Zone based booth pricing and how to draw the zone lines defensibly

The zone map is pinned to the wall in the sales office, and it is the whole of your zone based booth pricing. Premium is shaded in orange and covers the front third of the hall, standard is white, value is the grey strip along the back and down the east wall. Somebody drew it with a highlighter, probably in a meeting, and it has been redrawn twice since without anyone revisiting the logic.

A seller points at a block of orange behind the registration build. Those positions renew worse than anything in the white area, and she has been discounting them off the premium multiplier for two editions to get them sold. The map says they are premium. The contracts say they are not.

Zones are the right instrument. The trouble is almost always the lines.

What is a pricing zone for?

A zone is a way of admitting that positions differ in value while keeping the card small enough for a person to quote from. The continuous version of the truth is a surface, where every position has its own price, and nobody can sell from that.

So you approximate. Three or four classes, each with a multiplier, each drawn on a map an exhibitor can look at. The approximation is only useful if the positions inside a class are more like each other than they are like positions in the next class, which is a statistical statement with a statistical answer.

Drawing the lines by eye fails in a specific direction. People draw zones around geography, because geography is what a map shows: the front of the hall, the main aisle, the wall. Value follows geography loosely and breaks from it constantly, which is how a block behind the registration build ends up shaded orange.

What should you rank positions on?

Rank positions by value, then cut the ranking. The question is what to use as the value.

Paid rate per square foot from the last edition is the obvious candidate and it is contaminated, because your existing multipliers helped set it. A position in the orange zone was quoted 25 per cent above base, so of course it shows a higher paid rate. Rank on that and you will recover last year's map.

Two ways round it. If you have fitted a location model, rank on its fitted value, which strips out both the current multiplier and the negotiated discount. If you have not, rank on the components rather than the price: distance to the nearest main entrance, open sides, aisle traffic class, and distance to the nearest feature area. Those are attributes of the position that your card did not create.

Add one column that is not a price at all: the share of positions in each block that renewed last edition. Keep it out of the ranking score and use it afterwards as a check. If a block ranks high on attributes and renews badly, something is happening there that your attributes do not capture, and it is worth a walk round the hall before you shade it orange.

Aggregate to blocks rather than individual booths. A block is a run of positions along one side of one aisle, which is the unit exhibitors actually perceive and a unit big enough to have a stable mean.

Cutting at natural breaks

Once the blocks are ranked, the cut points should sit where the data already separates. Jenks published the classification argument in the International Yearbook of Cartography in 1967, in volume 7 across pages 186 to 190: choose class boundaries that minimise variance within classes while maximising variance between them.

Work it on twelve block-level mean rates, sorted: 38.2, 37.6, 36.9, 34.1, 33.8, 33.4, 33.1, 32.8, 28.9, 28.4, 27.6, 27.1.

Try cutting after the third and eighth values. The top class has a mean of 37.57 and its squared deviations sum to 0.85. The middle class has a mean of 33.44 and sums to 1.09. The bottom class has a mean of 28.00 and sums to 1.94. Total within-class deviation is 3.88.

Now try cutting after the fourth and eighth instead, which is what somebody would do to make the premium zone a bit bigger. The top class now contains 34.1, its mean falls to 36.70, and its squared deviations sum to 9.86 on their own. The middle class improves slightly to 0.55, the bottom is unchanged at 1.94, and the total is 12.35.

Compare both against the total squared deviation from the overall mean of 32.66, which is 166.01. The first cut explains 166.01 minus 3.88, over 166.01, or 0.977. The second explains 0.926.

The first cut is better, and you can see why without any of the arithmetic: there is a gap of 2.8 between 36.9 and 34.1, and only 0.3 between 34.1 and 33.8. The natural break is where the floor already stopped. The arithmetic matters because on a real hall you have sixty blocks rather than twelve, the gaps are not obvious by eye, and you need a number to defend the choice in a meeting.

A goodness of fit above 0.95 with three classes is a reasonable bar. If you cannot reach it, the honest reading is that your hall does not have three distinct grades of position, and you should be pricing two.

