Where dynamic booth pricing in exhibitions helps and where it breaks

Day two, the coffee stand at the top of aisle 600. Two exhibitors who have stood next to each other for six editions are comparing what they paid. One signed in March at the published rate. The other signed in August, after the hall had tightened, and paid 9 per cent more for a position two booths along.

By day three the same comparison has happened four more times, which is what dynamic booth pricing in exhibitions looks like to the people paying for it. By the following March, one of those accounts is asking your seller, on the record, whether the rate they are being quoted is the rate everyone is being quoted.

That is the failure mode of dynamic pricing on an exhibition floor, and it has nothing to do with whether the model was any good.

The assumption that does not survive the show

Airline repricing works on buyers who are anonymous to each other. Nobody on a flight compares fares with the person in the next seat, and if they did, they would not be back on the same route with the same neighbour next year.

Your exhibitors stand next to each other for three days, drink together at the same reception, and belong to the same trade association. They compare rates as a matter of routine, and the comparison is easy because your product is priced per square foot and everyone knows their own square footage. A rate difference of a few per cent is discoverable in one sentence.

The machinery of revenue management transfers to exhibitions well, and which parts of it apply, along with how protection levels get set, has its own treatment. What does not transfer is the assumption that price discrimination stays private. On a floor, it does not stay private for a full day.

What do buyers accept as a reason to change a price?

The useful literature here is not about optimisation. Kahneman, Knetsch and Thaler published a study of community standards of fairness in the American Economic Review in 1986, in volume 76 across pages 728 to 741, built from telephone surveys asking people to judge specific pricing and wage decisions.

Their finding, which became known as dual entitlement, is that people hold a reference transaction in mind, and that the reference gives the buyer a claim on the usual price and the firm a claim on its usual profit. A firm may raise prices to protect its reference profit when costs rise, and that is judged acceptable. Raising prices to exploit a shift in demand is judged unfair, even when it is straightforwardly profitable. The example that carried the point was a hardware store raising the price of snow shovels from 15 to 20 dollars the morning after a large snowstorm, which 82 per cent of their respondents judged unfair.

Map that onto a hall. Your rate goes up 8 per cent because drayage, hall rent and labour went up: that is a cost pass-through, it is defensible in a sentence, and your exhibitor has seen the same cost increases in their own budget. Your rate goes up 8 per cent because the hall filled faster than expected: that is the blizzard, and your exhibitor will read it exactly the way the survey respondents did.

Treat that as a hard constraint on which repricing moves are available to you, sitting outside the optimisation rather than inside it as one more term to be traded against yield. A model that ignores it will produce a recommendation that earns money this edition and costs renewals for three.

Confine it to the genuinely new buyer

The workable version of dynamic pricing in exhibitions runs on stock where no reference transaction exists.

That means late inventory, in the value zone, being sold to accounts that did not exhibit last edition. A first-time exhibitor buying an unsold 200 square foot unit 70 days out has no prior rate, no neighbour they have known for six years, and no expectation about what the position should cost. Whatever you quote is the reference.

It also means that renewing accounts are outside the scheme entirely. Their rate is the published card rate for the edition, fixed, and it does not move because the floor filled. That is the fence, and unlike an airline's Saturday night rule, it is a fence your customers will regard as fair rather than as an obstacle to be gamed.

Anything in the premium zone should stay outside the scheme too, because premium stock is where your long-standing accounts sit and where a discovered rate difference does the most damage. Premium stock has a different instrument, which is holding some of it back.

What is repricing upward actually worth?

Put numbers on the tempting move so the size of the prize is on the table.

Say 9,400 square feet of value zone stock is unsold at 90 days out, listed at 25.50 per square foot. Across the last three editions you have cleared about 55 per cent of the remaining value stock inside that window, so expect 5,170 square feet to sell, which at list is 131,835 dollars.

Reprice upward by 8 per cent, to 27.54, and assume clearance holds. The same 5,170 square feet now bills 142,382, a gain of 10,547 dollars.

