Summary

  • NANOG's audited 2024 accounts recorded $1,576,133 in meeting sponsorship, 52.490% of total revenue and support by calculation. The organization also reported 74 sponsors, making commercial support economically central without establishing that any company bought a program decision.
  • Public sponsor benefits start at $5,000 and sell visibility, access and event inventory. The Program Committee separately says it selects presentations, while current guidelines allow vendor expertise but bar promotional and proprietary pitches.
  • The separation is incomplete as a public accountability record. The detailed prospectus is available only through a request form, the current Program Committee page includes sponsor relationships among its goals, and no current sponsor-specific conflict, recusal or acceptance record was found.
  • The proportionate remedy is not a vendor ban. NANOG could publish a sponsor-rights matrix, revenue-concentration bands, current sponsor-status handling during proposal review, aggregate recusals and clear labels for content that enters by a route other than ordinary Program Committee selection.

The wrong question is whether money exists

An engineering conference without money is not more independent; it is often simply smaller, more expensive to attend, less reliable or unable to exist. Hotel space, food, production, connectivity, registration systems and staff time do not become free because the program is technical. NANOG's 2024 audited financial statements make the scale visible. Meeting expenses reached $2,888,886. Future hotel and food commitments could reach $918,545. Connectivity and cloud or system services worth $116,001 arrived as in-kind support.

The wrong question, then, is whether sponsors are present. NANOG advertises them, counts them and designs event inventory around them. The useful question is whether a reader can distinguish the rights bought with money from the decisions that still require technical judgment—and whether that distinction is observable when a sponsor's employee also has something valuable to teach.

That is harder than looking for logos. A logo proves recognition. A lounge proves an event benefit. A company name beside a speaker proves employment or affiliation. None of those facts alone establishes instruction, a paid talk, a favourable vote or operator adoption. Conversely, a page that assigns selection to a committee does not prove that every possible conflict is identified and handled. Both capture and complete insulation require mechanisms, not impressions.

NANOG's public record is stronger than a blank page. It separates sponsor benefits, proposal review, presentation rules and privacy obligations. It is weaker than a fully executable firewall because some decisive cells—the detailed package, current sponsor-status treatment inside review, company-level revenue concentration and exception records—remain unavailable to public inspection.

What the public product actually sells

NANOG's sponsorship value page starts with a price: opportunities begin at $5,000. The language presents sponsorship as a way to support the technical community while showcasing technologies and solutions. It promises brand visibility and reach, amplification of a sponsor's message, opportunities to connect with influential people and decision-makers, and a role in empowering or inspiring the community.

Those are meaningful goods. Visibility is valuable precisely because NANOG gathers engineers, vendors, operators and institutional leaders in one place. A message delivered near a technical program may acquire more attention than the same message on a generic advertising surface. Access to people who make infrastructure decisions can be commercially important even when nobody promises a purchase or a policy result.

The public bullets do not, however, promise a technical presentation, keynote, policy outcome or adoption by network operators. The distinction is textual and should not be inflated. It tells us what the public offer says; it does not disclose every term in a signed agreement. NANOG's prospectus request page places the detailed document behind a form. A reader cannot compare the full package schedule with the public program process without requesting commercial material.

One published benefit does approach the edge of the agenda: an affiliate event. NANOG says an organization must have at least $8,000 in sponsorship or contracted space to qualify, and the event may not conflict with the scheduled NANOG program. That wording simultaneously creates a paid adjacent opportunity and preserves a formal priority for the scheduled program. It does not say who allocates competing slots, how refusals work, whether exceptions exist or what an affiliate event may claim about its relationship to NANOG.

The NANOG 97 sponsor roster shows why one generic “sponsor” label is inadequate. It distinguishes premium tiers, connectivity, a lounge, Beer n' Gear, peering, showcases, mixers, socials, meals and services. Paid and in-kind value is distributed throughout the event experience. A connectivity provider contributes a different good from a social host; a showcase offers a different relationship from a logo; a meal slot creates a different form of proximity from a technical session.

