Summary
- The five regional Internet registries publish engagement spending in categories that are too different for a clean league table. RIPE NCC budgeted EUR 9.6 million for External Engagement and Community in 2024; APNIC recorded AUD 5.06 million for its Engagement pillar in 2024; ARIN budgeted USD 1.875 million for travel and meetings in 2024; and LACNIC's 2023 accounts separated staff, board and commission travel, fellow travel, outreach and training. These figures describe scope, not comparative efficiency.
- Meetings, training, fellowships, travel and public-policy work can all support registry reliability. They can help members use services correctly, surface operational evidence, widen access, reduce policy error and protect the open Internet from poorly designed regulation. The problem begins when participation volume is presented as an outcome rather than an input.
- Each engagement activity should have a named beneficiary population, a denominator, a baseline, an intended change, a cost, an evidence method and a stop or redesign rule. A meeting should be tested against unique network coverage and decision quality; training against retained competence and operational application; travel against the decision or relationship that required physical presence.
- Representation must not be inferred from attendance. Funded fellows, repeat entities, staff, consultants and well-resourced operators may all contribute useful knowledge, but none acquires authority to speak for absent networks merely by being present. Engagement reporting should separate access, evidence, deliberation, authorization and institutional promotion.
- A common member-facing statement should report engagement cost by function, cost per completed intervention, the distribution of beneficiaries, repeated versus first-time participation, documented service or policy changes, unsuccessful activities and follow-up results. It should preserve privacy and avoid reducing every public good to immediate revenue.
- The right reform is not an arbitrary cut to meetings or training. It is outcome-accountable engagement: keep activities that improve competence, service, evidence or legitimate participation; redesign those that only reproduce the same room; and stop claiming that a busy institution has proved a broad mandate.
A large category can conceal several different public goods
Community engagement sounds like one activity because institutions need a short heading for a budget. In practice, the heading can contain very different things. A registry may organize its own member meeting, support an independent network operator group, send staff to a standards meeting, train engineers to use resource-public-key tools, fund a fellow from a small island economy, maintain multilingual communications, meet a ministry drafting Internet legislation, survey customers, support a policy mailing list or sponsor an academic researcher.
The case for each item is different. A technical course may reduce configuration mistakes. A member meeting may let fee payers challenge the budget. A policy forum may expose an allocation rule to operational evidence before adoption. A government meeting may prevent a law from accidentally treating number registration as content regulation. A fellowship may remove a travel barrier for someone who would otherwise be absent. A regional event may reveal service problems that a headquarters team would not see from tickets alone.
These are plausible benefits, not automatic results. The training course can be delivered to people who never use the knowledge. The member meeting can be dominated by the same employers. The policy forum can collect comments after the decisive option has already been selected. The government meeting can become general institutional promotion. The fellowship can improve access while being cited later as proof that a population was represented. The regional event can generate photographs, registrations and travel claims without changing a service or decision.
A useful audit therefore begins by refusing two shortcuts. The first shortcut says engagement is peripheral and should be cut because it is not the registry database. That is too simple. Reliable registries depend on competent users, accurate feedback, understood rules and political conditions in which technical coordination remains possible. The second shortcut says engagement is inherently valuable because community is part of the mission. That is also too simple. A mission statement cannot certify the marginal value of the next trip, conference day or communications campaign.
The budget line should be disaggregated by the public good being purchased. The test is not whether an activity looks communal. It is whether the institution can identify what changed, for whom, at what cost and with what remaining uncertainty.
The published numbers do not yet form a league table
RIPE NCC's 2024 Activity Plan and Budget assigned EUR 9.6 million to External Engagement and Community, about one quarter of its EUR 38.2 million total expense budget. Within that division, EUR 5.75 million was assigned to Community Building and Engagement, EUR 1.8 million to Community Learning and Development, and EUR 2.05 million to Community Coordination and Collaboration. Those headings covered staff as well as operating expenditure. They did not represent cash handed to meeting attendees.
