Summary

  • A standard non-member registration for NANOG 97 plus three nights in the official hotel came to USD 1,727 before tax, transport, parking, visa-related expense, incidental meals or working time. It is a transparent minimum-price scenario, not an estimate of what an average attendee spent.
  • NANOG’s audited 2024 accounts show that meeting fees supplied only about 43% of reported meeting-program expense. Cash sponsorship exceeded meeting-fee revenue, so the meeting model drew on both fee revenue and sponsorship rather than on the fee line alone.
  • Student registration at USD 100 and virtual registration at USD 150 created materially cheaper routes, but the public evidence does not show who used them, who still could not participate, or whether remote attendance produced the informal relationships and repeated visibility available in person.
  • Membership and attendance are different institutional choices. A USD 100 annual membership brings governance rights and a USD 25 meeting discount; the discount alone does not recover the annual fee across the usual three-meeting year.
  • The available numbers justify better access reporting, not a verdict that NANOG excludes a defined population. Useful disclosure would add employer payment, travel origin, repeat attendance, assistance, participation mode and longer-term outcomes without asking attendees to surrender unnecessary personal detail.

The bill that appears before the journey begins

The published venue was in Bellevue, Washington. NANOG 97 was scheduled for three days, from 1 to 3 June 2026, at the Hyatt Regency Bellevue. The published standard non-member price was USD 950. The official room block listed a standard room at USD 259 a night. Three nights therefore added USD 777. Before a hypothetical attendee buys a flight, boards a train, pays hotel tax, parks a car, arranges care at home, obtains travel documents or accounts for three working days, the visible scenario is already USD 1,727.

That number is worth calculating because registration prices are often discussed in isolation. It is also easy to misuse. USD 1,727 is not NANOG’s estimate of the cost of attendance, and it is not an observed average. It is simply the sum of two published prices under one explicit scenario: standard non-member registration and three nights in the official hotel. Some attendees will stay fewer nights, share a room, use another hotel, live nearby or receive a corporate rate. Others will arrive early, depart late, pay considerably more for travel or face currency-conversion and visa-related expense.

The scenario excludes tax, transport, the USD 37 daily self-parking or USD 47 valet charge, meals outside those included in registration, and the value of time. Its analytical virtue is not completeness. It is that every included assumption is visible.

The published minimum changed with timing. An early non-member ticket of USD 800 plus the same three hotel nights came to USD 1,577. Waiting until on-site registration raised the ticket to USD 1,250 and the two-line scenario to USD 2,027. The difference between early and on-site versions is USD 450 before any travel cost changes. A person whose employer can approve travel six months ahead can therefore face a different price from someone whose shift rota, visa appointment, family circumstances or budget authority becomes clear only near the meeting.

Again, timing does not prove unfair treatment. Later prices can serve legitimate purposes: encouraging early commitments, improving catering forecasts and reducing the risk of holding costly venue capacity for uncertain demand. NANOG itself carries meeting commitments. Its audited statements disclose maximum hotel and food-underutilisation commitments for future meetings through 2026 of approximately USD 918,545 at the end of 2024. That figure is not an amount currently due, and it should not be presented as a bill. It shows that reserving a physical meeting carries risk well before every attendee has made a decision.

The entity carries timing risk too. NANOG 97’s published terms stated a USD 50 cancellation fee in the early and standard windows and USD 100 in the late window. After the stated late-window deadline, cancellation was not refundable. A registration could not be rolled into a future meeting, although a substitute could attend. The hotel required cancellation at least 48 hours before arrival to avoid a one-night charge. For an employer sending several staff, these may be manageable business contingencies. For a person paying privately, an illness or rejected travel plan can turn an access decision into a sunk cost.

The visa-letter sequence makes the asymmetry especially concrete. For NANOG 97, an applicant had to register and pay before NANOG would issue an invitation letter. The request asked for passport details, an email address, hotel reservation information and the person’s company name and address. An invitation letter is supporting documentation, not a guarantee that a visa will be granted. The public pages do not tell us how often people use this process, how long approval takes, how many applications fail or whether fees are refunded after a refusal. It would be unreasonable to infer an exclusion outcome from the sequence alone.

It is nonetheless a real allocation of uncertainty: the entity must make several commitments before the state has decided whether travel is possible.

Even smaller conditions reveal how access is managed as a bundle of obligations rather than a single purchase. In-person registration included general and breakout sessions, breakfast each day, lunch on the first two days, and official social and networking events. The N97 attendee communication stated that losing or forgetting the conference badge attracted a USD 800 replacement fee. The public text did not explain the rationale, how often it was applied or whether exceptions existed. It should therefore be reported as a published term, not used to invent a story about enforcement.

The defensible opening conclusion is modest. NANOG makes the first line of the access bill unusually legible. The registration page allowed prospective attendees to see four in-person tiers, a stable student price and a virtual option. The venue page made one hotel rate and parking charges visible. These disclosures improve budgetability. But the cost of entering the room is produced by more than the organisation that sells the ticket, and the price of becoming consequential in the room cannot be read from a checkout page.