This is not evidence of impropriety. It is evidence that the rights matrix needs more than two columns marked sponsor and non-sponsor. NANOG is already selling differentiated surfaces. Public accountability should describe the boundary with comparable precision.

Who selects the technical program

The current Program Committee page gives the committee responsibility for developing the meeting program. Its duties include recruiting content, reviewing submissions and voting on what appears. NANOG's proposal-review process describes a chain: a proposer submits an abstract and preferably draft slides; the PC performs an initial review; a shepherd may help; reviewers assess slides; the committee selects presentations and builds the agenda.

Draft slides matter because they let reviewers see the proposed substance rather than infer it from a polished abstract. That preference is a practical control. It narrows the gap between what is promised and what would reach the stage. It is not a blind-review guarantee, a conflict rule or a record of why one proposal defeated another.

The 2024 Annual Report, available through NANOG's annual-report page and 2024 report PDF, provides a denominator that many conference reports omit. The PC considered 206 submissions and accepted 105, an acceptance rate of 50.971%. The three event rows—70 submitted and 34 accepted, 75 and 36, 61 and 35—sum exactly to the annual totals. NANOG reports about 1,600 minutes of technical content per meeting and 4,800 minutes across the year.

Those figures show scarcity and work. Roughly half of the proposed content did not become an accepted presentation. They do not show who sponsored the event, which proposers worked for sponsors, whether anyone recused, how conflicts were resolved or whether a rejected proposal was deferred or withdrawn. The final agenda is an output, not the complete review ledger.

This distinction matters because the same PC page that assigns program authority contains a phrase that complicates a simple firewall story. Among its goals is developing mutually rewarding agreements with sponsors and presenters. The phrase may describe ordinary relationship-building for a successful event. Registration and attendee experience are understandably among the committee's measures of success. But placing sponsors and presenters inside one published relationship goal makes it especially useful to state how sponsor status is treated when content is reviewed.

The phrase is not proof of influence. It is a reason not to substitute institutional confidence for an executable rule.

Vendor expertise is not a defect

A crude independence test would exclude vendors from the program. NANOG's current presentation guidelines take a more useful approach. Vendor talks are allowed. Their purpose may not be to promote a vendor or a proprietary solution, and speakers should not promote their employer. Logos are generally confined to the first and last slide, with limited use where a logo is actually part of an illustration. An official announcement of the guideline update makes the current rule easier to date.

The guidelines prefer multi-vendor or open-source examples. A presentation built around a single-vendor configuration receives lower priority, but it is not automatically excluded. Submitted material also may not carry confidentiality notices that would prevent redistribution.

That design recognises a real feature of network operations: engineers employed by companies often possess the most direct knowledge of failures, deployments and tooling. Removing them would not remove commercial power from the market; it might remove the evidence needed to understand the market. The appropriate unit of scrutiny is the claim and its evidence, plus any decision-relevant conflict—not the mere fact of a salary.

The rules still leave enforcement questions. Lower priority is not a numerical score. “Promotional” can be obvious when a deck is a product pitch and difficult when a case study highlights capabilities available from only one supplier. A logo limit can be counted; the balance between operational lesson and brand advantage requires judgment. That is why proposal-level process and aggregate disclosure matter more than a rhetorical promise of neutrality.

Article 027 does not attempt to grade individual vendor talks. It asks a prior governance question: which route admitted the content, which rule applied, and what public trace distinguishes reviewed technical material from purchased adjacent inventory?

The financial centre of gravity

NANOG's financial-report index links the audited accounts. The 2024 figures deserve careful reading because they can support two opposite exaggerations.

Meeting sponsorship brought in $1,576,133, down from $1,811,504 in 2023. Meeting fees contributed $1,244,354, down from $1,376,413. Total revenue and support was $3,002,722, and membership dues contributed $64,526. In-kind support was $116,001: $56,001 of connectivity and $60,000 of cloud and system services.

