# Patent Sustainability Map: World Patent Body Class Map Flags 33% Overlap—Report or Not?

Priya Menon · October 12, 2026

> WIPO's IPC-to-SDG patent mapping flags a 33% overlap. Learn how this concordance works and verify totals before publishing sustainability claims.

This guide explains how WIPO's IPC-to-SDG mapping methodology classifies patent filings and what the reported one-third SDG overlap means for portfolio reporting decisions.

It provides a verify-before-you-commit check: compare like-for-like totals and terms, then confirm the live, complete mapping before any sustainability claim is published.

![This guide explains how WIPO's IPC-to-SDG mapping methodology — Patent Sustainability Map](https://static.mm-ais.com/article-images-ai/patent-sustainability-map-world-patent-b-ai-0c3507b4.jpg)

## How It Works

WIPO's IPC-to-SDG mapping is a concordance: a published lookup table that pairs patent classification symbols with Sustainable Development Goal targets. The mechanism runs from the goals toward the codes, not the reverse. The SDG framework supplies the destination list — 17 goals, each broken into numbered targets with a short statement of the subject matter it covers. Analysts then work through the IPC's symbol definitions and ask, symbol by symbol, whether the technology a symbol is defined to cover corresponds to the technology a target describes. Where the answer is yes, the pair is recorded. The output is a static table; classifying an individual filing is a separate step that happens afterward.

That sequence is why the mapping is code-based rather than text-based. Nothing in the mechanism reads a filing's title, abstract, or claims. A filing acquires an SDG tag only because a patent office assigned it an IPC symbol the concordance lists. Two consequences follow. The tag reflects a classification decision, and classification practice varies across offices and examiners. The tag is also inherited, so a portfolio's SDG profile is only as durable as the symbols sitting in it.

Five terms carry the rest of the analysis, and each one is a place where two reports can silently diverge:

| Term | What it means in practice |
| --- | --- |
| Concordance | The published symbol-to-target lookup; the artifact you are actually citing. |
| IPC symbol | A classification code nested from section down to subclass and group. The mapping resolves at one chosen depth, and that depth sets how coarse the tags are. |
| SDG target | A numbered sub-item inside one of the goals. Mapping to targets is what allows a code to roll up to a goal. |
| Multi-tag | One symbol may list several targets, and one patent family may carry several symbols. The relationship is many-to-many, not one-to-one. |
| Counting basis | Whether the unit of count is a publication, a family, or a symbol — and whether a multi-tagged item is counted once per tag or once overall. |

The counting basis is where totals move. Ask whether a family tagged to several goals is counted once per goal or once in the portfolio denominator, and whether the same family appearing as multiple publications is deduplicated. To verify, pick one goal, count distinct families tagged to it, then count symbol-level hits for the same goal. If the two totals differ by more than the multi-tag rate you can explain, the bases are mismatched and the comparison is not like-for-like.

Finally, only a fraction of IPC symbols reach any goal at all — the one-third overlap this guide uses as its working figure. That remainder is a scope boundary, not a finding about any invention's impact, because the concordance tests technical subject matter while the goals describe policy outcomes. Before you commit a portfolio number to a report, confirm which symbol depth the mapping used, which counting basis produced the total, and whether the number describes symbols or filings.

![How It Works — Patent Sustainability Map](https://static.mm-ais.com/article-images-ai/patent-sustainability-map-world-patent-b-ai-6f916301.jpg)

## Key Factors to Consider

Before you commit a portfolio number to a sustainability report, apply one rule: verify the live, complete option, then compare like-for-like totals and terms. Three criteria decide whether an IPC-to-SDG figure survives that test — coverage of your symbol set, denominator discipline, and mapping vintage. Run all three before the number leaves your draft.

| Decision criterion | The check to run | What it protects |
| --- | --- | --- |
| Coverage of your symbol set | List every distinct IPC symbol on the filings in scope and mark each as mapped or unmapped in the named concordance edition | Keeps an unmapped filing from being read as an SDG-free filing |
| Denominator discipline | Fix office, filing-year window, family definition, and unit of count before dividing anything | Makes two totals actually comparable |
| Mapping vintage and granularity | Record the concordance version and extraction date; test whether target-level rows sum cleanly to goal totals | Makes the reported figure reproducible |

**Criterion 1: coverage, not library size.** The useful question is not how many SDG targets the concordance lists, but how many of your filings sit on symbols the edition actually maps. Pull the distinct symbols in scope, mark each mapped or unmapped, and publish the unmapped share beside the SDG total. An unmapped symbol signals a gap in the concordance, not evidence that the filing has no SDG relevance.