A meeting financed by more than the people wearing badges

A high registration price can invite a simple story: entities pay for the meeting. NANOG’s audited 2024 statements show a more complicated bargain. Meeting fees brought in USD 1,244,354. Cash meeting sponsorship brought in USD 1,576,133. In-kind sponsorship added USD 116,001, consisting of USD 56,001 in connectivity and USD 60,000 in an enterprise cloud or systems contribution. The three meeting-linked revenue lines together were USD 2,936,488.

Meeting fees were therefore not the largest of those lines. They represented approximately 42.38% of the three-line total, cash sponsorship 53.67% and in-kind sponsorship 3.95%. Compared with the reported meeting-program expense of USD 2,888,886, meeting fees alone covered about 43.07%. Cash sponsorship and meeting fees together came to USD 2,820,487, USD 68,399 below programme expense. Add the in-kind contribution and the three event-linked lines exceed that expense by USD 47,602.

This arithmetic needs careful boundaries. It does not prove that NANOG made a profit on its meetings. Revenue categories, timing, restrictions, overhead and organisation-wide activity cannot be collapsed into one improvised profit-and-loss account. NANOG reported total 2024 revenue and support of USD 3,002,722, total functional expense of USD 3,635,406 and, after other income, a USD 287,784 decrease in net assets. End-of-year net assets were USD 2,539,602. The meeting commitment maximum mentioned above was about 36.17% of those net assets, but it was a contingent maximum rather than a current liability.

These statements describe an organisation managing both annual operations and future exposure, not a single event clearing a cash register.

Nor does sponsorship revenue establish sponsor control. The annual report says 74 companies sponsored NANOG in 2024. The accounts measure financial and in-kind contribution; they do not measure influence over programme selection, speaker choice or meeting governance. A claim that sponsors control the programme would require separate evidence about decision rights and outcomes. The finance data support a different and important conclusion: cash sponsorship supplied a larger share of the three identified meeting-revenue lines than meeting fees did.

That changes how the admission price should be interpreted. If entity fees supplied the entire reported meeting-program expense, a conversation about affordability might be framed as a direct exchange between attendees and the costs they create. They do not. The 2024 model dispersed the burden. Attendees bought a bundle. The accounts record sponsors supplying cash and services, but they do not specify the complete exchange attached to each contribution. NANOG used organisational capacity and accepted commitments. Employers may have paid an unknown portion of the remaining attendee-side cost; individuals may have paid another unknown portion.

The public record cannot divide that burden between them.

The expense side also makes the physical bundle less abstract. The audited schedule reports USD 2,126,916 in detailed meeting expense, USD 612,755 in salaries allocated to the meetings programme and USD 148,851 in contract services allocated to it. These are not costs that disappear because an observer wishes registration were cheaper. A meeting with rooms, connectivity, food, staff, contractors, live delivery and social space has a material cost. The question is how that cost is distributed, what benefit each contributor receives and whether the distribution advances NANOG’s stated purpose.

Historical evidence shows that this tension is not new. A November 2016 announcement said the registration fee had not increased since February 2008 and that the existing fee did not cover meeting delivery without external sponsorship. Beginning with NANOG 69, the published non-member early price rose from USD 450 to USD 550, standard from USD 525 to USD 650, late from USD 600 to USD 750 and on-site from USD 675 to USD 950. The USD 25 member discount continued. The announcement was management’s explanation, not an independent audit, but its account of sponsorship dependence is consistent with the later audited pattern.

The nominal schedule has continued to rise. A 2008 NANOG 43 reminder listed USD 525 registration before the late date and USD 600 thereafter. In 2012, NANOG 56 listed USD 575 for a late member, USD 600 for a late non-member, and USD 650 and USD 675 respectively on site. NANOG 89 in 2023 listed USD 675 and USD 700 for early members and non-members, USD 775 and USD 800 standard, USD 875 and USD 900 late, and USD 1,075 and USD 1,100 on site. By N97, the comparable non-member numbers were USD 800, USD 950, USD 1,050 and USD 1,250.

These are nominal observations, not an inflation-adjusted affordability series. They say nothing by themselves about wages, company budgets, travel prices, event quality or the share of cost borne by sponsors. Choosing the endpoints that make an increase look most dramatic would create heat without clarity. The meaningful finding is institutional: sponsorship has been part of the meeting’s stated economic design for years, and the audited 2024 numbers confirm that the current participation price is not a full-cost charge.

This is good reason to avoid two symmetrical errors. One is to look at a USD 950 ticket and assume NANOG is collecting more than the meeting costs. The audited organisation-wide result does not support that conclusion. The other is to point to sponsorship and imagine that the attendee price no longer matters. It matters intensely to a person or small organisation facing it without subsidy. The same nominal price can be a routine professional-development expense for one employer, a hard managerial decision for a small network and an impossible personal purchase for an independent engineer.

Public data do not show how many attendees occupy each position.

The invisible employer in the access bargain

The largest missing actor in the published attendance picture may be the employer. A professional meeting occurs during working days. Someone authorises the time, absorbs the absence from ordinary duties and decides whether registration, hotel and travel are business expenses. A delegate from a large operator may experience the ticket as an approved line in a departmental budget. A person from a small provider may be both the engineer who wants to attend and the operator who must cover a shift. A consultant, student or person between roles may face the price directly.