From those base values:

  • cash sponsorship supplied 52.490% of total revenue and support;
  • cash plus in-kind support supplied 56.353%;
  • cash sponsorship exceeded meeting fees by $331,779;
  • cash sponsorship was 24.426 times membership dues, or 26.224 times when in-kind support is included;
  • cash sponsorship fell by 12.993% from 2023, while meeting fees fell by 9.594%; and
  • cash sponsorship equalled 54.559% of the $2,888,886 in aggregate meeting expense.

These are calculations from audited categories, not ratios printed by NANOG. They establish economic centrality. They do not establish cause. The accounts do not say why sponsorship and fee revenue declined, and a simultaneous decline cannot be turned into a market narrative without other evidence.

Cash sponsorship plus meeting fees was $68,399 below aggregate meeting expense. That arithmetic does not prove a meeting-level deficit funded by a particular source. Organization-wide accounts contain other revenue and expense categories, and the statements do not trace each dollar into a named session, meal, network service or fellowship.

The same caution applies to timing. Deferred sponsorship of $810,333 is not extra sponsorship earned in 2024. It is a liability or timing item associated with future service delivery under the accounting policy. The statements treat sponsorship as revenue from event-service arrangements. A large deferred balance confirms that contracts and performance periods matter; it cannot be added to earned revenue to make sponsorship look larger.

The audit reports no donor restrictions on the $116,001 of in-kind support. That does not reveal the private service specification or let a reader assign cash sponsorship to particular goods. It establishes the audited category and its two components.

NANOG's net assets decreased by $287,784 in 2024, compared with a $363,395 decrease in 2023. This is organizational context, not a reason to infer pressure on a specific PC decision. The stronger inference is narrower: when more than half of annual support is sponsor-related, clarity about purchased rights and non-purchased authority protects both the community and the sponsors who do not seek program control.

Seventy-four sponsors, no concentration measure

The Annual Report counted 74 sponsors in 2024. That number signals breadth, but breadth in names is not the same as diversity in revenue. One company may contribute much more than another. Cash, in-kind service and event-specific benefits may be distributed unevenly. The audited statements do not publish company-level amounts.

Exact company figures may be commercially sensitive. A proportionate disclosure need not expose every contract. Concentration bands could show, for example, the share contributed by the largest sponsor, the top five and the top ten, while keeping names and prices confidential. A reader could then distinguish a broad revenue base from dependence on a small group.

The 2024 Annual Report also records strategic goals to maximise sponsor inventory and regain seven premium sponsors. It asks whether badge scanning provides value or revenue and whether NANOG should explore a full-event expo room. Those are questions and goals. They are not evidence that badge scanning was deployed, that a lead product was sold, or that a permanent expo room opened.

They do show the organization actively developing the commercial surface. That is ordinary management. It also means the boundary document should evolve as the product evolves. A privacy statement written for registration data may not, by itself, tell an attendee what happens after a voluntary badge scan at a booth. A program guideline may not explain how an expo demonstration is labelled if it later appears in an agenda interface.

Data access is a separate right

NANOG's Privacy Policy says it does not sell or rent Personal Data and does not share it for a third party's own marketing without consent. That is a substantive boundary. Sponsor support is not, on the public wording, a general licence to take the attendee database.

The policy also permits service providers to process data and allows NANOG to use information for its own stated purposes. It leaves room for consent. A voluntary scan can create a specific exchange that differs from a bulk transfer, depending on the screen, the fields, the recipient and the retention terms. The public strategy question about badge scanning does not answer those implementation details.

The correct conclusion is therefore neither “sponsors get the list” nor “no sponsor can ever receive contact information.” The policy supports a narrower rule: third-party own marketing requires consent, and Personal Data is not sold or rented. A current sponsor-rights matrix could add operational clarity: whether scanning exists, what the attendee sees, which fields move, who retains them, how consent can be withdrawn and whether NANOG receives revenue.

