Webconf workflow

Web Conference Workflow

The Web Conference runs one research-tracks cycle per year, anchored in early autumn, with the conference the following spring or summer. The 2026 edition adds two planning lessons no other pack in this collection teaches as vividly: the same edition offered later submission windows into the same proceedings (short papers, Web4Good), and the conference itself moved by ten weeks mid-cycle (April 13-17 → June 29 - July 3, 2026), so travel and money plans need slack.

The 2026 cycle as a planning template

Milestone	2026 edition date	Offset from abstract
Abstract (research/industry)	Sep 30, 2025	0
Full paper	Oct 7, 2025	+1 week
Short papers abstract / paper	Nov 10 / Nov 17, 2025	+6 / +7 weeks
Web4Good abstract / paper	Nov 23 / Nov 30, 2025	+8 / +9 weeks
Rebuttal window	Nov 24 - Dec 1, 2025	+8 weeks
Notification	Jan 13, 2026	+15 weeks
Camera-ready	Jan 25, 2026	+17 weeks
Author registration (post-postponement)	Mar 30, 2026	+26 weeks
Conference	Jun 29 - Jul 3, 2026	+39 weeks

Every date above is a 2026 anchor for shape, not a prediction: pull the current edition's important-dates page before committing the team.

Backward planning from the abstract deadline

T-12w  Claim set frozen; datasets acquired; ethics/ToS posture settled
        (data-access negotiations are the long pole for web papers)
T-8w   Experiment matrix running (webconf-experiments); baselines reproduced
T-6w   First full draft; web-native framing test (webconf-writing-style)
T-4w   Internal review by someone from the track's *other* discipline
T-2w   Freeze authors + track; sweep co-author EasyChair profiles;
        count per-author submissions against the cap of 7
T-1w   Page-zone audit, anonymity sweep (webconf-submission)
T+0    Abstract registered -- author list now immutable
T+1w   Paper in, with hours of margin (deadlines are strict, AoE)

Two locks distinguish this venue's timeline: the author-list freeze at T+0 and the track choice, both irreversible a week before the paper itself is due.

The intra-edition fallback tree

A miss or a weak draft at the research deadline is not a lost year here:

1. Short papers (+6 weeks; 4 pages including references; main proceedings) — right when the core idea is sound but the full evidence program isn't. Compressing an 8-page draft to 4 is a rewrite, not a trim; decide by T+2w.
2. Web4Good (+8 weeks; main proceedings) — right when the work's center is societal benefit; it is a distinct scope, not a consolation lane.
3. Workshops / companion venues (per-workshop deadlines, companion proceedings) — visibility and feedback without burning the idea's archival slot.
4. Hold for the next edition — WWW 2027 is announced for Dublin, Ireland, so the next cycle's deadlines will land in roughly the same autumn window.

Each fallback has different proceedings placement and prestige; choose by what the paper needs (archival main-proceedings status vs. fast feedback), not by sunk cost.

Capacity planning quirks

• The rebuttal week (late November in 2026's cycle) collided with a major US holiday; name the responding author at submission time and check their November calendar.
• Notification-to-camera-ready was twelve days; pre-draft the de-anonymized version and the artifact deposit during the December quiet period (webconf-camera-ready).
• Budget approvals for registration + travel should start at notification, not at the registration deadline — and the 2026 postponement shows dates can move after booking, so prefer refundable fares and watch official channels, not rumor.

Portfolio planning under the cap

Groups submitting several papers per cycle hit constraints individuals never see. The 7-per-author cap binds on senior authors first: an advisor on nine drafts must drop off two before abstracts are registered, which is an authorship-integrity decision, not a bookkeeping one — being removed to dodge the cap while continuing to contribute violates the spirit of both the cap and the authorship policy. Rebuttal week is the second portfolio bottleneck: five papers times a one-week window means five simultaneous response drafts; assign one owning author per paper in September and let the PI review rather than write. Track collisions are the third: two similar papers from one group in one track invite the panel to compare them against each other — differentiate their framings deliberately or route one to a different track or venue.

After rejection

Route by the review packet's agreement structure (webconf-review-process): a converged evidence objection means running the missing experiments and returning next edition; a fit objection means the sibling circuit — WSDM (autumn deadline, web mining methods), CIKM (spring, broad IR/KM), SIGIR (winter, retrieval), KDD (winter/summer dual-cycle, mining) — inherits the paper with modest reframing. Never resubmit unchanged into an overlapping reviewer pool; panels remember.

Cycle-shape volatility to plan around

Three shape changes the series has actually made, which a multi-year plan should treat as live possibilities rather than history:

• Conference month moves. Recent editions ran in late April/early May (Singapore 2024, Sydney 2025) while 2026 landed in late June/early July after its postponement — the submission window has stayed autumn-anchored, so plan research capacity around September-October and treat the conference date as the flexible end.
• Platform switches (OpenReview 2024-25 → EasyChair research tracks 2026) change co-author onboarding costs; budget a week for profile wrangling whenever a new edition's call names a different system.
• Special-track churn (Web4Good appeared for 2026) means the fallback tree must be rebuilt from the current edition's calls page each year, not reused.

Output format

[Cycle position] T<±n>w relative to current-edition abstract deadline
[Locks ahead] author freeze / track choice / cap count -- dates
[Fallback decision] research full / short / Web4Good / workshop / next edition
[Capacity flags] rebuttal-week owner, camera-ready pre-work, travel slack
[Next actions] <this week's moves, with owning author>