These examples are plausible, but the public evidence does not tell us their proportions. NANOG’s annual report offers employment and organisation charts, yet some displayed categories have unclear denominators or cannot safely be treated as mutually exclusive. They should not be converted into a false census of who pays. The public sources do not report attendee income, employer size, operator size, career level, travel origin, reimbursement policy or personally paid share.

They also do not count people who considered the meeting and declined because of price, visa difficulty, care responsibilities, disability, timing or a refused travel request.

Without those denominators, “affordable” and “unaffordable” become assertions about unseen populations. A USD 1,727 published minimum may sound prohibitive or ordinary depending on the reader’s institution and location. Neither reaction is a measurement. A sound assessment would ask at least four separate questions: Who received financial approval? Who attended without it? Who used a lower-cost channel? Who wanted to participate but did not? The current record answers none of them.

Employer support is not merely a private convenience. It can shape whose presence compounds over time. Network-operations communities can reward accumulated familiarity: knowing whom to approach after a routing incident, recognising the people behind an autonomous system, being invited into a technical discussion, learning which questions can be resolved in a hallway rather than through a formal escalation. This is not a claim that repeat attendance automatically produces authority, or that occasional attendees lack influence. It is a reason to measure recurrence rather than treating each meeting as an isolated transaction.

The 2024 annual report gives total and first-time attendance for three meetings. NANOG 90 reported 734 entities, including 629 in person and 105 virtual; 125 were described as first-time attendees. NANOG 91 reported 738 total, 673 in person and 65 virtual; 184 were first-time. NANOG 92 reported 869 total, 808 in person and 61 virtual; 209 were first-time. Calculated against total attendance, those first-time counts are approximately 17.0%, 24.9% and 24.1%.

Those figures show substantial entry. They do not show retention. The report does not specify enough about the first-time denominator beyond the displayed counts to support finer interpretation, and it does not disclose how many people attended two or three meetings that year, returned the next year, became members, spoke, volunteered for a committee or stood for office. A meeting can welcome hundreds of first-time entities while still having highly uneven return rates. It can also have a stable repeat community without that stability being problematic. The missing fact is the distribution, not a predetermined verdict about it.

The practical distinction matters because employer subsidy can operate repeatedly. One approved trip can introduce a person to the community. Three approved trips a year can make participation habitual. Approval across several years can create a durable network of relationships. If NANOG wants to understand institutional access, it needs to know not only how many new people cross the threshold, but how the opportunity to return is distributed. That inquiry can be conducted without demanding salaries or naming employers in public.

Anonymous ranges—self-paid versus partly or fully reimbursed, organisation-size bands, broad travel distance and number of NANOG meetings attended—would already make the access picture more intelligible.

Such disclosure would also prevent employer support from being mistaken for NANOG support. Where an employer covers registration as well as hotel, airfare and paid time, it is financing an employee’s access in addition to whatever sponsorship or membership it may provide. Only the amount paid to NANOG can enter NANOG’s accounts; hotel, airfare and paid time sit outside them. A financial statement can therefore show revenue received by NANOG without showing who ultimately bore the full attendance cost.

The inverse is also true. A self-funded attendee and a fully reimbursed attendee may produce identical revenue entries. Their exposure to cancellation, currency fluctuation and lost work is radically different. A price schedule can be transparent at the seller’s boundary and opaque at the entity’s. Institutional legitimacy does not require NANOG to solve every employer decision. It does require care in any claim that an open registration page produces an equally open participation opportunity.

Membership is not a cheaper ticket

NANOG’s annual individual membership costs USD 100. Student membership is USD 50, while terms of three years or more are priced at USD 90 per year. Membership reduces each in-person meeting registration by USD 25. Read narrowly as a discount purchase, one year of ordinary membership requires four discounted registrations before USD 100 in fee savings equals the membership price.

NANOG normally holds three numbered meetings in a year. A person attending all three at equivalent tiers would save USD 75, still USD 25 below the annual membership cost. That does not make membership a bad bargain. It shows why the bargain must not be reduced to ticket arithmetic. Membership carries institutional rights, including voting, eligibility for office and committee participation. NANOG says its members elect the board. A member is buying or supporting a continuing relationship with the organisation, not only a coupon.

The distinction helps interpret the attendance figures. At NANOG 90, the reported 629 in-person attendees were 215 members, 410 non-members and four students. At NANOG 91, the 673 in-person attendees comprised 218 members, 445 non-members and ten students. At NANOG 92, 219 members, 573 non-members and 16 students added to 808 in-person attendees. The member, non-member and student categories align with the in-person totals, not the totals that include virtual attendees.

Non-members were consequently about 65.2%, 66.1% and 70.9% of reported in-person attendance at the three meetings. They were not a marginal population. But “non-member” is a registration category, not a synonym for outsider, novice or person without influence. A non-member may be a long-standing attendee whose employer pays admission, a speaker, a sponsor representative, a highly connected operator or someone at a first meeting. Conversely, holding membership does not tell us whether a person can afford regular travel or participates actively.