Now price the risk. A mid-sized account holding 800 square feet at 30 dollars is worth 24,000 dollars a year to you. The entire upside of the repricing scheme is 44 per cent of one such account for one edition. Lose that account for three editions because they found out that the newcomer two aisles over got a better start, and the scheme has cost you 72,000 dollars to earn 10,547.

The asymmetry is the whole argument, and it does not depend on the exact numbers. Upward repricing on a floor produces a small, certain gain and a large, uncertain loss, and the loss lands on the accounts you can least afford to test.

The direction that does work, and its arithmetic

Repricing downward on unsold late stock has the opposite shape, provided it is confined to inventory your renewing accounts do not want.

Publish a release schedule for value zone stock: 25.50 to 90 days out, 23.50 from 90 to 30 days, 21.50 inside 30 days. Print the dates and the rates on the card at the start of the cycle.

Take the same 9,400 square feet. At flat list you expect 55 per cent to clear, which is 5,170 square feet at 25.50, or 131,835. Say the published ladder lifts clearance to 70 per cent, which is 6,580 square feet. The extra 1,410 square feet clears at the 30 day rate of 21.50, which is 30,315 dollars of revenue that otherwise never existed.

Then charge yourself for the cannibalisation. Some accounts that would have paid 25.50 will wait for 21.50. Assume a third of the original 5,170 do, which is 1,723 square feet losing 4.00 per square foot, or 6,892 dollars. Net gain is 23,423.

Test the worst case. If every one of the 5,170 square feet waits for the bottom rate, the loss is 5,170 times 4.00, or 20,680 dollars, against 30,315 of new revenue. The scheme still clears 9,635 dollars ahead. A published markdown ladder on stock that would otherwise go empty cannot lose money on that stock, which is a rare property in pricing and worth having. The early booking discount sits at the other end of the same cycle and answers a different question.

The condition attached to it is strict. The moment the ladder touches inventory a renewing account would have bought, the arithmetic changes, because now you are discounting sales you already had.

Publishing the rule is most of the work

The mechanism matters less than the fact that it is written down before the cycle opens.

Three lines on the card do the job. Rates for renewing accounts are fixed at the published card rate for the edition and do not vary with availability. Late release rates apply to specified positions in the value zone on published dates. Any account may see the release schedule on request.

Publishing costs you the ability to improvise, which is the point. An unpublished scheme requires your sellers to explain, individually and defensively, why this exhibitor is paying more than that one, and they will not do it. They will discount instead, off the record, and your realised rate will drift down while the model reports that yield improved.

A published schedule also changes what the coffee stand conversation sounds like. Two exhibitors comparing rates and finding a difference that matches a printed rule are having an entirely different exchange from two exhibitors finding a difference nobody can account for.

Where this stops

You never observe the counterfactual, and this is the honest limit on every claim above. The 23,423 dollar gain assumes a clearance lift you cannot verify, because you only ever run one version of the sales cycle. Tracking the clearance curve across editions is ordinary exhibitor analytics, and it gets you a rough read after three years, by which time the market has moved. Estimating elasticity from five editions is the nearest thing to a check on the assumption.

The fairness constraint is real and it is not measurable on your floor. Nobody can tell you how many renewals a discovered rate difference costs, and the accounts that leave rarely say that was the reason. You are pricing against a risk you can name and cannot size, which argues for staying well inside the boundary rather than optimising up to it.

There is also a commitment problem in the good version. Once you publish a markdown ladder, you have to run it, including in the year the hall sells out early and the late release rate looks absurd. Breaking a published schedule to capture a few thousand dollars will cost you the credibility that made the schedule work, so write in a floor: the release schedule applies only to positions still unsold at each date, and the show reserves nothing.

Take last edition's contract file, filter to the value zone, and plot committed square feet against days to show open. Find the 90 day mark and read off what was still unsold. That number, multiplied by your value zone rate, is the entire pool any dynamic pricing scheme on your floor is allowed to touch, and on most shows it is smaller than the meeting assumes.