**Criterion 2: like-for-like denominators.** Two SDG totals are comparable only when the office, the filing-year window, the family definition (priority-based versus publication-based), and the unit of count match. One filing can carry several IPC symbols and therefore land on several SDG targets, so decide once whether you are counting filings or symbol–target pairs. Label the unit on every table; if you shift between the two, the totals stop being comparable.

**Criterion 3: vintage and granularity.** The concordance is a published table tied to an IPC edition, and results can be rolled up at goal or target level. Name the concordance version and the extraction date on the report. Then check whether your target-level rows sum cleanly to your goal totals — if a filing appears under two targets, the goal figure is not a sum of unique filings.

**The numbers that matter.** The roughly one-third SDG overlap is the figure that drives the reporting decision: the concordance accounts for about a third of filings, leaving the remainder outside it. A report that shows the SDG total without the uncovered remainder reads as complete when it is partial. Request three figures from one extract — total filings in scope, filings carrying at least one mapped symbol, and the mapped/unmapped split by symbol. If those three cannot come from the same extract, you are not yet comparing like with like. Treat the one-third figure as a screening flag, not a reportable SDG patent share, and hold any portfolio claim until the like-for-like comparison is complete.

![Key Factors to Consider — Patent Sustainability Map](https://static.mm-ais.com/article-images-pixabay/patent-sustainability-map-world-patent-b-30d76262.jpg)

## Common Mistakes

Most bad SDG figures in patent portfolio reports are not mapping errors. They are counting errors, and they survive review precisely because the mapping was applied correctly, which makes the final total look credible. The pitfalls below are the ones that quietly inflate or distort a number before anyone checks it against the underlying records.

The first is summing symbol rows instead of distinct filings. Consider a battery thermal-management application whose symbols both resolve to the same goal: a spreadsheet that counts rows reports that invention twice, and the error compounds as the portfolio grows. The check is mechanical. Deduplicate on the publication number before you total, then verify that the deduplicated count is never larger than the raw row count. If it is, your join key is wrong.

A close cousin is counting family members rather than inventions. A single filing extended across several jurisdictions can appear as several SDG filings, inflating the numerator and the denominator together and flattering every ratio built on them. Run the same query twice, once at the document level and once at the family level, and compare the outputs. The gap is your jurisdictional duplication — a figure you can only disclose if you actually measured it.

The second pitfall is comparing figures produced under different versions of the reference tables. Classification schemes and mapping releases get updated, so a growth claim built by placing last year's number beside this year's number may be comparing two different rulebooks. The check: re-run the prior period through the same release you used for the current period, or, if that is not possible, label each figure with its release and drop the growth claim. Reproducibility is the test, not plausibility.

Finally, a missing match is not a finding. Portfolios weighted toward general-purpose instrumentation often contain large blocks of codes outside the mapped set, and reporting those as zero sustainability activity asserts something the lookup cannot support. Sample the unmatched block, record how many entries a technical reviewer would reasonably credit, and disclose an undercount rather than a hard zero. A corrected number is defensible in review; an inflated or mismatched one is not, and the difference is usually one verification pass made before the figure went into the report.

![Common Mistakes — Patent Sustainability Map](https://static.mm-ais.com/article-images-pixabay/patent-sustainability-map-world-patent-b-e217479a.jpg)

## Insider Tactics

The most useful insider move is to stop treating the concordance as a search tool and start treating it as a filter you run in reverse. Most teams pull every SDG target that looks relevant, then collect the IPC symbols attached to each one, and report the union. That inflates the count, because a single patent family can sit under several targets at once. Instead, take your portfolio's actual IPC symbols first, look each one up in the live concordance, and record only the targets that the table itself assigns. The overlap you are left with is the defensible figure — and it is usually the smaller one. Confirm the live, complete mapping version before you commit, and compare like-for-like totals and terms.

Timing matters more than most analysts admit. The concordance is a living document: WIPO revises it as the SDG framework and the IPC evolve, so a figure you pulled last year may not reconcile with the same portfolio today. Before you commit a number to a report, re-pull the mapping and re-run the count in the same session. If a prior-period figure is already published, recompute it under the current table and disclose the restatement rather than silently carrying the old total forward. Like-for-like means same table version, same portfolio snapshot, same counting rule. Re-run the comparison whenever the mapping version changes, re-checking the same IPC classes, SDG mapping terms, and filing window.