The numbers suggest that NANOG’s meeting community is broader than its formal membership. That can be a strength: technical exchange need not be limited to the electorate. It also creates a governance question. If the meeting is the primary gateway to relationships and membership is the gateway to formal rights, what fraction of non-members eventually join, and what fraction of members repeatedly attend? There is no public transition table linking attendance category to later institutional activity.

The USD 25 discount is also stable across N97’s four in-person tiers. It therefore softens the ticket price by a smaller percentage as timing gets later: about 3.1% against the early non-member price of USD 800, about 2.6% against the standard USD 950, 2.4% against the late USD 1,050 and 2% against the on-site USD 1,250. Those percentages are calculations, not statements by NANOG. The more important observation is that membership cannot neutralise the timing penalty or the hotel and travel bill. It may reward commitment to the institution, but it is not the principal affordability mechanism.

Membership accountability begins by presenting these choices honestly. A prospective entity should be able to understand that joining provides a modest registration discount and meaningful governance rights. An account that advertises only the discount understates membership. An account that points only to voting rights ignores the recurring economic threshold for exercising those rights in a community whose relationships are strengthened at meetings. Formal eligibility and practical presence are related but not interchangeable.

The public record also leaves open whether participation outside the meeting sustains institutional voice. NANOG identifies digital spaces among its activities, and membership rights are not formally described as conditional on travel. The membership page does not make voting rights conditional on buying three hotel stays. Yet committee work, candidacy and deliberation may benefit from relationships built in person. That is a reasonable hypothesis, not a demonstrated causal chain. Measuring it would require linking participation modes and repeat attendance to later roles without converting governance research into surveillance.

The correct reading is therefore neither “pay to govern” nor “price is irrelevant because voting is remote.” NANOG has separate prices for attendance and membership, separate benefits attached to each, and incomplete evidence about how one leads to the other. Making those pathways visible would help members judge whether the institution is drawing authority from the full community it says it serves.

The USD 150 door is real, but it does not open into the same room

Virtual registration for NANOG 97 was listed at USD 150. Compared with the USD 950 standard non-member in-person ticket, that was approximately 84.2% lower. Choosing it removed the official hotel from the price scenario and avoided most transport costs and the need to be away from home in Bellevue. For someone whose principal purpose was hearing the technical programme, asking questions and following current operational discussion, the difference could be substantial.

NANOG’s attendee information said registered virtual entities could log in through a NANOG profile to watch the live stream, use chat and submit questions. Live transcription was available for general and breakout sessions. Those are meaningful forms of access, not a consolation prize. They can serve people whose budgets, health, location, visas, care responsibilities or work schedules make travel difficult. Earlier NANOG records also show continuity in remote provision: a NANOG 56 reminder in 2012 invited people who could not attend in person to join by webcast.

It would be equally misleading, however, to describe the virtual and physical products as equivalent. In-person N97 registration included official social and networking events as well as meals, and an attendee communication specified that a lightning-talk presenter had to be physically at the event. A live stream can carry a presentation and a question. It cannot place a newcomer beside a prospective peer over breakfast, reproduce the chance encounter after a breakout session or guarantee that a chat message will have the same relational effect as an introduction.

This distinction is difficult to value, precisely because the most consequential benefit of an operations meeting may not be scheduled. The published agenda is only part of what attendees buy. They also buy temporal concentration: several days in which people who normally coordinate through tickets, mailing lists and emergency calls are available in the same place. The social programme and shared meals lower the friction of beginning conversations. A entity may find a mentor, compare an operational practice, resolve a commercial misunderstanding or learn whom to call during a future incident. None of those outcomes is guaranteed.

They are nevertheless part of the possibility created by physical presence.

Remote participation has its own advantages. A virtual attendee can remain close to ordinary duties, choose sessions selectively and avoid travel disruption. A written chat or transcribed session may be more accessible than a crowded social event for some people. Access should not be described as a ladder on which in-person participation is always superior and remote viewing always inferior. The channels provide overlapping but different bundles.

The attendance counts illustrate both the existence and the limits of the remote route. NANOG 90 reported 105 virtual entities out of 734 total, roughly 14.3%. NANOG 91 reported 65 out of 738, about 8.8%. NANOG 92 reported 61 out of 869, about 7%. These percentages are calculations from NANOG’s counts. They show that virtual participation was used, but the decline across three meetings cannot safely be explained from the published data. Location, programme, post-pandemic behaviour, ticket timing or other factors may have mattered.

The counts do not tell us how long virtual entities watched, whether they asked questions or how many later attended in person.

Nor can the price difference be converted into a value difference. Saying that the virtual ticket is 84.2% cheaper does not mean it provides 84.2% less value. Value depends on purpose. An experienced operator wanting two technical sessions may find the virtual offer exceptionally efficient. A first-time attendee seeking professional relationships may find the missing informal layer decisive. A person whose alternative is no access at all may value the live stream greatly while still recognising what it cannot supply.