The RIPE NCC 2026 Activity Plan and Budget refined the presentation. Community Building and Member Engagement carried EUR 5.95 million across community building and membership engagement; Learning and Development carried EUR 1.9 million; Coordination and Collaboration carried EUR 1.95 million. The document also named major cost types. Community Building included EUR 1.565 million for RIPE meetings and events, EUR 430,000 for consultancy and EUR 220,000 for travel. Public Policy and Internet Governance included EUR 530,000 for consultancy and EUR 250,000 for travel. Technical Training and Development listed outreach, travel and information-technology costs.
APNIC uses another architecture. Its December 2024 financial report to the Executive Council recorded AUD 5.055 million of actual expenditure in the Engagement pillar against an AUD 5.309 million budget. Community Participation accounted for AUD 2.208 million, Stakeholder Cooperation AUD 1.071 million, Online Communication AUD 960,000 and Relationship Coordination AUD 816,000. APNIC Academy sat in the separate Development pillar at AUD 4.563 million, along with technical and security community support and APNIC Labs.
ARIN's 2024 budget presented narrower expense accounts. It budgeted USD 1.875 million for travel and meetings, split between USD 1.03 million of travel and USD 845,000 of meetings. It also budgeted USD 410,000 for community outreach, including USD 30,000 for On-the-Road and other training, and USD 645,000 for industry support. The total operating expense budget, including depreciation adjustments, was USD 31.707 million.
LACNIC's audited 2023 financial statements reported USD 844,337 for travel by staff, the Board and commissions; USD 314,130 for fellows and exhibitors' travel; USD 1.040 million for outreach; and USD 178,895 for training. Its 2025 budget later planned USD 1.484 million for travel, USD 1.301 million for outreach, USD 235,000 for training and USD 117,000 for community projects within USD 11.666 million of operating expenses.
AFRINIC's institutional crisis makes current comparison harder. Historical reports and activity records remain relevant, but they do not supply a stable, recent and directly comparable engagement series across ordinary governance years. That absence should be reported as a limitation, not filled with a stale number and a confident conclusion.
The figures above should not be converted into one ranking. RIPE NCC includes communications, policy coordination and substantial staff in a broad division. APNIC places Academy expenditure outside Engagement. ARIN publishes account-level travel and outreach costs but not the same pillar model. LACNIC separates categories in its income statement and budget. The currencies, years, labour markets, geographic distances, event models and accounting boundaries differ.
The comparison establishes something more useful: engagement is material, and the institutions already possess enough cost information to connect money to purpose. What is missing is a common outcome account.
An event count is an input, not a result
Annual reports often record how many events were held, economies visited, sessions delivered, entities registered or people trained. These measures are useful. They prove that an activity occurred and can reveal changes in volume. They are not, alone, evidence that the activity achieved its stated public purpose.
Consider a regional meeting attended by 500 people. The number can rise because the host city is accessible, registration is free, sponsorship is generous, a neighbouring conference is co-located or the same large employers send multiple staff. It can fall because visas are difficult, airfares rise, a conflict begins or remote participation improves. Neither movement, without more evidence, tells members whether a policy decision included a broader range of operational facts.
The first missing measure is unique organizational coverage. How many distinct networks, resource holders, universities, public networks, exchange points and civil-society organizations participated? How many were first-time organizations? How many economies and subregions were represented? How much of the attendance came from registry staff, suppliers, sponsors, fellows, speakers and repeat professional entities? These categories may overlap, so the report must explain its counting rules rather than produce a false total.
The second missing measure is the relevant denominator. Five hundred people can look large until compared with twenty thousand members, tens of thousands of autonomous systems or a service region spanning dozens of economies. The denominator does not convert an open policy forum into a referendum. It prevents the institution from describing the room as the region.
The third measure is decision connection. Which agenda items required member authority, operational evidence, general learning or merely institutional updates? What changed after discussion? Were objections answered with reasons? Did a budget, service plan, policy text or implementation date change? If nothing changed, was that because the original plan survived scrutiny or because the session had no decision path?
The fourth is retention. Did first-time entities return independently, contribute to a technical group, use registry services more accurately, or bring knowledge to a local operator community? Return is not the only success measure; a entity may obtain what was needed in one event. It is still useful when interpreted with care.
