Summary

  • ITU Document 33, dated 5 August 2026, proposes one starting indicator for Target 1.7, “Interoperability standards impact”: the percentage share of global video traffic carried by ITU-standardized codecs.
  • The paper says a large majority of Internet video uses ITU-standardized compression and that a share can change with competition from other codecs. It does not publish the proposed indicator’s baseline, numerator, denominator, traffic unit, observation method, data source, uncertainty, owner or 2031 target.
  • Council records had already called for specific, measurable and relevant indicators linked to financial and operational plans, based on existing tools, focused on results achieved through ITU activity and avoiding proxy indicators.
  • Codec prevalence is not automatically institutional attribution. ITU’s own histories identify H.264, H.265 and H.266 as products of joint ITU-T and ISO/IEC teams, while deployment also depends on platforms, devices, licensing, content libraries and network choices.
  • PP-26 can preserve the reach measure while requiring a public metric passport and two distinct ledgers: one for ecosystem prevalence and another for ITU-attributable contribution.

The new indicator replaces two unfinished candidates

Document 33 is not an adopted strategic target. It is a Secretary-General’s report submitted for consideration by the 2026 Plenipotentiary Conference, alongside the Council’s draft in Document 32. The conference, scheduled in Doha from 9 to 27 November, is the body expected to agree ITU’s strategic and financial roadmap for 2028–2031.

The paper responds to an explicit Council instruction. During the May session, the General Secretariat and relevant Bureaux were asked to improve Targets 1.5, 1.6 and 1.7 before PP-26. Target 1.7 retained the label “Interoperability standards impact,” but the proposed way to observe it changed.

Document 32 Addendum 1 had placed two bracketed candidates beneath the target. One counted ITU-T Recommendations included in the accreditation scope of testing laboratories recognized by ITU. The other counted ITU-T standards adopted in products or services, as reported by ITU membership. The brackets mattered: these were unresolved candidates, not settled performance measures.

Document 33 removes both from the Target 1.7 proposal and offers one starting point instead: “Share of global video traffic carried by ITU-standardized codecs (%).” Its explanation is concise. A large majority of Internet video, it says, runs on ITU-standardized compression. Expressing that use as a share would let the indicator move over time, register competition from other codecs and show that much of the world’s video depends on standards ITU helps set.

That is a clearer public story than two administrative counts. A percentage appears comparable across years and closer to real-world use. But a statistic can become rhetorically clear before it becomes operationally measurable. The replacement solves the presentation problem and leaves the evidence problem open.

The public paper does not define the percentage

“Share of global video traffic” has no self-evident denominator. It could mean bytes transported, viewing minutes, streams initiated, sessions, encoded assets, device playbacks or traffic observed at selected networks. High-resolution video would dominate a byte measure; short clips could dominate a session count. Cached delivery might appear at one observation point and disappear at another. Adaptive streaming can make one viewing session use several representations. Transcoding can place different codecs on the same content’s path.

Document 33 does not choose among these units. It does not identify which networks, content-delivery systems, platforms, countries or observation windows enter the estimate. It does not say whether encrypted traffic is classified directly, inferred from metadata or excluded. It gives no treatment for private networks, multicast, broadcast-adjacent delivery, downloaded files, edge caches or traffic that cannot be classified.

The classification side is equally important. “ITU-standardized codecs” needs a versioned registry. Does a codec qualify whenever an ITU-T Recommendation exists? How are profiles, levels, extensions and technically aligned twin texts handled? If a stream can be decoded under more than one label, which rule prevents double counting? If a platform changes containers but not compression, does the classification change? The public proposal answers none of these questions.

Nor does it publish the ordinary governance fields of a performance measure: a baseline value and date; a 2031 target or benchmark interpretation; an uncertainty interval; an update cadence; a responsible data owner; an independent review path; a correction policy; collection cost; or the resource line that sustains measurement. A text search of the published paper finds no baseline, numerator, denominator, data source, sampling rule or uncertainty statement for Target 1.7.

This is a bounded absence claim. ITU may possess internal data, a provider relationship or a developing methodology that is not in the paper. A Member State may submit a fuller method before PP-26. The point is narrower: the document asking the conference to consider the indicator does not yet give delegates or readers the contract needed to reproduce it.

Reach and institutional impact are different ledgers

Even a perfectly measured traffic share would answer only the first question: how prevalent is a defined class of codecs in a defined traffic universe? Target 1.7 asks a second question: what does that prevalence demonstrate about ITU’s impact?

The distinction is visible in the standards’ provenance. ITU’s official history says the Joint Video Team brought together ITU-T’s Video Coding Experts Group and ISO/IEC’s MPEG and produced H.264 | ISO/IEC 14496-10. Later joint work produced H.265 | ISO/IEC 23008-2, and the current JVET covers H.264, H.265 and H.266 | ISO/IEC 23090-3. ITU’s contribution is real. It is also expressly collaborative.

Deployment has additional principals. Device vendors implement decoders and encoders. Platforms select formats and migration schedules. Patent and licensing conditions influence adoption. Content owners retain large encoded libraries. Browser, operating-system and silicon support shape what can be delivered economically. Network operators and content-delivery providers determine where traffic is visible. A large installed base in 2028 may reflect standardization work completed years earlier as much as activity performed during the 2028–2031 plan.