This is where outcome measurement would improve the discussion. NANOG could ask virtual attendees, at suitable intervals, whether they interacted through chat, made a useful connection, attended a later meeting, became a member, submitted a talk or joined community work. It could ask in-person attendees comparable questions. The purpose would not be to rank people by devotion or to claim that membership is the desired endpoint for everyone. It would be to learn whether lower-cost entry routes connect entities to the institution over time.

The financial evidence cannot answer whether virtual participation subsidises or is subsidised by physical attendance. The audited statements do not allocate programme cost by delivery channel. Streaming, transcription and remote interaction have costs; a physical venue, catering and social programme have others. Shared staff and systems serve both. Any confident assertion that one group pays for the other would outrun the accounts.

The reviewed public record establishes live remote participation—streaming, chat, questions and transcription—but does not establish the timing or terms of any later recording access. An access account should therefore distinguish the three routes documented here: admission to the physical meeting, live remote participation and continuing membership in the organisation. It should not fill the documentary gap with assumptions about later distribution.

That distinction could make NANOG’s own openness claim more precise. The organisation describes its community as open to all and lists meetings, events, scholarships, fellowships and digital spaces among its activities. “Open” can mean that no invitation is needed to register. It can mean that technical content is later available. It can mean that membership has broad eligibility. It can also be heard as a claim about practical ability to participate. Those meanings should not be treated as interchangeable.

The USD 150 virtual ticket strengthens NANOG’s access case because it gives a defined live route at a fraction of the in-person cash price. It does not complete that case. To know whether it functions as a bridge rather than a separate audience, NANOG would need to observe transitions and outcomes. The intellectually honest position is to credit the door that exists while asking where it leads.

The student price and the danger of turning history into policy

The most striking number in the N97 price table may be the one that remained unchanged across the sampled schedules: student registration was USD 100. It was USD 100 in early, standard, late and on-site windows. NANOG 89 in 2023 also listed a USD 100 student rate, as did the archived prices for NANOG 56 in 2012. A 2007 Steering Committee record discussed the same nominal student price.

That continuity is significant, but its meaning must be bounded. The N97 page establishes that event’s price. It does not, in the reviewed material, establish a current cap, verification method, number of applications, number turned away, average travel support or the distribution of students by institution and location. An archived student-policy page described verification and a 20-person limit, but that is historical policy. It cannot be attached to the 2026 ticket merely because the price happens to match.

The 2024 attendance figures show a small group reported as students; they do not establish which price each person paid. Four students were counted among 629 in-person attendees at NANOG 90, ten among 673 at NANOG 91 and 16 among 808 at NANOG 92. Those shares are approximately 0.64%, 1.49% and 1.98% of in-person attendance. The increase across those three meetings is visible. It does not tell us why it occurred or whether every student used the student rate.

Student counts are not a measure of early-career participation. Many people entering network operations are not enrolled in education. Some students may already have substantial experience. Apprentices, career changers, technicians at small providers and junior engineers do not fit one educational label. A policy aimed only at verified students can be valuable while leaving other people at the start of a career on the ordinary price schedule.

The historical minutes are revealing because they show that NANOG’s predecessors understood the student price as a subsidy rather than a naturally low-cost seat. In 2007, the Steering Committee recorded that USD 100 did not cover food and that each student at the Toronto meeting cost the meeting budget approximately another USD 100. Entities debated a cap, described students as a long-term investment and considered finding a sponsor dedicated to the student discount. A prospective Bellevue hotel subsidy was also discussed with the aim of keeping attendee room cost below USD 200 a night.

Those deliberations belong to an earlier operating arrangement. They do not tell us how NANOG Inc. currently calculates the student offer. Their value is conceptual. They show that the access question was already understood as cost allocation: if a student pays less than the cost attached to participation, someone else bears the difference. The institution can regard that difference as an investment, seek a dedicated sponsor, limit the number of places or absorb it across the meeting budget. Each choice expresses a view about who the future community should include.

An archived fellowship page provides another historical example. It offered hotel accommodation and round-trip economy airfare up to USD 500 to as many as two people per meeting, with regional eligibility rules and preference for applicants new to NANOG. That was a concrete attempt to address travel as well as registration. It must not be advertised as available in 2026. Current fellowship amounts, recipient numbers, selection methods and coverage were not established from a current policy page in the reviewed public material.

NANOG’s current description of its activities mentions scholarships and fellowships. The audited functional-expense schedule reports USD 40,000 in scholarship expense for 2023 and USD 0 for 2024. Zero in that line does not prove that no assistance existed in 2024. Support could have been supplied in kind, classified elsewhere, attached to a different programme or absent; the statement does not explain which. A responsible account reports the line and the uncertainty together.

The gap between a broad current description and unavailable current terms is precisely where transparency would help. A single annual table could state the number of student registrations, applications for each assistance programme, awards, typical categories of cost covered, funding source and broad outcome. Where a scheme is not offered in a given year, NANOG could say so. Where support is embedded in another financial line, it could explain the classification. This would prevent historical pages, accounting categories and present-day offers from being blended into an inaccurate promise.