An event can succeed without producing a vote. It may resolve confusion, expose a risk, build trust among operators who later coordinate an incident, or make a remote member willing to challenge a future proposal. Qualitative evidence belongs in the account. It should be gathered systematically through structured follow-up, decision records and samples rather than selected testimonials alone.
Training needs a competence measure and a use measure
Technical training has one of the strongest claims on engagement money. A registry can reduce operational risk by helping networks maintain accurate contacts, create valid route-origin authorizations, secure accounts, manage reverse DNS, understand transfer procedures and use public registry data correctly. A region with uneven access to commercial training can gain substantial value from a trusted technical curriculum.
Completion counts remain inadequate. A person can attend every session and retain little. An online learner can pass a simple quiz without changing a production system. A course can be popular because certification is useful for a career while producing no measurable improvement in registry-facing operations. None of these possibilities proves the programme failed. They show why delivery is only the beginning of the evidence chain.
Each course should identify a competence objective. For RPKI, the learner might need to explain route validity states, create a safe authorization, avoid an overbroad maximum-length setting and recover from an error. For registry account security, the objective might be multi-factor authentication, role separation, credential recovery and timely contact maintenance. For IPv6, the objective might be a deployable address plan rather than general awareness.
Assessment should occur at more than one point. A baseline indicates what the entity knew before the course. A completion assessment tests immediate understanding. A later check tests retention. A voluntary operational follow-up asks whether the entity applied the skill and what prevented use. The institution should not demand sensitive network configurations as proof. Aggregated, privacy-protective evidence is enough to distinguish learning from attendance.
Cost should be reported per completed learner and per learner demonstrating retained competence, with regional and delivery-mode distributions. An expensive in-person course may still be efficient if it reaches operators who cannot use the online alternative and produces durable skill. A cheap webinar may be poor value if almost everyone leaves early. The objective is not to prefer one format in advance. It is to understand the trade.
APNIC's fellowship model illustrates a useful condition. Its 2025 fellowship programme required substantial online participation before travel eligibility and offered courses, mentoring and conference participation. That design links travel support to preparation. The next evidence question is whether fellows retained and used the competence, not whether funded access made them representatives of their economies or generations.
Fellowship removes a barrier; it does not transfer a mandate
Fellowships can correct a real access failure. Travel from a remote economy can cost more than a small network can justify. Visa processes can be uncertain. Early-career engineers may lack employer support. Women, caregivers, people with disabilities and entities outside established professional circles can face barriers that an open registration page does not remove. Funding, mentoring and preparation can bring knowledge into a room that would otherwise be absent.
The public value claim should be stated precisely. A fellowship funds access and development. It may broaden the evidence available to the institution and build regional capability. It does not make a selection panel capable of appointing representatives for everyone who shares a country, age, gender, language or sector with the fellow.
The outcome account should therefore avoid two distortions. It should not treat later volunteer service, election or employment as the only success; that would reward programmes for feeding the institution rather than strengthening the wider network community. It should not treat the fellow's presence as proof that a constituency consented to an institutional decision.
Better measures include completion, competence, local knowledge-sharing, continued technical contribution, organizational diversity, repeat dependence on funding and the proportion of fellows whose employers are operating networks. Alumni surveys should ask whether the programme improved work, opened a professional path or enabled a local activity. They should also report attrition and negative experience. Privacy and safety may require aggregation, especially for small cohorts.
Selection criteria matter. If applications reward prior visibility, polished English, endorsements from known insiders and promises of future institutional participation, the programme can reproduce the network it says it is widening. Reporting should show applicant and award distributions without exposing unsuccessful applicants. Rotation should prevent a small group from converting recurring support into a durable participation advantage.
Travel requires an attributable purpose
Travel is easy to attack because every ticket is visible and the benefit is often diffuse. It is also easy to defend with general language about relationships. Neither approach is adequate. Some conversations cannot be replicated well online, especially where trust, confidential risk discussion, local language, field observation or access to decision-makers matters. Other trips persist because they are habitual.
