Space selection appointment scheduling and the queue that decides your floorplan

The exhibitor ranked 63rd arrives for their 11:20 slot with three positions shortlisted. All three went at 10:40. Your salesperson opens the current map, the exhibitor looks at it for a while, and then asks whether they can call back after lunch. Nine minutes of the slot are spent on a decision that will not be made in the room, and the person ranked 64th is now waiting.

Space selection appointment scheduling looks like an administrative job and it is not. The order and the timing decide which exhibitors are standing in front of a map with real choices on it and which ones are looking at what is left.

Selection order is a queue, and the slot is the service time

Everything about a selection sequence behaves the way a single server queue behaves. There is an arrival order you have fixed in advance, a service time per customer, and a finite window before the server closes. Two properties follow and both are worth stating in the language you will use with your sales team.

Total elapsed time is the sum of service times, so anything that shortens the average slot shortens the whole event, and anything that lengthens the tail lengthens it more than people expect.

The value of the service falls as the queue advances. In an ordinary queue every customer receives the same product. Here the product degrades continuously, because each exhibitor removes inventory that the next one might have wanted. The exhibitor at position 400 receives a materially worse service than the exhibitor at position 40, and no amount of scheduling changes that. What scheduling can do is stop position 400 from also receiving a worse experience.

How long should each appointment slot be?

The default is a flat slot for everyone, usually somewhere between five and ten minutes, and it is wrong in both directions at once.

Band the slot by booth size instead. Take a floor of 450 selecting exhibitors, split as 25 above 1,000 square feet, 90 between 300 and 1,000, and 335 below 300.

Give the large band 15 minutes, which is 375 minutes. Give the middle band 8 minutes, which is 720. Give the small band 4 minutes, which is 1,340. Add those and you have 2,435 minutes, or 40.6 hours.

Now price the flat alternative. Eight minutes for all 450 is 3,600 minutes, or 60 hours. Banding saves 19.4 hours, which at five and a half selling hours a day is three and a half days of your sales team's show week.

The saving is only half the argument. The other half is that the flat eight minute slot was actively harming the top of the queue, because a 2,000 square foot exhibitor deciding whether to take two 1,000 foot islands or one 2,000 foot island needs to see the plan, walk the plan, and phone somebody. Cutting that conversation short at eight minutes gets you a smaller contract, and a smaller contract from your largest account is the most expensive thing on this page.

Build the bands from last edition's contract file. You already know how much space each exhibitor took, and booth size predicts decision time better than points, tenure or anything else you hold.

Who actually owns the first day?

Run the concentration figure before you publish anything, because it tells you what your schedule is really allocating.

Take a hall selling 120,000 net square feet across those 450 exhibitors. If your top 20 per cent hold 60 per cent of the space, which is roughly the shape of most B2B floors, then 90 exhibitors hold 72,000 square feet between them.

Under the banded schedule those 90 exhibitors are the 25 in the large band at 15 minutes and 65 of the middle band at 8 minutes. That is 375 plus 520, so 895 minutes, or 14.9 hours. At five and a half selling hours a day, 60 per cent of your floor is allocated in the first two and three quarter days.

Two consequences follow. Anything you want to influence about the shape of the floor, a category adjacency, a new zone, a pavilion, has to be settled before those first fifteen hours, because after that the negotiable inventory is gone. And the remaining 360 exhibitors, who between them hold 48,000 square feet, are selecting from a plan that is already 60 per cent committed, so the map they see matters far more to them than the map the first 90 saw. That bites harder on a floor that is not growing. CEIR put net square feet in the third quarter of 2025 at 8.3 per cent below the same quarter of 2019, and a hall with no new premium positions in it has nothing to offer the back of the queue except what the front declined.