Assistance also raises a problem of discretion. A public fee schedule is impersonal: anyone who meets a stated category and deadline can budget against it. A fellowship or waiver may target need more effectively, but if criteria, available places and award patterns are unclear, prospective entities cannot estimate their chance of support. Neither model is inherently superior. A low fixed student price offers predictability. A selective travel grant can reach costs the ticket does not. The strongest access design may combine transparent categorical prices with clearly described, measured assistance.

Selection itself deserves care, though it is a different question from the total cost assessed here. Publishing recipient names or detailed financial circumstances can create privacy and status risks. Useful accountability does not require that. NANOG could disclose aggregate applications, awards, broad regions, career stages, first-time status and covered cost categories. It could publish conflict rules and describe decision responsibility without exposing applicants.

The historical record also warns against romanticising subsidy. The sampled continuity of the nominal student price is meaningful, but the hotel component in the N97 scenario is USD 777 for three nights. Even a free registration would leave travel, lodging and time. A USD 100 ticket cannot by itself make Bellevue accessible from every part of North America, still less from outside the region. The official 2023 NANOG 88 hotel announcement, for example, listed USD 299 a night plus 15.7% occupancy tax and a USD 4 tourism fee. Location and local price can dominate the discounted ticket.

At the same time, the student rate should not be dismissed because it does not solve everything. Against the N97 standard non-member ticket, it removes USD 850 from the registration line. Against on-site non-member registration, it removes USD 1,150. That is a material intervention. The appropriate question is how many people can use it, what other burdens remain and whether comparable routes exist for people whose access constraints do not fit student status.

History gives NANOG a language for this conversation. Earlier stewards openly discussed the amount by which student participation exceeded the student payment, who might finance it and why the community could treat that cost as a long-term investment. A contemporary account could be equally explicit while reflecting today’s institution, prices and channels. What it should not do is borrow an archived fellowship promise to fill a present-day evidence gap.

Attendance totals reveal entry, not the distribution of opportunity

NANOG’s annual report does more than many event organisations by publishing meeting-by-meeting counts. For 2024, it shows in-person and virtual attendance, member status, student registrations, first-time participation and a reported count of women. This creates a factual base for scrutiny. It also demonstrates how quickly a number can be made to answer a question it was never designed to answer.

Consider first-time attendance. A total of 125 first-time entities at NANOG 90, 184 at NANOG 91 and 209 at NANOG 92 suggests that each meeting brought a substantial new cohort. The temptation is to call that evidence of inclusion. It is evidence of entry under NANOG’s definition. It does not show who considered attending but could not, how the cohort differed from the established community, or whether first-time attendees returned.

Consider non-members. They were a clear majority of the reported in-person population at all three meetings. That can demonstrate that formal membership is not required for physical attendance. It cannot show that institutional voice is evenly distributed. We would need to know how non-members participate, whether they later join and how membership intersects with speaking, committee work and voting.

Consider the reported gender figures: 97 women at NANOG 90, 101 at NANOG 91 and 125 at NANOG 92, displayed as 13.2%, 13.6% and 14.3% of total attendance. These provide one reported dimension, not a complete representation measure. The categories, collection method, response rate and treatment of people outside a binary are not sufficiently described in the relevant presentation to support a broader claim. Repeating the numbers is legitimate; pretending they close the question is not.

The same restraint is needed with employment and organisation demographics. If categories overlap or denominators are unclear, adding them as if they were mutually exclusive produces mathematical certainty where none exists. A well-designed access report should publish a short methodology beside every chart: who was counted, whether a response was optional, whether multiple choices were allowed, which participation modes were included and what “first-time” means.

Methodology is not decorative. Imagine that first-time status is calculated across total attendance, while member status is broken out only for people in person. Comparing the two without noticing their different populations can generate false ratios. The 2024 report’s member, non-member and student counts add exactly to in-person totals, which is useful. The first-time denominator beyond the displayed total-based calculation is not clear enough for more detailed inference. Stating that limit protects the value of the underlying numbers.

The most useful missing measure is a recurrence distribution. Instead of publishing only total and first-time counts, NANOG could report how many entities were at their first, second, third-to-fifth and sixth-or-later meeting. It could do so separately for in-person and virtual participation, and show transitions without naming individuals. That would reveal whether entry is followed by return and whether a small core accounts for an increasing share of attendance.

Recurrence should not be treated automatically as concentration. Repeated attendance is often desirable in an operations community. Trust develops, knowledge accumulates and experienced people help newcomers. A large returning cohort can be evidence of value. The accountability question is whether routes into that cohort are broadly available and whether costs systematically interrupt them. Answering it requires matching return patterns with voluntary, aggregate information about payment and barriers.

An annual access survey could ask whether registration, accommodation and travel were paid by an employer, sponsor, educational institution or the attendee; whether approval covered working time; and whether the person would have attended without that support. It could ask broad travel-distance bands rather than precise origin. It could invite non-attendees who started but did not complete registration to select a reason, while making clear that no answer affects future participation.

There are obvious cautions. Response rates may be uneven. People may not know which departmental budget bore a cost. Small samples can make combinations of region, career stage and employer size identifying. The answer is not to collect every conceivable attribute. It is to publish a narrow set of questions, aggregation thresholds and non-response rates, then resist conclusions that the sample cannot support.