ARIN provides a useful transparency mechanism through its quarterly travel reports. Since 2018, it has published ARIN-paid travel costs to significant Internet events so the community can understand the costs and value of outreach to other regions. The existence of a report does not prove the trip was worthwhile, but it creates the basis for review.
Every material trip should have one primary purpose before approval: service delivery, technical coordination, policy evidence, government engagement, member governance, training, incident support or organizational administration. Secondary purposes can be recorded. The traveller should identify why remote participation would be limited public evidence, the intended counterpart or audience, and the expected follow-up.
Afterward, the institution should record outcomes at an appropriate level. A government meeting may be confidential, but the report can state the policy subject, jurisdiction, institutional position, next action and whether a proposed measure changed. A standards meeting can identify the document or operational issue addressed. A network event can identify member problems collected, service changes opened or local partnerships established. A training trip can use the competence measures already described.
Not every trip will succeed. A ministry may ignore the evidence. A planned partner may withdraw. A session may be poorly attended. Reporting failure is important because it lets the institution improve selection. If only successes appear, staff learn to describe every trip as relationship building and the Board loses the denominator needed to judge the portfolio.
The audit should also detect travel concentration. How many trips are taken by executives, public-policy staff, trainers, Board members, community leaders and technical specialists? Which destinations recur? Are the same external events attended by several staff without distinct roles? Are trips combined, as APNIC's 2024 report says it did to save travel cost? Concentration is not automatically waste. It identifies where justification should be strongest.
Public-policy engagement should report the risk, not claim the world was saved
RIRs operate in political environments even though Internet number coordination is technical. Sanctions, cybersecurity law, data protection, digital-sovereignty proposals, telecommunications regulation and content-control debates can all affect registry services. An institution that never speaks with public authorities may leave technical dependencies unexplained until a harmful rule is adopted.
The outcome is difficult to measure because policy change has many causes. A registry should not claim that one meeting prevented a bad law. It can report a contribution. It can identify the draft provision, the technical risk explained, the written submission, the coalition or counterpart, the change between drafts and the remaining issue. Where no public text exists, it can describe the topic and stage without disclosing confidential officials or advice.
This is more rigorous than counting meetings with governments. Fifty meetings can reflect one unresolved issue or fifty new relationships. One carefully timed technical memorandum can have more value than a year of conference panels. The unit should be the policy risk and the evidence supplied, not the number of badges collected.
The institution must also separate defence of the registry function from defence of institutional expansion. Explaining why authoritative records must remain available is different from arguing that one organization should receive a permanent broader role. Public-policy staff may legitimately defend an open, interoperable Internet. They should not use the public importance of uniqueness coordination to place every organizational preference beyond member review.
RIPE NCC's 2026 plan says its public-policy and Internet-governance activity will hold roundtables, engage governments and support regional forums. Those are commitments to activity. The year-end account should connect them to defined risks, written outputs, affected service regions and observed changes. The same discipline applies across the other registries.
Communications are infrastructure only when they improve action
Websites, translations, mailing lists, podcasts, reports and social media are often included in engagement. They can reduce information asymmetry. A member who understands a charging proposal, security change or election deadline can act. A public archive can make governance review possible after the meeting ends. Translation can widen practical access to material that is formally public but inaccessible to much of the service region.
Traffic statistics are weak on their own. Page views can rise because a deadline is confusing or a controversy is growing. Email opens can be distorted by privacy controls. Podcast downloads do not prove comprehension. The right measures depend on the communication's purpose.
For a fee notice, the result might be fewer avoidable billing tickets, fewer missed deadlines and correct use of relief options. For an election notice, it might be delivery coverage, registration completion and a lower rate of credential failure. For a security notice, it might be completed protective action. For a translated policy document, it might be use by affected organizations, substantive comments and fewer interpretation errors.
Communications should also be tested for distribution. A global total can hide concentration in English-speaking, highly connected audiences. The report should show language, economy and organization-type coverage where lawful and technically reliable. It should state when the institution cannot know who read or acted, rather than turning imprecise analytics into certainty.
The same entity can appear in several success stories
Engagement reports often count activities separately. One person can attend a meeting, complete a course, receive a fellowship, join a policy list, speak on a panel and answer a survey. Each programme may then cite that person as a beneficiary. The institution delivered several services, but it did not reach six people.