Run one reconciliation check before you sign off. Take your total filings, subtract the filings whose IPC symbols appear nowhere in the concordance, and confirm the remainder equals the sum of your per-target counts after de-duplication. If the two sides disagree, you have double-counted a family, not mis-mapped a code. The arithmetic is simple; the discipline is not.

| Check | What to verify | Pass condition |
| --- | --- | --- |
| Table version | Live concordance date matches the date cited in the report | Same version for every figure |
| Direction | Symbols looked up from the portfolio, not targets collected first | No union-of-targets counting |
| De-duplication | Each patent family counted once per target | Sum of targets equals de-duplicated total |
| Restatement | Prior-period figures recomputed under the current table | Any change disclosed, not hidden |

Two habits keep the number honest across reporting cycles. First, freeze the portfolio snapshot and the table version together, so the pair can be reproduced on demand. Second, keep a short audit note for each figure: which symbols were included, which were excluded, and why. When a reviewer asks why your SDG total differs from a peer's, the note answers in seconds — and the answer is almost always the counting rule, not the mapping. Do not commit a sustainability portfolio claim on the one-third overlap alone; hold the claim until the like-for-like comparison is complete.

None of this requires new data. It requires running the check before the commitment, not after.

![Insider Tactics — Patent Sustainability Map](https://static.mm-ais.com/article-images-pixabay/patent-sustainability-map-world-patent-b-ca7be774.jpg)

## Comparison

Put the two reporting options side by side and the arithmetic does the arguing. Option A is the headline SDG share: total filings in the portfolio, divided by filings whose IPC symbols appear in WIPO's published IPC-to-SDG concordance. Option B is the verified subset: only filings whose symbols map to a specific SDG target, counted once per target, with the concordance version and retrieval date recorded. Run both on the same portfolio and the gap is the story. Work the two options on the same portfolio and the gap is the story. Take a portfolio of 1,200 filings in which 400 carry at least one mapped symbol: Option A reports 400 divided by 1,200, or 33.3 percent. Now suppose 60 of those 400 filings map to more than one target and you count each mapping: Option B's numerator becomes 460 and the share rises to 38.3 percent. Same portfolio, two defensible-looking numbers, a five-point spread created entirely by counting rules. Before you commit either figure to a report, reconcile the denominators first — same filing set, same concordance version, same retrieval date — then reconcile the numerators. If the denominators differ, stop; the comparison is void.

That spread is the one-third overlap in miniature. WIPO's concordance runs from goals toward codes, so a single IPC symbol can land under several SDG targets; the overlap is structural, not an error. The practical consequence: any total that counts mappings rather than filings will exceed the total that counts filings, and the two cannot be compared.

| Check | Option A: headline share | Option B: verified subset |
| --- | --- | --- |
| Unit counted | Filings with any mapped symbol | Filings per mapped SDG target |
| Overlap handling | Collapsed | Counted per target |
| Version control | Often unrecorded | Concordance version and date logged |
| Best used for | Internal trend lines | External sustainability reporting |

Each option wins in a different room. Option A wins for internal trend lines: it is fast, it tolerates a stable concordance version, and it answers "is our mapped share moving?" without claiming target-level precision. Option B wins for external reporting, assurance review, or any disclosure where a reviewer will ask which targets and which concordance version produced the number. The rule: if the audience will test the figure, use Option B; if the audience only tracks direction, Option A is enough — provided you label it as a mapped-symbol share, not an SDG contribution.

One more comparison decides the winner in practice: reproducibility. Hand Option B to a colleague with the concordance version, retrieval date, and counting rule, and they should regenerate your total exactly. Hand over Option A without those three items and they cannot. When two options produce different totals, the one a second person can rebuild is the one that survives review — commit to that one.

## What to do next

| Step | Action | Why it matters |
| --- | --- | --- |
| 1 | Open the live, complete IPC-to-SDG mapping version from the World Patent Body (WIPO) and confirm it is the current release, not a partial class map. | A partial class map invalidates the comparison, so the live, complete option must be verified before any commitment. |
| 2 | Check whether the 33% IPC-to-SDG class overlap you are looking at is the screening flag or a claimed SDG patent share. | The one-third overlap is a classification signal, not a reportable SDG patent share, and must not be committed as a portfolio claim on its own. |
| 3 | Compare like-for-like totals: same IPC classes, same SDG mapping terms, same filing window across the datasets you intend to report. | Unmatched scope makes the patent count non-comparable, so totals only hold when the terms and windows align. |
| 4 | Report SDG totals only against the live, complete IPC-to-SDG mapping version you verified in step 1. | Reporting against anything else breaks the canonical rule of verifying the live, complete option before committing. |
| 5 | Do not commit a sustainability portfolio claim on the one-third overlap alone; hold the claim until the like-for-like comparison is complete. | The 33% figure is a screening flag, and the thesis treats it as a signal requiring verification before portfolio reporting. |
| 6 | Re-run the comparison whenever the IPC-to-SDG mapping version changes, re-checking the same IPC classes, SDG mapping terms, and filing window. | Only a live, complete mapping supports a defensible SDG total; a stale or partial version silently invalidates the comparison. |

## Frequently Asked Questions

**In which direction does WIPO's IPC-to-SDG concordance actually operate?**

The mechanism runs from the goals toward the codes, not the reverse.

