Tips on workshop and conference organization
Tips on workshop and conference organization
I have spent a significant amount of time organising workshops and conferences over the past few years and I believe I learnt things that could be useful to others.
My first advice if you decide to organise an event, or if you have been invited to join a team, is to think carefully about your time commitment. Organizing is time consuming throughout the process and you need to factor in the emails, the decisions and negotiations, the technical aspects (review process, etc) and the learning time (how to do things, asking others). It is reasonable to see it as 1-2h work per week, which will spread a bit unevenly but you should plan it like this.
When you are new to it, it is tempting to be willing to do more, to invent new processes or new features, etc. But keep in mind that anything new or non previously tested will need a lot more time to double check and test every single details, otherwise any mistake will cost you a lot of time when the consequences impact all your authors and reviewers. Thus, a corollary of my first advice is to keep things simple and easy the first time you organise something. Stick to standard processes, always lean toward the easiest way to do things well and if you want to do some things differently, just pick one novelty that you will take time to implement. Do count extra time on your planning for this special thing you want to do.
The main steps towards organising a workshop or conference are:
Find a team. Make sure to choose or join a team you know and trust, people you feel comfortable working with and who have a reasonably clear idea of what they want to do. If you are the lead organiser, make a very clear plan of what the event is about and who you need to succeed. As a rule of thumb, the smaller the team the better. Only scale up the team where you see a need for it. The minimum is 2 people, ideally 3-4 for small events involving a review process.
Choose a timeline and stick to it. Early on you need to decide when is the even, and then backwards the decision date (announcement), the duration of the review phase (plan for margins between reviews and decisions in case of delays), the submission date (plan margin for the technical parts of review assignments), and enough time before for you to get ready and announce your event. A rule of thumb is to have about 6 months between this step and the event, more is great.
Create your OpenReview. (if you want to use this platform, otherwise CMT is also fine for small events). OpenReview will ask you a ton of questions when you create a venue: do you want ACs, SACs, should the reviews be public, etc. Ideally do this step together with your core team so there's more than one pair of eyes. The support can help you out if you make a mistakes, but some decisions are irreversible. Generally, have a look at the OpenReview guidelines if you've never done this before. You will need a contact email for this. Assign a member of your team to this: you will receive a decent number of emails (see below) and someone needs to be in charge of this (and potentially less in charge of other things).
Announce your event. This is crucial, make sure all your target community is aware and has enough time to prepare. Double check all communication text as anything unclear will trigger emails and this is exactly what you want to avoid. Make sure the links work and point to the right pages, make sure the website is visible and looks as expected... From this point on, you cannot make major changes anymore: dates are fixed, call for papers are clear and fixed (number of pages, template...).
Submission and review process. OpenReview lets you change the submission and review forms via a JSON text box. This is a real pain but language models can help you a bit. Do test your chosen features before you open submissions or before the review phase starts. Read the Open Review instructions on the review stage. Hire your reviewers as early as possible, typically 1-2 months before the submission deadline. I wrote a short tutorial on how to extract Open Review IDs from existing venues. Reviewer assignments will take you some time so plan enough days to do it in your timeline. Plan weekdays for this, not weekends, because the Open Review support only works on week days and you may need them if something odd happens. Also it's just a good idea to plan to do work on week days.
Reviewers will submit last minute. You can send a reminder, but don't harass them 1 week before the deadline. Send an email to all reviewers with pending tasks 1-2 days before the deadline, then again 1 day after so you can plan how much extra reviews you'll need.
Make decisions. Take the venue size into account, plan enough time to do this. You can split the workload but eventually it works best to book 2h with all the core organizing team to finalize things together online. Again, several pairs of focused eyes are important at this stage.
The fun part starts :) Now the event is close, it is all about final local arrangements (hopefully you have a local team and some help for this). You will get emails from accepted authors. If that gets too much to handle, you can consider having an FAQ page on your website and making sure all latest instructions are public and clear.
Try your best to come relaxed and open to the event itself. It is a lot of work to put an event together but ultimately the most important is to have kind, available, excited people around. If catering is late, it really doesn't matter. You did all this work to get a bunch of people together, not to make sure they get coffee. Come early to the venue and try the IT system (do it the day before if possible), make sure there are signs to orient participants, and sleep early before it starts ;)
Key links:
My short tutorial on how to extract Open Review IDs from existing venues (previously accepted papers for instance)
Open Review guides: https://docs.openreview.net/getting-started/hosting-a-venue-on-openreview
The Bandit Workshop, Imperial College London, UK (2019)
The Summer School in Theoretical Computer Science (EPIT in French), CIRM Marseille, France (2021)
ICLR Blog Post Track (co-creator and Program chair) (2022, 2023, 2024)
Algorithmic Learning Theory (international conference, Program Chair) (2023)
CoLLAS Workshop Track (2025)
ICML Tutorials (2025, 2026)
RLC Workshops (2025, 2026)
Mannheim RL Workshop (2025, 2026)
European Workshop on Reinforcement Learning (EWRL) 2025 in Tuebingen, Germany
AISTATS Workshops (2026, inauguration year)