NANOG could also distinguish reach from depth. Reach measures how many people can see or hear programme content through the documented channels: in-person attendees and live virtual entities. Depth measures opportunities to interact: questions asked, chat participation, mentoring, social attendance, speaking and volunteer roles. Continuity measures return and progression over time. No single metric needs to be a target for every person. Together, they would show whether lower-cost channels connect to richer forms of participation for people who want them.

The organisation’s 2024 aspiration to average 1,000 in-person attendees per meeting makes this measurement timely. The report also describes a goal of meeting profitability shaped by location, size, sponsorship and local cost. Growth, profitability and access are not naturally identical objectives. More attendance can spread fixed costs and enlarge the community. It can also increase venue commitments, favour large destinations or make informal navigation harder for newcomers. A location that improves the budget may impose greater travel costs on some entities.

Those tensions do not demand that NANOG choose access over financial durability. An organisation with declining net assets and large venue commitments cannot promise participation without financing it. They demand that decisions be assessed across both ledgers. A meeting budget can improve while the attendee-side burden rises; attendee prices can remain stable while hotel and airfare costs change; a sponsor can reduce the ticket burden while receiving legitimate commercial visibility. Governance improves when these effects are reported separately.

NANOG may already hold much of the information needed to start. Registration records may permit analysis of participation mode, ticket category and repeat history, subject to identity and retention practices. Financial records contain fee, sponsor and expense lines. Post-event surveys can ask narrowly drawn questions about payment and value. The missing move is not an elaborate ranking of people. It is a repeatable public account that links prices, support and participation outcomes without claiming more than the data show.

A proportionate access account

What, then, would better accountability look like? It should begin with the strengths of the present model rather than treating disclosure as an indictment. NANOG publishes its registration tiers. It gives a material student reduction and a much cheaper virtual route. It describes what in-person admission includes. It publishes annual reports and audited statements. It uses sponsorship so that attendee fees do not carry the reported meeting-program cost alone. These are substantial foundations.

The next step could be a compact annual access account built around five ledgers.

The first is the entity-price ledger. For each meeting, it would reproduce the registration tiers, official hotel range, included meals, cancellation conditions, virtual price and the date each tier became available. It would add transparent sample scenarios rather than an asserted average: perhaps a local attendee, a three-night hotel attendee and a virtual attendee. Every excluded cost would be listed. Historical prices could be shown in both nominal and inflation-adjusted terms if NANOG wanted to discuss trends.

The second is the cost-incidence ledger. It would place meeting-fee revenue, cash sponsorship, in-kind sponsorship and meeting-program expense beside each other, with a note explaining that line arithmetic is not profit. Where possible, NANOG could show which sponsor benefits are part of the exchange and which services were provided in kind. This would make clear that lowering attendee prices is partly a question of who else finances the meeting, without inviting unsupported claims about influence.

The third is the assistance ledger. It would state the number of student registrations, the current terms of any fellowship or scholarship, application and award totals, cost categories covered and aggregate funding. If an audited scholarship line is zero because assistance is classified elsewhere, the account could explain that. If no programme is available, it could say so. Historical schemes would remain in an archive clearly dated, valuable for institutional memory but not confused with current offers.

The fourth is the participation ledger. It would separate in-person and live virtual reach, adding any later content reach only on documented terms; publish first-time and recurrence bands; and show member, non-member and student categories against explicit denominators. Optional demographic information would include methods and non-response. This ledger would answer who entered and who returned, not why.

The fifth is the barrier-and-outcome ledger. A brief voluntary survey could record broad payer type, travel-distance band, whether cost or approval affected the attendance decision, and whether the entity achieved the purpose that led them to register. Follow-up at a suitable interval could ask about return, membership, speaking and volunteering. NANOG should publish response rates and treat the results as respondents’ experience, not a census of everyone who might have attended.

These ledgers need not expose individual finances. Nor should they create a bureaucratic burden that costs more than the insight is worth. Much of the price and finance material already exists. A few stable questions can be more valuable than a sprawling annual questionnaire whose categories change each year. The core design principle is comparability: publish the same denominators often enough to observe change.

Targets should be chosen after a baseline, not before. It would be premature to decree a correct share of students, virtual entities or employer-paid attendees without understanding NANOG’s purpose and population. A useful first-year commitment might simply be completeness: report participation mode for all registrations, repeat bands where records allow, assistance applications and awards, and survey response rates. Later goals could address demonstrated bottlenecks.

Any intervention should also preserve the economics that make the meeting possible. If evidence showed that late approval disproportionately burdens small operators, NANOG could test deadline or cancellation changes without assuming every tier must be flat. If hotel cost emerged as the main barrier, it could compare venue strategies or targeted lodging support. If virtual entities reported strong technical value but weak social connection, the response might be facilitated small-group sessions rather than pretending a stream can reproduce a reception.

Sponsorship could be evaluated in the same practical manner. The question is not whether commercial support is pure or corrupting. It is how much attendee cost it offsets, what benefits sponsors receive and what safeguards preserve independent programme decisions. The financial contribution is observable. Programme influence requires separate governance evidence. Keeping those records distinct protects sponsors from insinuation and the community from complacency.