A privacy-preserving unique-entity measure would reveal concentration without building an intrusive profile. The institution can use consented identifiers or aggregated matching to report how many beneficiaries were first-time, recurring and heavily recurring. It can also count organizations independently of people.
The concentration question is not an accusation against repeat entities. Institutional memory is valuable. A skilled volunteer may contribute for years. A network may reasonably send different staff to different technical sessions. The question is whether the budget keeps lowering barriers for new operational evidence or primarily subsidizes an established circuit.
Employer concentration matters as much as individual concentration. Ten attendees from one multinational operator do not create the same coverage as ten distinct small networks. Three consultants with multiple affiliations should not be counted as three independent constituencies. Reports should preserve names where events are public but analyze representation with role and organizational context.
A cost-to-outcome map can remain honest about uncertainty
The common objection to outcome budgeting is that community value cannot be reduced to a spreadsheet. The objection is partly right. Trust, knowledge, professional relationships and political understanding are not fully captured by a number. A registry that funds only easily counted activities may neglect slow public goods.
That is a reason for mixed evidence, not for no evidence. Each programme can combine quantitative indicators, structured qualitative findings and explicit limits. A meeting report can include unique organizational coverage, decision changes and an independent sample of entity accounts. A training report can combine assessment results with examples of operational use. A policy report can trace submissions and text changes while acknowledging multiple causes.
The map should contain seven fields.
First is purpose: the service, capability, decision or risk the activity addresses. Second is beneficiary: the population intended to gain, not only the people who attended. Third is denominator: the relevant total against which coverage is understood. Fourth is cost: staff time, travel, venue, supplier, grant and overhead using a stated allocation method. Fifth is result: the observable change. Sixth is uncertainty: what cannot be attributed or measured. Seventh is action: continue, expand, redesign, merge, pause or stop.
This structure permits a legitimate finding that an activity is valuable but not yet measurable. The response should then be a limited evaluation, not invented precision. It also permits a finding that an activity produced no observable result while remaining worth another trial. The Board should state the evidence threshold and review date.
Efficiency includes the alternative that was not chosen
An engagement budget cannot be evaluated by outcomes alone. A programme can produce benefit and still be inefficient if a cheaper method would produce the same benefit. Every material activity should identify at least one alternative: remote delivery, regional partner delivery, a smaller cohort, a combined trip, an open online resource, a targeted consultation or no intervention.
The counterfactual is especially important after the pandemic proved that many governance and training functions could continue remotely. That experience did not prove physical meetings obsolete. It created evidence about which functions degrade online and which do not. Institutions should use that evidence rather than returning automatically to the pre-2020 pattern.
Remote access has costs. Platforms, captioning, moderation, security, translation and time-zone duplication require resources. A hybrid meeting can cost more than either format alone. Remote entities may struggle to enter informal conversations where options form. Efficiency analysis should include those losses rather than treating a video link as free equivalence.
Local partners can improve reach and reduce travel, but they introduce another accountability question. A network operator group or training partner may understand the context better than headquarters. It may also have its own sponsors, membership boundaries and selection patterns. Contracts or grants should state deliverables, independence, beneficiary data and conflicts without turning the partner into a branch office.
Member authority needs a budget view before commitments harden
Publishing an annual plan is valuable only if members can influence it before major venue contracts, employment commitments and recurring sponsorships make change expensive. RIPE NCC publishes a draft Activity Plan and Budget for member comment before Board approval. APNIC's Executive Council approves an activity plan and budget. ARIN's Board adopts the strategic plan and budget. LACNIC's financial and assembly structure provides annual budget scrutiny. The precise authority differs, but the governance problem is common.
Members need a functional engagement statement at draft stage. It should show prior-year actual cost, current-year forecast, next-year proposal, outcome evidence and change rationale for meetings, training, fellowships, travel, communications and policy work. A large increase should identify the additional result expected. A flat budget should explain whether inflation reduces delivery or efficiency offsets it. A cut should identify which beneficiary or risk loses coverage.