Treating data as a separate row prevents a common form of mandate laundering. Access to a room does not imply access to registrations. A branded table does not imply permission to market to everyone present. Consent to one scan does not imply endorsement of a product or authority over the technical program.

A labelled sponsor item tests the boundary

The NANOG 89 agenda includes one 30-minute hackathon recap/presentation explicitly marked as sponsor-linked. That is useful evidence because it refuses to hide the relationship. It is not enough to establish the admission route. The public agenda does not say whether the item was submitted and selected through the ordinary process, attached to contracted inventory or invited as a report from an adjacent activity.

One item cannot support a pay-to-play accusation. It also prevents a categorical claim that sponsor-labelled material never appears on an agenda. The accountability need is small: label the route as well as the relationship. “PC selected,” “invited technical report,” “affiliate event,” “sponsor showcase” and “contracted session” tell readers different things even when all the content is useful.

Communications around NANOG 95 provide another bounded example. The attendee archive for June 2026 and the September 2025 announcement archive describe branded tables or banner placement and a half-day sponsor room. These are concrete event benefits. They do not expose the full contract or establish a permanent rule across meetings.

The NANOG 97 event page presents agenda and sponsor information on separate surfaces. That separation is useful navigation. It remains information architecture, not an audit of private influence.

Conflicts: a historical rule is not a current guarantee

An official NANOG mailing-list archive from January 2005 described Program Committee conflict disclosure, recusal and blind rating. It shows that the institution has previously articulated proposal-review controls. It does not, without a version chain, establish that the exact rule operates unchanged in 2026.

Current financial controls are more recent but govern a different surface. NANOG's financial-policies page links an Internal Financial Controls Policy that prohibits self-dealing and requires approvals and conflict signatures. Its Financial Reserves and Investment Policy establishes responsibility for another part of the organization's money. These controls matter, but they are not evidence of proposal-level recusals.

NANOG's Board responsibilities include budgets, committees and institutional affiliations. That gives the Board an organizational oversight role. It does not reveal how a sponsor-affiliated PC member is identified in the review system, when that reviewer withdraws, who receives the reassigned proposal or how many recusals occur.

Public proposal confidentiality is a legitimate constraint. Releasing proposer names, comments and votes could discourage candid review and expose competitive information. Aggregate execution records need not do that. NANOG could state the current rule, whether sponsor status is visible at each review stage, how conflicts are declared and reassigned, and annual counts of recusals or exceptions without naming proposers.

The Code of Conduct adds one useful outer boundary: removal of sponsorship can be a consequence. Sponsorship is therefore conditional rather than an untouchable institutional status. Conduct enforcement is not program independence, but it shows that payment does not purchase immunity from every NANOG rule.

A six-row public rights matrix

The clearest disclosure would fit on one page.

Brand exposure. List recognition, logos, showcases, lounges, social functions, meals and service categories. State whether placement is guaranteed or subject to change. No technical authority follows from these rights.

Affiliate activity. Publish the $8,000 eligibility floor, scheduling constraint, allocation actor, labelling rule and exception path. Make clear whether the activity is part of the PC-selected program.

Technical program. Name the PC as selector, link the review and presentation rules, state how sponsor status and conflicts are handled, and publish aggregate execution figures. Keep vendor expertise eligible under the same evidence and anti-promotion standards.

Sponsor-linked agenda content. Where a recap, showcase or invited report appears in the agenda interface, label both the sponsor relationship and the admission route. A useful report does not become less useful because its route is visible.

Attendee data. State whether lead capture is offered; present the fields, recipient, consent and retention terms; separate a voluntary scan from bulk access.

Financial use. Continue audited aggregate reporting, add sponsor concentration bands and distinguish cash from in-kind support. Do not imply restricted allocation where none is documented.