None of that makes the reach statistic useless. It makes its label decisive. “Ecosystem reach of jointly standardized video codecs” is a defensible descriptive object. “Impact achieved by ITU during the plan period” requires another link: the named ITU action, output or support mechanism that contributed to a measured change, plus an account of other causes.

Council records had already drawn this line. At the fifth plenary meeting, one councillor said impact measurement should focus on results achieved through ITU activities rather than Member State performance, should use existing tools and should avoid proxy indicators. The four-country contribution by Bulgaria, France, Italy and Sweden likewise asked for indicators that reflect ITU as an administrative organization and for a chain from goals to priorities, outcomes, indicators, finance and operations.

The video share risks becoming a proxy if prevalence is presented as institutional causation without that chain. It may be large even when current ITU activity is unchanged; it may fall because an efficient competitor succeeds even while ITU’s own standardization process performs well. The same number can therefore describe market position accurately and institutional performance badly.

The metric must also have a resource owner

The Council’s test was not only statistical. It was managerial. Document C26/83 said proposed indicators should use data already available within ITU or the UN system and should not require additional human or financial resources to obtain, assess or apply. Council summaries called for coordination among the strategic, financial and operational plans. Councillors later requested itemized amounts for each Bureau and the General Secretariat, while the CWG-SFP chair said the resource-allocation table would be updated after the draft financial plan was approved.

A global traffic measure is not free because its output is one percentage. It may require commercial telemetry, agreements with platforms or network operators, recurring sampling, codec inference, data cleaning, geographic weighting, confidentiality controls and methodological review. If an existing dataset supplies these functions, the plan should name it and state its limits. If ITU will build or purchase the capability, the relevant owner and cost should appear in the financial and operational chain.

Without that link, the indicator can fail in two ways. It can become ceremonial—a percentage announced once, without a stable series. Or it can become dependent on one opaque supplier whose coverage changes while the headline number appears continuous. Neither outcome satisfies the Council’s call for usable decision data.

A public metric passport can make the proposal decision-ready

The fix does not require a large new bureaucracy. It requires a small, versioned record published before the first performance claim.

The metric passport should define the numerator and denominator, the unit of traffic, included delivery modes and the observation window. It should publish the codec-family and version registry, including joint ITU-T and ISO/IEC provenance, and explain how profiles, transcoding, multiple representations and unknown traffic are treated.

It should then name data sources and their coverage: provider types, networks, platforms, regions, sampling method and exclusions. Classification, inference, weighting, deduplication and missing-data rules belong in the same record. The first estimate should carry a baseline date and uncertainty interval, not only a point value.

Governance fields matter just as much. The passport should identify the accountable owner, reviewer, update cadence, source-contract dependencies, collection cost and resource linkage. Revisions should preserve prior values, method changes and comparability breaks. If the data supplier changes, the public record should say whether the series can still be compared.

Finally, reporting should keep two linked ledgers. The ecosystem-reach ledger records the measured share and its uncertainty. The institutional-contribution ledger records ITU actions: convening a joint team, approving or maintaining specifications, publishing conformance material, supporting implementation, resolving defects or documenting adoption pathways. A link between the two may support an impact claim; one ledger should never silently substitute for the other.

This is where Heng Lu’s distinction between evidence and authorship applies. A record is trustworthy when it describes a bounded reality and exposes the authority behind its state. It becomes unsafe when the label creates the very status it claims merely to observe. Calling a traffic percentage “impact” should not manufacture institutional causation. The measurement should show the observed change, who contributed, what remains uncertain and who accepted the method.

What the evidence does not establish

Nothing reviewed here refutes ITU’s statement that a large majority of Internet video uses compression standardized with ITU participation. This Article produces no competing market-share estimate. It does not recommend one codec, challenge the technical quality of H.264, H.265 or H.266, or minimize ITU’s role in their development.

It also does not treat the earlier bracketed indicators as better. Laboratory accreditation scope can count recognition without deployment. Member-reported product adoption can be incomplete, inconsistent and vulnerable to self-reporting bias. Replacing them may be sensible. The question is whether the replacement is fully specified.

Document 33 is a proposal, and its own “starting point” language leaves room for refinement. TSAG’s standards-success-stories work may later provide additional evidence. PP-26 can amend, condition, combine, replace or reject the indicator. The present decision is therefore not whether video traffic is important. It is whether a strategic performance measure can be adopted before the public knows exactly what will be measured and what kind of impact the result is allowed to prove.

Sources

  1. ITU — Document 33: Strategic and Financial Plans 2028–2031
  2. ITU — Document 32: CWG-SFP report and draft Strategic Plan
  3. ITU Council — C26/83, four-country contribution on the draft Strategic Plan
  4. ITU Council — C26/122, fifth Plenary summary record
  5. ITU Council — C26/124, seventh Plenary summary record
  6. ITU-T — Joint Video Team
  7. ITU-T — Joint Video Experts Team
  8. ITU-T — TSAG liaison record on Standards Success Stories
  9. ITU — What you need to know about PP-26
  10. Heng Lu — On When the Bookkeeper Auditions for Olympus