The vote or comment mechanism should not require members to choose between endorsing every engagement activity and attacking community itself. Functional detail allows a member to support technical training while questioning executive travel, or defend regional policy work while asking whether two conferences can share infrastructure.
Board decisions should record how material objections were treated. Consultation is not a veto, but unexplained non-response teaches members that participation is decorative. The year-end report should return to the commitments made at approval and state what was delivered.
A common engagement statement would improve all five registries
The five institutions do not need identical programmes. Their regions differ radically in distance, language, market structure, travel connectivity, regulation and training capacity. They do need a minimum common account if they want to claim that engagement supports the legitimacy and performance of the registry system.
The statement should begin with a reconciliation from audited expenses to engagement functions. It should identify what is included in staff cost, consultancy, travel, venue, sponsorship, communications, grants and shared overhead. Where a cost serves more than one function, the allocation method should be stated.
It should then report a small set of common indicators: distinct organizations reached; operating networks reached; economies and languages covered; first-time and repeat participation; funded and unfunded participation; completed training and retained competence; decisions or services changed; policy risks addressed; cost per intervention; failed or discontinued activities; and follow-up due.
Each indicator needs a caveat. Organization counts do not equal votes. Operating-network status can be difficult to classify. Economy coverage does not prove national representation. Competence assessment can alter entity behaviour. Decision changes may have several causes. Cost allocation is approximate. Stating these limits improves the report because it prevents the measures from becoming a new mythology.
An independent evaluator should sample a few high-cost activities each year. Independence need not mean a permanent external bureaucracy. A rotating reviewer can test the evidence trail, entity distribution, attribution and alternatives. The contract and findings should be public, and management should answer recommendations. The evaluator should not be paid according to a positive result.
The objective is disciplined public value, not an empty calendar
Regional registries should not retreat into silent database administration. Operators need training. Members need places to challenge decisions. Technical communities need coordination. Governments need accurate explanations. People excluded by cost, geography or professional networks need routes into knowledge and discussion. These functions can protect the open Internet and improve the records on which networks rely.
Their value is weakened when the institution treats delivery as proof. A thousand attendees are not a mandate. A hundred meetings are not a policy result. A completed course is not retained competence. A fellowship is not representation. A trip is not a relationship outcome. A communications impression is not an informed member.
The strongest defence of engagement is therefore an honest account of what it did and did not achieve. If a programme improved account security, show the protected action. If a meeting changed an implementation risk, show the decision trail. If a fellowship built capability, show the later use without appropriating the fellow's identity. If public-policy work prevented a technical error, show the text and acknowledge other contributors. If an activity failed, say so and change it.
The institutional incentive matters. A registry receives fees because networks need a recognized registration relationship in its service region. Engagement should not become a way to transform that dependent revenue into a self-validating claim of ever broader authority. It should remain a funded means to improve service, evidence, access and accountability.
That is the budget line worth defending: not community engagement as a ceremonial category, but a portfolio of specific public goods whose beneficiaries, costs, results and limits members can see.
Distribution matters more than the average entity
An engagement programme can improve its average result while leaving the most dependent networks behind. Suppose a training portfolio reports high satisfaction, strong completion and low cost per learner. The average may be driven by employees of large operators who already have English-language materials, experienced colleagues and reliable access to online platforms. A small island provider, rural network, public university or new entrant may face the highest need and the highest delivery cost. Efficiency measured only at the mean can reward the institution for serving the easiest audience.
The report should therefore show distributions. Completion and retained competence should be grouped by delivery mode, language, subregion, first-time status and organization type where cohort size permits. Travel and meeting support should show which barriers were addressed. Service changes should identify whether they benefited a broad user population or a narrow set of repeat entities. Small cells must be suppressed or combined to protect privacy; that is better than publishing an apparently precise table that identifies individuals.
Distribution also matters inside organizations. A registry may reach a member's public-policy representative while missing the engineer who maintains resource records. It may train a technician without reaching the person authorized to change production configuration. It may send an election notice to a corporate contact who has left. Organization coverage should not be treated as complete merely because one email address or attendee appears.