The constraint that keeps premium meaningful

The statistical cut will usually put too much space in the top class, because value on a good floor is broadly distributed and the algorithm has no view about scarcity.

Hold premium under a quarter of net square feet sold. That is a convention rather than a measured optimum, and it earns its place because a premium zone covering 40 per cent of the hall has stopped being a premium and become the standard rate with a surcharge attached.

Put the constraint against the cut. Suppose the show sells 128,000 net square feet, so premium is capped at 32,000. The three blocks that the break analysis put in the top class hold 41,000 square feet between them, which is 32 per cent of the floor. The weakest of the three has to come down to standard, which leaves premium at 29,500 square feet, or 23 per cent.

That is the constraint overruling the statistics, and it should. Record it, because next year the same block will look like a candidate again and somebody will ask why it is not orange. A block that keeps failing to earn the top class has a demand problem, and pricing the back of the hall is a different job from zoning it.

The denominator has to be stable for any of this to mean anything across editions. CEIR's Index measures net square feet of exhibit space sold as one of its four components, alongside professional attendance, exhibiting companies and real revenues, and reports it quarterly. Use the same definition internally, and fix it in the report template, or your zone shares will move because somebody changed what counts as sold rather than because the floor changed. That definition is load-bearing for every other piece of exhibitor analytics you run off the contract file.

Multipliers, and the base rate nobody re-solves

Multipliers of 1.25, 1.00 and 0.85 are the common set, and applying them is where zone schemes quietly become price rises.

Take the zoned floor above: 29,500 square feet premium, 66,500 standard, 32,000 value. At a 30 dollar base the premium zone bills 37.50, so 1,106,250. Standard bills 30.00, so 1,995,000. Value bills 25.50, so 816,000. Total space revenue is 3,917,250, and the blended rate is 30.60.

A flat 30 across the same 128,000 square feet gives 3,840,000. The zone scheme adds 77,250, or 2.0 per cent, and nobody announced a 2.0 per cent increase.

That happens because the multipliers are not weighted to your actual mix. If you want the scheme to be revenue neutral, solve for the base. The weighted footage is 29,500 times 1.25 plus 66,500 times 1.00 plus 32,000 times 0.85, which is 36,875 plus 66,500 plus 27,200, or 130,575. Divide 3,840,000 by 130,575 and the neutral base is 29.41.

Whether you want neutrality is a separate decision, and it belongs with the annual increase. What you do not want is to make that decision by accident, in a spreadsheet, while thinking about geography.

Positions that sit on a line

Every zone map generates boundary disputes and they are all the same dispute.

Decide three things in advance and write them on the map. A booth whose area falls across a boundary takes the zone holding the majority of its square feet. A booth whose position is upgraded because a feature area moved keeps its contracted zone for that edition. Zone assignment is a property of the position, published before sales opens, and does not change during the sales cycle for any reason. Moving price inside the cycle is a separate instrument with separate rules.

That last one is the expensive discipline and the one worth holding. An exhibitor who signed in March at standard and finds in September that their position was reclassified to premium will not renew, and the reclassification was worth a few thousand dollars.

Where this stops

Zones are a coarse approximation and they will always misprice the positions nearest the lines. A block at 34.1 that just missed premium is being sold at the same multiplier as a block at 32.8, and the exhibitor in the 34.1 block is getting a better deal than anyone. That is the cost of a card people can quote from, and it is usually worth paying.

The ranking also depends on attributes you chose. Distance to an entrance and open sides are easy to measure and they are not the whole of value. Category adjacency, sightline to a hanging sign, and whether an aisle is on the route from registration to the keynote room all move demand and most floors record none of them.

The quarter rule has no evidence behind it. It is a scarcity heuristic that keeps the top class small enough to feel scarce, and if your waitlist for premium positions is four times the available stock, a quarter is too generous a cap rather than too tight. Set it from your own waitlist rather than from the convention where you can.

This week, take last edition's contract file, group the booths by aisle block, compute the mean paid rate per square foot for each block, and sort the list. Print it next to the zone map. Every block whose rank contradicts its colour is either a mispriced position or a discount nobody recorded, and you will know which within about ten minutes of looking at them.