Organising a conference is a chain of constrained decisions, and the useful thing to know about each one is not the best answer but the point at which it stops scaling. What follows is that chain, in the order it actually gets made, with the constraint that drives each link and the size at which the usual answer breaks.
Venue and date
These are one decision. The venue’s free weekends are the tightest constraint in the process, and everything after this point is measured from the date it produces.
At regional scale the venue is borrowed or nearly free: a university lecture theatre, a company auditorium, a civic room. The 2007 edition ran in a corporate auditorium and the 2008 edition in a university room, which is the entire reason both could be priced the way they were. A borrowed room comes with conditions, a fixed capacity, a caretaker’s hours, and no control over the layout, and those conditions become the programme’s constraints.
Where it breaks: when the room’s capacity is smaller than demand two years running. Renting changes the budget by an order of magnitude, and the ticket price follows, which changes who attends.
Track count
How many talks can run at once is decided by the rooms available, not by preference. It then fixes the size of the programme: a single-track day holds six to eight sessions, and a second track roughly doubles it while halving the audience each speaker gets.
A single track has one property that is easy to underrate: everyone at the conference has seen the same talks, so the conversations at lunch have a shared reference. Two tracks buy programme capacity and spend that.
Where it breaks: when the call for papers returns far more good proposals than slots. At that point the choice is a second track, a longer event, or a harder rejection rate, and the third option costs goodwill with exactly the people an event needs next year.
The call for papers window
Three dates matter and they are usually set too late: when the CFP opens, when it closes, and when decisions go out. The gap between closing and deciding is the one that gets compressed, and it is the one that determines whether every proposal is read properly.
Speakers need four to six weeks to write something considered, and accepted speakers need two to three months before the event to book travel and prepare. Working backwards from the date, a CFP that closes less than four months out is already squeezing someone. The mechanics of what the document should say are on the call for papers page.
Where it breaks: when the committee is large enough that “read everything twice” needs a rubric and a schedule rather than an evening: see the review process.
Money
A regional conference has three sources (tickets, sponsors, and a paid workshop or tutorial), and the mix determines how independent it is. Tickets priced to cover catering, sponsorship covering the rest, and occasionally a pre-conference workshop that funds a scholarship or a charity, as in 2007.
The structural risk is a single sponsor covering most of the budget: it is the easiest money to raise and the most expensive to lose, because the programme quietly begins to accommodate it. Several smaller sponsors cost more work and keep the programme independent, which is the trade-off set out under community versus corporate.
Where it breaks: when the budget needs a signed sponsor before the venue can be booked. From that point the event has a cash-flow problem, not a fundraising problem, and it needs a reserve or an underwriter.
People
Volunteer organisers are the norm at regional scale and the most common single point of failure: one person holds the CFP, the schedule, the sponsors and the catering, and the event ends when that person changes jobs. The durable arrangement is the one the Hoedown demonstrated between 2007 and 2008, the event as a format that a different user group can pick up, rather than as a committee.
On the day, the roles that actually need a named person are the ones nobody volunteers for: someone on the door, someone running the clock in the room, and someone whose only job is to answer questions. Everything else can be shared.
Which rooms actually suit an event of this size, with a real example of each, is on venues.
Questions
How many people does it take to run a regional conference?
Three or four across the planning months, and six to eight on the day. The planning roles are programme, sponsorship and logistics; the day roles are registration, room-running and a floating person for whatever breaks.
When should the call for papers close?
At least four months before the event, and further out if the committee is small. Speakers need weeks to write, the committee needs weeks to read, and accepted speakers need months to arrange travel, all of which comes out of the same gap.
Is it better to have one large sponsor or several small ones?
Several, if the work can be found. One sponsor covering most of a budget is the cheapest to raise and the most damaging to lose, and the programme tends to bend towards it without anyone deciding that it should.
What should a first-time organiser cut first?
Scale, not quality. A one-day, single-track event done properly is a better first outing than two days of two tracks done thinly, and it leaves the obvious next move available for the year after.