**How is the SDG framework itself structured for the purpose of the mapping?**

The SDG framework supplies 17 goals, each broken into numbered targets with a short statement of the subject matter it covers.

**At what stage does an individual patent filing receive its SDG classification?**

Classifying an individual filing is a separate step that happens after the concordance produces its static lookup table.

**Does the concordance read any part of the patent filing's actual text?**

Nothing in the mechanism reads a filing's title, abstract, or claims.

**Under what exact condition does a filing acquire an SDG tag?**

A filing acquires an SDG tag only because a patent office assigned it an IPC symbol the concordance lists.

**Why is the mapping characterized as code-based rather than text-based?**

Because a filing receives a tag only through the IPC symbol a patent office assigned it, not through anything read from the filing itself.

## Quick answers

| What is WIPO's IPC-to-SDG mapping? | WIPO's IPC-to-SDG mapping is a concordance: a published lookup table that pairs patent classification symbols with Sustainable Development Goal targets. |
| --- | --- |
| Which direction does the mechanism run? | The mechanism runs from the goals toward the codes, not the reverse. |
| What does the SDG framework supply? | The SDG framework supplies the destination list — 17 goals, each broken into numbered targets with a short statement of the subject matter it covers. |
| Why is the mapping code-based rather than text-based? | Nothing in the mechanism reads a filing's title, abstract, or claims. |
| Why does a filing acquire an SDG tag? | A filing acquires an SDG tag only because a patent office assigned it an IPC symbol the concordance lists. |

Also worth reading: **WIPO GREEN Register: 2,100 Self-Declared Connections, Not Vetted**: [WIPO GREEN Register: 2,100 Self-Declared](https://iprs.cloud/blog/wipo-green-register-2100-self-declared-connections-not-vetted.php) · **API Triangulation and Clearance Velocity Data for IP Teams**: [API Triangulation and Clearance Velocity](https://iprs.cloud/blog/api-triangulation-and-clearance-velocity-data-for-ip-teams.php)

### Related reading

- [Weight loss drug patents: 2024-2025 longevity indications hit 37 — watch or challenge?](https://iprs.cloud/blog/weight-loss-drug-patents-2024-2025-longevity-indications-hit-37-watch-or-challenge.php)
- [AI Copyright Claims 2026: Model-Only Fails Human-Authorship Test](https://iprs.cloud/blog/ai-copyright-claims-2026-model-only-fails-human-authorship-test.php)
- [Meta monthly active users: Community Rights End at 700,000,001](https://iprs.cloud/blog/meta-monthly-active-users-community-rights-end-at-700000001.php)
- [Can Prompts Qualify for Copyright?: Section 202.1 in 2026 — Length or Human Expression?](https://iprs.cloud/blog/can-prompts-qualify-for-copyright-section-2021-in-2026-length-or-human-expression.php)
- [Trade secret disclosure rules: $407M verdict on binder vs clean team](https://iprs.cloud/blog/trade-secret-disclosure-rules-407m-verdict-on-binder-vs-clean-team.php)
- [Patent Maintenance Fees: 30% Lapse, Auto-Pay vs Triage vs Prune](https://iprs.cloud/blog/patent-maintenance-fees-30-lapse-auto-pay-vs-triage-vs-prune.php)

### Latest

- [Weight loss drug patents: 2024-2025 longevity indications hit 37 — watch or...](https://iprs.cloud/blog/weight-loss-drug-patents-2024-2025-longevity-indications-hit-37-watch-or-challenge.php)
- [AI Copyright Claims 2026: Model-Only Fails Human-Authorship Test](https://iprs.cloud/blog/ai-copyright-claims-2026-model-only-fails-human-authorship-test.php)
- [Meta monthly active users: Community Rights End at 700,000,001](https://iprs.cloud/blog/meta-monthly-active-users-community-rights-end-at-700000001.php)

Canonical: https://iprs.cloud/blog/patent-sustainability-map-world-patent-body-class-map-flags-33-overlapreport-or-not.php
Markdown: https://iprs.cloud/blog/patent-sustainability-map-world-patent-body-class-map-flags-33-overlapreport-or-not.php/index.md