This matrix would not publish negotiated prices, private reviewer comments or attendee records. It would tell each audience which institutional claim can be made from which benefit. It would also protect sponsors: a company buying a lounge should not be suspected of buying a talk because the public record fails to say otherwise.

What remains unknown

Several questions cannot be closed from the reviewed public sources. The complete current prospectus and signed package terms are not visible. Revenue concentration across the 74 sponsors is not published. The public process does not say whether sponsor status is shown to reviewers at every stage or provide a current proposal-level conflict and recusal procedure. The exact continuity of the 2005 practice is unresolved.

No public record reviewed here establishes whether a current package ever promises a stage slot. The route used by the sponsor-labelled NANOG 89 hackathon recap is unknown. Badge-scanning implementation after the 2024 strategic question is unknown. The accounts do not trace restricted sponsorship funds to named outputs. No reviewed case shows a sponsor requesting a program alteration, the conditions of a rejected sponsorship or affiliate-event request, or acceptance rates for sponsor-affiliated and other submissions.

These are public-record limits, not findings of misconduct or proof that private controls do not exist. A serious article must be willing to stop at that boundary.

The strongest defence of the existing model

The strongest defence is practical and substantial. NANOG stages expensive meetings and produces thousands of minutes of technical programming each year. Cash sponsorship covered only 54.559% of aggregate meeting expenses in 2024. Even after adding meeting fees, the two categories were $68,399 below that expense line. In-kind connectivity and cloud services may deliver capabilities that would otherwise require cash purchases. Vendors employ engineers with direct operational knowledge.

Sponsors may support NANOG because the program is valuable, not make the program valuable by controlling it. A committee that ignored attendee experience, presenter relationships and financial sustainability would not be more technically serious; it could simply fail to deliver an event. A firewall designed as social separation would be counterproductive.

That defence strengthens, rather than weakens, the case for bounded disclosure. When the legitimate relationship is visible, allegations require better evidence. When paid rights are enumerated, benefits outside the list can be questioned without treating every logo as suspicious. When current conflict handling and aggregate recusals are published, the institution does not have to disclose private reviews to show that a rule executes.

The boundary NANOG can prove

Three shortcuts that would make the record worse

The first shortcut is to infer power from proximity. Sponsors are visible because visibility is part of the product. Their employees will appear in corridors, social events and technical sessions because companies are part of the network industry. A photograph, roster or employer line can identify that proximity. It cannot identify the content of a conversation, an instruction to a reviewer or a right in a contract. Treating proximity as the missing mechanism would make careful disclosure dangerous: every additional name would become raw material for insinuation.

The second shortcut is to treat a content rule as a complete governance system. The ban on promotion is valuable, but it answers what a presentation should contain, not how a reviewer handles a financial relationship. The logo rule can be enforced even if a conflict is undisclosed; a perfect recusal can occur even if a later slide becomes promotional. Content, conflict and commercial rights are separate controls. NANOG needs no single grand principle when three narrower statements would be more accurate.

The third shortcut is to equate absence from the public site with institutional absence. The full prospectus may contain restrictions that are not visible. The review system may capture conflicts that are not reported. Badge scanning may have been rejected after the Annual Report posed the question. None of those possibilities can be claimed as fact, but neither can their opposite. The public-accountability problem is not that private controls have been disproved. It is that a reader cannot distinguish a documented control from a hoped-for one.

These shortcuts point to the same editorial discipline. A relationship requires an actor, a right or act, a route and an observable result before it can support a claim of influence. A firewall requires a rule, a responsible actor, a trigger, a response and some execution evidence before it can support a claim of independence. The evidentiary burden should be symmetrical.

What an annual sponsor-boundary note could report

NANOG already collects much of the information needed for a compact annual note. The Annual Report counts sponsors, submissions, acceptances and program minutes. The audit reports cash and in-kind categories. Event pages identify inventory. The privacy policy states the outer data rule. Bringing those fragments together would be a reporting exercise, not a new bureaucracy.

