How to host a conference

Most advice on hosting a conference is a checklist that starts with the venue. That is the wrong first step, and it is why so many first-year events are logistically competent and hard to explain.

Start with the premise. Everything below is derived from it.

Twelve months out: the premise

Write one sentence: who this is for, and what they get that they cannot get elsewhere.

Good: “A one-day conference for people who run internal platform teams at companies over 500 people, where they can talk honestly about what did not work.”

Bad: “A conference for the industry, bringing together thought leaders to explore the future of the space.”

The test is whether somebody could decide, from that sentence alone, that they are or are not the audience. If your sentence needs a comma and an “and” to include a second group, you are planning two events and should pick one.

From the premise you derive nearly everything: how many people, which city, what a ticket costs, what the programme covers, and which sponsors have a reason to be there.

Set the number. How many people, realistically. First-year events overshoot this consistently. It is easier to grow a sold-out 200-person event into a 400-person one than to recover from a half-empty room, and a half-empty room is visible to everybody who came.

Eleven months out: date and city

The date. Check it against your industry’s existing calendar, school holidays in your main markets, religious observances, and the two or three large events your audience already attends. Clashing with a bigger event in the same field is the single most avoidable planning error.

The city. Three factors, in order: can your audience get there without a connecting flight, is there enough hotel capacity within walking distance of your venue, and is there a venue at your size that is not booked. City choice is more constrained by the second and third than most first-time organizers expect, particularly in cities where a large recurring show absorbs capacity for a whole week each year.

Ten to nine months out: the venue

Visit before you sign. Photographs are marketing and floor plans are optimistic.

What to check on the visit

• The main room at your capacity, with the seating layout you want, not the layout it is set in.
• Where the breaks happen, and whether people can actually stand there. This is where your event’s networking lives.
• Distance and route between the main room and the breakouts. A three-minute walk between tracks becomes a ten-minute changeover.
• Natural light, or the honest absence of it. Two days in a windowless basement is a real cost to your programme’s energy.
• Step-free routes, walked, not described.
• Where the loading bay is and how early you can get in.
• Mobile signal inside the rooms, on the networks your attendees use.

What to read in the contract

• Attrition. The minimum room nights or catering spend you commit to, and what you pay if you fall short. This is where the financial risk of a low-attendance year actually sits.
• Cancellation, on both sides, with the dates and the percentages.
• What is not in the room rate: power, rigging, cleaning, security, overtime, and the charge for running past the contracted end.
• Exclusive suppliers. Many venues mandate their own AV and catering, which removes your ability to shop those two lines. Know before you budget.
• Catering guarantee deadlines, and the number you must confirm by.

Nine months out: the budget

Build it per attendee, not in total. Fixed costs first, then per-head costs, then 10% contingency, then three attendance scenarios. The budget template has the full structure and a worked 500-person example.

Two rules from that page are worth repeating here because they change the plan rather than the spreadsheet:

Model the low case and confirm it survives. Fixed costs dominate conference economics, so a 30% attendance shortfall hurts much more than a 30% overshoot helps.

Sponsorship is flat across scenarios. It is sold on a promise and collected before the audience arrives, which is exactly what makes the low case survivable. That is why it comes next.

Eight months out: sponsorship

Sell it before tickets. Sponsors buy the promise of an audience, and the earlier you sell, the more of your fixed cost is covered before you have any attendance evidence.

Build packages from limited assets rather than from logo sizes, keep to four tiers, and price from your own cost base rather than another conference’s rate card. The full structure, with a worked four-tier card and the add-ons that outsell tier upgrades, is in sponsor package ideas.

One number to hold: at 500 attendees, eight to twelve sponsors in total is comfortable. Twenty is a trade show with a conference attached.

Seven to six months out: the programme

Decide the shape before the content. Single track or multi? Half-day or two days? Workshops or not? These decisions constrain the venue, the budget and the ticket price, so make them consciously rather than by accumulation.

Single track is under-rated. It is easier to run, it produces a shared experience that makes networking easier, and it removes the changeover problem entirely. Multi-track exists to serve a genuinely heterogeneous audience — which, if you wrote a good premise, you may not have.