Visa-related access deserves its own count. NANOG could report how many invitation letters it issued, how many requests arrived after relevant deadlines and, where voluntarily disclosed, how many applicants ultimately attended. It should not collect immigration outcomes that entities do not wish to share. Even a simple count could show whether the register-and-pay-first sequence is affecting a handful of people or a recurring portion of the community.

The badge replacement charge is another candidate for proportional review. A published fee of USD 800 is large enough to matter. Reporting how often it was charged and, if applicable, waived or refunded would show whether it is mainly a deterrent or a frequent attendee expense. The same principle applies to cancellation fees: conditions can be justified by real commitments, while aggregate use helps members judge their practical effect.

An access account should finally explain what NANOG cannot control. Airfares, government visa decisions, employer budgets and individual care duties sit outside the organisation’s authority. Venue selection, ticket tiers, assistance terms, cancellation policy, virtual design and information disclosure do not. Legitimacy grows when an institution distinguishes these boundaries instead of either claiming responsibility for every barrier or treating all external costs as someone else’s problem.

What the price proves—and what it does not

The facts support a firm conclusion about the structure of access. NANOG’s registration price is only one component of attending. Under an explicit N97 scenario, standard non-member registration and three official-hotel nights reach USD 1,727 before many unavoidable or possible costs. A virtual ticket reduces the cash threshold dramatically, but it offers a different participation bundle. Student pricing lowers one line even more, while current public evidence is limited public evidence to describe the whole assistance landscape.

The accounts support an equally firm conclusion about financing. Reported meeting fees did not carry the meeting programme alone. In 2024, meeting-fee revenue was smaller than cash sponsorship and equal to about 43% of reported meeting-program expense. Physical meetings have material costs, and NANOG was managing future hotel and food commitments while reporting an overall annual decrease in net assets. Any affordability proposal that ignores those facts is not serious.

What the evidence does not support is a finding that NANOG excludes poor operators, discriminates through price or gives sponsors control. We do not know who pays personally, which employers reimburse attendance, who declined to come, how income or organisation size is distributed, or how assistance reaches entities. We do not know whether recurring in-person visibility causes later institutional authority. The proposition may be plausible enough to investigate; it is not established enough to declare.

This boundary is not timid journalism. It is the centre of the governance issue. Institutions often answer a question about distribution with an average, a total or an open-registration claim. Yet a total attendance record cannot identify absent people. An average cost can conceal radically different travel origins. An open form can coexist with unequal capacity to use it. Equally, a high published price can coexist with generous employer support and growing entry. Without the missing denominators, every sweeping conclusion tells us more about the speaker than the meeting.

NANOG’s strongest response would be neither defensive nor confessional. It would be empirical. Show the full published-price scenarios. Show how meeting expense is divided among fees, sponsorship and organisational resources. Show current assistance terms and aggregate use. Show first entry, return and transitions between virtual, in-person and membership participation. Ask who ultimately bore the attendee-side cost. Explain the limits.

That approach would also give the community a better argument about value. A meeting is not merely three days of presentation content. If it were, the live stream might be a sufficient substitute for many entities. Its distinctive value lies in concentrated interaction, operational trust and the possibility that a first conversation becomes a durable professional relationship. The price of that opportunity should be assessed against its value, not caricatured as rent for a chair.

Value, however, does not dissolve burden. The benefits of repeated presence may be greatest precisely because they accumulate. If access to that accumulation depends heavily on employer size, geography or personal means, NANOG should want to know. If the evidence instead shows broad employer support, strong virtual-to-in-person transitions and healthy return across organisation sizes, that would be important evidence in favour of the current design. Measurement leaves room for either result.

The title of this article deliberately refers to entering “the room,” but the evidence describes several thresholds. There was the ballroom in Bellevue, with breakfast, sessions and the conversations between them. There was the live online space, where questions and chat travelled without airfare. There is the membership electorate, with voting and eligibility for institutional roles. A person can cross one threshold without crossing all three.

Good access policy does not require making every room identical. It requires saying what each contains, what it costs, how people move between them and where the institution has evidence of a blocked route. NANOG already publishes enough to reveal the outlines of that system. Its audited accounts show shared financing. Its attendance report shows entry at meaningful scale. Its price pages show real lower-cost choices. Its missing denominators show the next responsibility.

The price of entering the room is therefore neither just USD 950 nor automatically USD 1,727. The first is a ticket under one timing and membership condition. The second is a bounded hotel scenario. The fuller cost includes travel, time, risk and whatever support a entity can command. The institutional price includes the sponsorship, staff and commitments that make the room possible. Until those burdens and outcomes are reported together, NANOG can demonstrate openness of offer more confidently than equality of practical access.

That is not a verdict against the meeting. It is a case for a clearer bargain. Publish enough of the access ledger that members, sponsors, employers and prospective attendees can see what each is carrying. Preserve the lower-cost routes, but measure whether they connect to the relationships and rights people seek. Treat historical assistance as history and current support as a current obligation to explain. Above all, let the evidence determine whether the room is widening, rather than asking the price tag to answer a question it cannot.

Sources