Build a buffer into the same fifteen hours. A no show rate of 8 per cent across 450 exhibitors is 36 appointments that will not happen as scheduled, and most of them will want to be rescheduled within the same window. Leaving one empty slot in every twelve costs you 37 slots of capacity, which is about three hours, and it buys you somewhere to put a rescheduled account without pushing everybody behind them. Without the buffer, a rescheduled 1,500 square foot account either goes to the back of the entire queue, which they will escalate, or gets squeezed in ahead of someone with more points, which the someone will escalate.

Stage the map so nobody selects against stale availability

The failure at the top of this post is a data freshness failure, and it has a fix that costs nothing.

Hold the map static within a batch of appointments, publish the times at which an updated map is released, and send exhibitors the version they will be selecting from at the moment they get their appointment confirmation. An exhibitor who arrives having shortlisted from the 10:00 map and is selecting at 11:20 should have been sent the 11:00 map at 11:00.

Inside the batch, lock a position the instant it is verbally claimed, before any paperwork. The International Roofing Expo 2027 space draw guidelines require the exhibitor to select the space, complete and sign the booth application, and provide the deposit at the scheduled appointment time, and they commit to giving at least one week of notice of that time and date. Requiring the signature and the deposit in the slot is what makes the map trustworthy, because a position marked as taken is actually taken.

The same guidelines take the opposite view on staging, telling exhibitors the live floor plan will be updated continually throughout the process. That is defensible when every selection is locked at the desk in real time, because there is no window in which the plan is wrong. It is the wrong choice when exhibitors are shortlisting from a plan hours before their slot, which is the situation most organisers are actually in, and the reason to publish release times instead.

What do the published sort keys tell you?

Organisers who publish their sort keys are doing themselves a favour, and reading somebody else's is the fastest way to audit your own.

The roofing show sets its keys out plainly. One point for every year the company has exhibited, one for association membership held in good standing at the time of space assignment, one for every 100 square feet of booth space from the previous show year, and one for every $2,500 of sponsorship spend from the previous show year. Two further clauses settle the awkward cases. Companies exhibiting together or in contiguous space are assigned on the points of the company with the lowest total, with no aggregation between them. In a merger or acquisition the surviving company may elect to take the higher of the two totals in place of its own, again with no aggregation.

Three things there are worth copying. The keys are published, so nobody has to ask why they are 63rd. Two of the four are bought with recent behaviour, space and sponsorship from the previous show, which gives your sales team something to sell in the six weeks before the draw. And the co-exhibiting and merger clauses are written down in advance, which is the difference between a rule and an argument at the desk with an account that has just acquired a competitor.

A call based sequence also solves the geography problem. If a third of your exhibitors are overseas, an onsite draw quietly ranks them below the domestic accounts who happened to be in the building, whatever your points say. The same guidelines assume the telephone throughout, asking exhibitors for alternate contact details and working a missed appointment back into the schedule when the exhibitor is able to call in. Running the draw itself by scheduled call removes the geography penalty, at the cost of the exhibitor not being able to walk the position before committing.

Where this stops

Scheduling cannot change who is at the front of the queue. If your standings are wrong, a beautifully banded appointment grid delivers the wrong exhibitors to the good space faster and more politely than before, and the fix belongs in the formula, where no calendar change can reach.

The bands also assume last edition's booth size predicts this edition's decision. For most exhibitors it does. For an account in the middle of a strategy change it does not, and those are precisely the accounts you most want to give time to. Ask your sales owners to flag the twenty accounts they expect to move up or down a band, and hand schedule those.

There is a harder limit underneath all of this. Appointment scheduling allocates a fixed set of positions, and if the positions themselves are badly drawn, the queue just distributes the problem in points order. A hall with 40 positions nobody wants produces 40 unhappy exhibitors whatever their slot time, which is a question about how the plan was drawn, and selling it in a different order changes nothing.

This week, take last edition's selection log, plot cumulative net square feet against elapsed selling hours, and mark the point where the curve crosses 60 per cent. If that point sits later than the first third of your schedule, your slot lengths are wrong at the top. Joining that log to your contract file is also the smallest useful piece of work you can hand exhibitor analytics, because without it nobody can tell you what the queue cost you.