |
| The company's own patents | High — describes their approach | Public |
| Marketing describing function | Moderate | Public |
| Marketing describing benefits | Low | — |
| Assertion without a source | None | — |
The company's own patent applications are the most underused source. Companies describe their products in their own filings, in technical detail, in public, and those descriptions can be cited directly.
Marketing material describing function is usable; material describing benefits is not. "Uses a spring-loaded poppet" is evidence. "Delivers superior reliability" is not.
Choosing which patent and claim
Mapping everything is wasteful; mapping the wrong thing is worse.
| Selection step | Rule |
|---|---|
| Which patents | Those in force, in the relevant jurisdiction |
| Which claims | Independent claims first |
| Order | Broadest first |
| Dependent claims | Only if the independent is met |
| Expired patents | Skip entirely |
| Pending applications | Map published claims separately, noting they may change |
Broadest first is efficient. If a product falls outside the broadest independent claim, it falls outside everything narrower.
Pending applications need their own treatment. Published claims usually narrow before grant, so a mapping against them is a worst-case view rather than a current one. See patent is pending.
The chart
| Column | Contains |
|---|---|
| Element | The exact claim words for that element |
| Accused feature | The specific corresponding feature |
| Evidence | Source, with a page or figure reference |
| Grade | Present / absent / arguable / unverified |
| Notes | Prosecution history, alternative readings |
Quote the claim language exactly. Paraphrasing in the element column is how mappings drift toward the conclusion the author wanted.
Cite a locatable source. "Service manual p.14, fig. 3" is a citation; "the product manual" is not.
See patent claim chart for the artifact itself and how it is used in negotiation and litigation.
Grading
| Grade | Meaning | What follows |
|---|---|---|
| Present | Clearly within the claim language | Continue |
| Absent | Clearly not there | Defensive analysis can stop |
| Arguable | Similar but not clearly within | Check estoppel; this is where the case lives |
| Unverified | Cannot be observed from public material | Say so; do not guess |
Unverified is a legitimate grade and an underused one. Marking an internal component as unverified is more useful than assuming it either way, and it tells the reader exactly what further evidence would settle the question.
Arguable rows drive everything else. They are where the doctrine of equivalents argument lives and where prosecution history estoppel decides whether it is available.
Worked example: a mapping
Claim 1 above, against a competitor's valve.
| # | Element | Accused feature | Evidence | Grade |
|---|---|---|---|---|
| 1 | A housing | Cast aluminium body | Spec sheet p.2 | Present |
| 2 | Defines inlet and outlet | Ports marked IN and OUT | Spec sheet fig. 1 | Present |
| 3 | Poppet within the housing | Internal poppet shown | Service manual fig. 4 | Present |
| 4 | Biased toward inlet by a spring | Magnetic bias assembly | Service manual p.11 | Arguable |
| 5 | A bypass channel | Secondary passage shown | Teardown photo 3 | Present |
| 6 | Between inlet and downstream region | Runs inlet to post-poppet cavity | Teardown photo 3 | Present |
The arguable row
| Question | Finding |
|---|---|
| Function-way-result | Magnetic bias performs the same function, arguably the same way |
| Prosecution history | Claim originally read "a biased poppet" |
| Amendment | Narrowed to "biased by a spring" to overcome prior art |
| Estoppel | Applies — equivalents likely barred |
| Conclusion | No literal infringement; equivalents estopped |
The file history decided it, not the technical comparison. The applicant gave up non-spring biasing to get the patent, and that scope generally cannot be recaptured.
Which is why step seven is not optional. A mapping that stops at "arguable" without checking prosecution history leaves the most important question unanswered. See doctrine of equivalents.
Mapping method claims
| Difference | Handling |
|---|---|
| Elements are steps, not components | Map each step |
| Who performs each step | Add a column |
| Divided performance | Note it explicitly |
| Evidence | Behaviour, logs, documentation of operation |
The performer column is essential and frequently omitted. A method claim split between a service operator and an end user may escape liability entirely, and the mapping is where that becomes visible.
Map the actor for every step, then ask whether one party directs or controls the others. Anything short of that generally defeats direct infringement.
Mapping for assertion versus defence
| Asserting | Defending | |
|---|---|---|
| Completeness required | Every element | Stop at the first absent element |
| Tone | Establish presence | Establish absence |
| Unverified rows | Weakness — fill them | Acceptable |
| Prosecution history | Check before asserting | Check first |
| Purpose | Support a demand or contention | Assess exposure |
A chart with gaps invites the response you were trying to pre-empt. If you are asserting, an unverified row tells the other side exactly where to push.
Defensively, one clearly absent element is a complete answer. There is no need to map the rest, though doing so is useful if the claim might be construed differently.
Documenting what you could not see
Every mapping has gaps, and stating them is what makes it credible.
| Gap | How to record it |
|---|---|
| Internal component not observable | "Unverified — requires teardown" |
| Software behaviour not documented | "Unverified — requires testing" |
| Manufacturing process not public | "Unverified — process not disclosed" |
| Configuration varies by customer | "Varies — verified in configuration X only" |
An honest gap is more useful than a confident guess. It tells the reader exactly what evidence would resolve the question, which is actionable.
It also protects the analysis. A mapping asserting presence for an unobservable element collapses the moment anyone asks how it was determined.
Common errors
| Error | Consequence |
|---|---|
| Mapping against the abstract | Wrong scope entirely |
| Paraphrasing claim language | Drift toward the desired conclusion |
| Grouping several limitations in one row | Missing a requirement |
| Treating extra features as avoiding the claim | Wrong — "comprising" is open |
| Assuming an unobservable element is present | Unsupportable |
| Skipping the prosecution history | Missing estoppel |
| Uncited assertions | The chart carries no weight |
Mapping against the abstract is the most common and most consequential. The abstract has no legal effect on scope, and it usually describes the invention more broadly than the claims do.
Treating extras as a defence is the second. Almost every claim uses "comprising", so a product containing everything claimed plus more still infringes.
Automation and verification
| AI does well | Still human |
|---|---|
| Drafting a first-pass chart | Verifying every citation |
| Finding candidate documentation | Reading the prosecution history |
| Suggesting element mappings | Grading arguable rows |
| Processing many patents at once | The legal conclusion |
Verification is not optional. A tool can cite a document that does not say what it claims, and a chart built on a misread source is worse than no chart because it looks credible.
Automation makes portfolio-scale mapping affordable, which changes renewal decisions more than it changes litigation. Checking whether anyone practises a patent's claims used to cost more than the maintenance fee. See patent portfolio management.
Product mapping: the checklist
- Confirm the patent is in force before starting.
- Map the independent claims first. Falling outside one excludes everything dependent on it.
- Parse into numbered elements, splitting at every limitation rather than every clause.
- Note the transition word. "Comprising" means extra features never help.
- Quote claim language exactly. Paraphrasing lets conclusions drift.
- Cite a locatable source for every row — page, figure, photograph.
- Grade honestly, and use "unverified" rather than assuming.
- Check the prosecution history on every arguable row. Estoppel usually decides them.
- Add a performer column for method claims, and ask who directs or controls.
- Stop at the first absent element when defending, and fill every row when asserting.