A first table could reconcile the year. It would show the number of sponsors, cash sponsorship, in-kind support, meeting fees, concentration bands and material categories of delivered benefits. The 2024 line would retain the audited $1,576,133 cash value and $116,001 in-kind value without pretending that accounting recognition equals unrestricted cash available on the same date.

A second table could describe program-boundary execution. It would keep the existing 206/105 proposal totals and add only non-identifying information: whether sponsor status is available in the review tool, number of declared conflicts, number of recusals, number of reassigned proposals, and number of agenda items admitted by invitation or a contracted route. Zero would be meaningful only if the field and reporting rule existed; a blank should not be converted into zero.

A third table could cover attendee data. It could state whether badge scanning or another lead product operated during each meeting, which fields were optional, whether the transfer went directly from attendee to sponsor, and where consent and deletion information could be found. NANOG need not publish who scanned whom. It would merely distinguish an offered mechanism from the 2024 strategic question.

The note should also identify exceptions. If a sponsor benefit changed because of schedule, safety, conduct enforcement or production failure, an aggregate count and general category would demonstrate that the agreement remains subject to rules. The Code's sponsorship-removal consequence would become more than an outer possibility without exposing a private case.

Such reporting would not measure “influence” as a single number. Influence can be legitimate, diffuse and difficult to isolate. It would measure the more tractable conditions under which money receives a right, a reviewer steps away, data moves by consent or a program label signals a different route. Those are things an institution can actually record.

Why the recommendation is narrower than full transparency

Commercial terms sometimes need confidentiality. A sponsor may negotiate a package around availability, timing and service contributions. Publishing every price could weaken future bargaining. Program reviewers need space for candid criticism, and unsuccessful proposers need not have their ideas or identities exposed. An attendee who shares a business card or scans a badge does not thereby agree to public disclosure.

The proposed boundary respects each of those interests. Concentration bands replace company amounts. Aggregate conflict counts replace names and comments. Route labels describe institutional status rather than commercial detail. Data-product notices describe the transaction before consent rather than revealing the resulting contact.

This is also why a maximal “transparency portal” would be the wrong symbol. A short, versioned statement tied to the Annual Report is easier to maintain and test. When a new benefit such as an expo room is introduced, one row changes. When the review system changes, the conflict row gains a new effective date. When no lead product exists, the data row says so for that event year. Versioning prevents a current page from being projected backwards, the same mistake that would otherwise turn the 2005 conflict record into an unbroken 2026 guarantee.

The value of this limited record appears during disagreement. A sponsor can point to the package category it bought. A proposer can point to the review route. NANOG can point to the rule that controlled data or conflicts. Critics still may investigate influence, but they must do so with evidence of a mechanism rather than by converting every visible commercial relationship into a verdict.

The public record supports a measured conclusion. NANOG sells valuable exposure and event access. Sponsorship is central to its finances. Its published Program Committee process separately assigns presentation selection to reviewers, and its content rules reject employer promotion and proprietary pitches while keeping vendor expertise eligible. Its privacy policy denies a general right to sell or rent Personal Data or share it for third-party marketing without consent.

The same record does not prove a complete sponsor-program firewall. Detailed package rights, concentration, current proposal-conflict execution, the route for every sponsor-labelled agenda item and badge-scan implementation remain unclear. Nor does it prove that a sponsor purchased a talk, controlled a vote, obtained unrestricted attendee data or caused an operating decision.

That evidentiary balance is the point. Money does not become authority merely because it is necessary. A committee page does not become independence merely because it names a selector. NANOG can make the boundary stronger by publishing the rights that money buys and the controls that still apply after the invoice is paid. The result would credit sponsors for enabling the meeting, protect vendor engineers as legitimate technical contributors and give the community a record sturdy enough to distinguish proximity from power.

Sources