The institution can classify the intended role for each intervention. Governance information should reach authorized member contacts. Operational training should reach staff able to apply it or transfer the knowledge internally. Security notices should reach both technical and accountable contacts. Public-policy evidence may need executives, engineers and legal teams. The outcome account should test whether the right role was reached, not only whether someone from the organization attended.
Unequal cost is not automatically inequitable. Reaching a remote or underrepresented network may cost more per person and still produce higher marginal value because the institution has little other evidence from that environment. The Board should publish when it intentionally accepts a higher unit cost to widen capability or reduce a material blind spot. That is a defensible subsidy when the purpose and result are explicit.
Engagement claims need an expiry date
Programmes acquire constituencies. A meeting has a host network, regular attendees and suppliers. A fellowship develops alumni. A training course has instructors and certifications. A public-policy initiative forms relationships. Once these structures exist, an annual renewal can feel safer than a fresh test. The benefit claimed in the first year may be repeated long after the original problem changes.
Every programme should have a review horizon. A new activity can receive a limited period to establish a baseline and test demand. A mature activity should state the result that justifies continuation and the condition that would trigger redesign. A long-running event should periodically retest location, frequency, format, audience and relation to decision-making. A course should be retired or revised when the relevant technology, service or error pattern changes.
Expiry does not require automatic abolition. It requires a new reason. The institution should not say that a regional meeting remains necessary because it has existed for fifteen years. It should show the current service, evidence or coordination need. A fellowship should not survive solely because alumni value it; it should show which access barrier remains and whether the selection design addresses it. A travel circuit should not renew because the same invitations arrive; each destination should remain linked to a current purpose.
The review should include the option of transfer. A registry may incubate training or a local event and later support a capable regional partner to deliver it. Transfer can lower headquarters control while preserving value. The partner needs its own accountability, and the registry should avoid presenting independent community work as institutional output. A clear handoff can be evidence of success rather than loss of territory.
What the Board should ask before approving the line
Board scrutiny improves when questions are stable across years. Which exact risks or member needs justify the proposed portfolio? Which populations are currently missing? What changed because of last year's expenditure? Which activities failed, and what was learned? Where is participation concentrated? Which costs are fixed, which can stop, and which contracts are already committed? What alternative delivery methods were tested?
The Board should also ask what the institution will not claim. Will attendance be cited as regional support? Will fellows be described as representatives? Will a government meeting be counted as influence without evidence of a policy connection? Will a training completion be called deployment? Stating the limits in advance prevents promotional language from outrunning the evidence at year end.
Finally, directors should ask who can challenge the result. Members need the underlying non-personal totals and methods. Entities need a route to correct their classification or report a harmful experience. Staff need permission to report a failed activity without losing budget automatically. Independent reviewers need access to samples and decision records. Accountability works when negative evidence can travel upward.
An engagement budget that survives these questions becomes easier to defend. It is no longer a soft category protected by the word community. It is a set of measured interventions whose public value can be argued, corrected and, when necessary, ended.
Sources and analytical limits
The RIPE NCC 2024 Activity Plan and Budget and 2026 Activity Plan and Budget support the cited divisional costs, staff-inclusive activity scopes and named meeting, travel, outreach and consultancy expenses. Budget values are plans, not audited actual outcomes.
APNIC's December 2024 financial report, 2024 Budget Submission and 2025 fellowship description support the Engagement and Development amounts, stated travel savings, programme structure and fellowship conditions. The workstream accounts do not disclose every beneficiary-level outcome.
ARIN's 2024 budget, quarterly travel reporting page and 2024 Annual Report support the cited travel, meeting, outreach, training and reporting practices. Budgeted categories do not establish the value of an individual trip.
LACNIC's audited 2023 financial statements and 2025 budget support the travel, fellow, outreach, training, community-project and total-expense figures. The audited 2023 accounts and 2025 budget should not be read as one continuous actual series.
The article does not rank the registries, convert currencies, infer waste from a high category share or infer success from a low one. Accounting boundaries, labour, geography, event models and years differ. It proposes an outcome method for future member scrutiny and treats missing current AFRINIC comparability as a disclosed evidence limit rather than a number to be estimated.