Programme the breaks first. Lay out the meals and breaks, then fit sessions between them. Doing it the other way round is how conferences end up with fifteen-minute breaks that are a coffee queue. Two twenty-five-minute breaks beat three fifteen-minute ones and cost the same programme time.

Protect one unstructured hour. It will be the first thing you cut when the programme gets crowded, and it is the thing most attendees came for.

Getting speakers. Invite the ones you want directly before you open any call. A call for proposals is a good way to find people you do not know and a poor way to build a spine. If you run one, say what you are looking for specifically, and tell everybody the outcome on the date you promised.

Session lengths that work: 30 minutes plus 10 of Q&A for a talk, 45 for a keynote, 90 for a workshop, 45 for a panel with a moderator willing to interrupt. Nothing over 45 minutes without a break, at any point in the day.

Six months out: registration opens

• Publish the programme, or at least the shape and the confirmed keynotes. Selling into an empty programme is hard and it teaches your audience to wait.
• Early bird for eight weeks, then standard. A discount with no end date is a price.
• Have a concession or student rate and mean it.
• Decide your comp policy now: speakers, sponsors, staff, press. Count them in the budget; they eat lunch.
• Set the refund and transfer policy before the first sale, and publish it.

Three months out: operations

This is where the event becomes real work.

• Contract the event app and confirm who submits it to the app stores, which sets your timeline. The full sequence is in the launch checklist.
• Confirm AV requirements room by room, itemised. The base package covers a room, not your event.
• Brief speakers with a one-page speaker brief, and send it again a week out.
• Build the run of show, one row per action, one named owner per row.
• Confirm catering numbers against registration, allowing for no-shows.
• Recruit and brief the crew. Registration and Q&A moderation are the two roles most often under-staffed.
• Sign off signage, with room names exactly as the venue’s doors read.

The last month

• Freeze the schedule two weeks out and check every room and time against the venue’s own booking paperwork.
• Invite attendees to the app one week out, with one reason to open it now. Chase the bounces the same day.
• Rehearse the transitions in the actual room, with the actual crew.
• Print the run of show and the day-at-a-glance. Nothing else.
• Write the incident card: who changes the schedule, who sends notifications, who to call, on what number.

The days themselves

Three things carry a conference operationally, and none of them is the programme.

Registration. Open early, staff it heavily, and put one person on helping people install the app rather than scanning. The queue at the door is where your attendees form their first opinion and where app adoption is won or lost.

Transitions. Between sessions, more goes wrong than during them. Rehearse the break reset — three rooms, two crew, twenty-five minutes — because that is the one that fails.

One person who owns the schedule. Every change goes through them, and every change is followed by a notification that deep-links to the session that moved. Editing without notifying sends people to the wrong room; notifying without editing leaves the app contradicting itself.

Beyond that: watch the arrivals count in the morning, staff Q&A moderation in every room that has it, and send something useful on the first morning so that notification permission is worth having later.

Afterwards

• Session feedback to each speaker within a week, with the questions they were asked.
• Post-event survey within 24 hours, twelve questions, once. The question list is a menu.
• Archive the Q&A, polls and photos before anything expires. The most-upvoted questions across the whole event are next year’s programme, and they are better research than any call for papers.
• Record the twelve numbers that matter, so next year has a baseline. See event metrics worth tracking.
• Write one page of notes while it is fresh: what to keep, what to change, what nearly went wrong.
• Announce next year’s date at the closing session if you have it. It is the cheapest marketing you will ever do.

The four mistakes that cost the most

Starting with the venue. An event with a room and no premise is much harder to fix than an event with a premise and no room.

Over-sizing year one. A full 200-person room builds an event. A half-empty 500-person room damages one, and everybody in it can see the empty chairs.

Under-staffing registration and moderation. Both are visible to every attendee and both are cheap to fix in advance.

Filling every minute. The programme is not the event. The conversations between the sessions are, for most attendees, and they need time and a place to happen.