Working out the true cost per hosted buyer before you scale it
Cost per hosted buyer divides the programme's total spend by the buyers it hosted, but the figure that matters divides the same spend by the buyers who completed their meeting commitment. Twenty per cent failing to complete turns a headline unit cost of 1,400 into a real cost of 1,750.
The board pack says the hosted buyer programme costs 1,400 per buyer and the recommendation is to grow it from 200 buyers to 300. The arithmetic on the slide is 280,000 over 200, and it is correct.
The cost per hosted buyer that matters uses a different denominator, and on most programmes it lands 20 to 30 per cent higher than the number on the slide. The reason is that some of those 200 buyers arrived and did nothing. They flew, they slept in the hotel you paid for, they collected a badge, and they held two of their twelve meetings. Their cost is in the numerator and their output is not in anything.
What actually goes into the numerator
Six lines, and the one that gets left out is always the same one.
Travel is the visible cost, and its shape varies more than people expect. Business Travel Show America's 2026 hosted buyer travel policy offers fully hosted buyers up to 400 US dollars of transportation reimbursement, applicable to airfare, train, parking or ground transportation, and a standard room at a four star hotel in Times Square for two nights, room and tax only. That is a capped contribution model. Other programmes book the flights themselves and carry the full fare, which produces a very different cost per buyer for the same experience.
Accommodation is usually two or three nights at whatever the show week rate is, which is not the rate anybody quoted in March.
Transfers, catering and lounge access are small individually and add up to something real across 200 people.
Concierge staffing is the line that goes missing. Qualifying applications, chasing profiles, building schedules, handling changes on site. If two people work the programme for three months and four people work it during the show, that is a meaningful share of the total and it is usually sitting in a departmental salary line where nobody attributes it to the programme.
Platform cost, meaning the matchmaking and scheduling system, sits in the same place and has the same problem.
Take a worked total. Two hundred buyers, 700 travel contribution each, 300 of hotel, 60 of transfers and catering, which is 1,060 of direct cost, or 212,000. Add 55,000 of staffing and 13,000 of platform, and the programme costs 280,000. Divide by 200 and the headline unit cost is 1,400, of which 24.3 per cent is not travel or hotel at all.
Which denominator should you divide by?
Buyers hosted is what you paid for. It is the wrong denominator for every decision you are going to make with this number.
Define completion first, and define it as the programme's own commitment. If the terms say a buyer attends 12 meetings, then a buyer who attended 12 completed. A buyer who attended 9 did not. Partial credit is defensible too, and if you use it, say so.
Now suppose 20 per cent of your 200 buyers fail to complete. That is 40 buyers, and 160 completed. The cost per completing buyer is 280,000 over 160, which is 1,750. The 350 difference between 1,400 and 1,750 is the cost of non-completion, spread across the buyers who did complete, and it is 25 per cent on top of the headline.
Take it one step further, to cost per meeting held. If 160 buyers held their full 12 and the 40 non-completers held an average of 4, meetings held is 1,920 plus 160, which is 2,080. Cost per meeting held is 280,000 over 2,080, which is 134.62. Compare that with the planning figure of 280,000 over 2,400 scheduled meetings, which is 116.67. The gap between 116.67 and 134.62 is what non-completion costs you per meeting, and it is 15.4 per cent.
Those three numbers, 1,400 and 1,750 and 134.62, should all appear in the same table. Reporting only the first is the version of this that gets a programme scaled into a problem.
There is a fourth denominator that some programmes reach for and I would avoid: cost per meeting scheduled. It looks like a productivity measure and it is trivially gamed, because scheduling more meetings per buyer lowers it without anybody attending anything. A programme that raises its quota from 12 to 16 improves cost per scheduled meeting by 25 per cent on the same spend and will usually see completion fall, so the measure moves in the opposite direction to the outcome. Use scheduled meetings for capacity planning and held meetings for cost.
Does scaling the programme improve the number?
Partly, and it is worth separating which parts move.
Travel and hotel are variable, so they do not improve with scale at all. Six hundred of hotel for two buyers is exactly twice 300.
Staffing is partly fixed. The person who designs the qualification process does that once whether you host 200 or 300, so the 55,000 might become 68,000 for 300 buyers rather than 82,500. That is a saving of 14,500, or 48 per buyer.
Platform cost is mostly fixed, so 13,000 covers 300 buyers as easily as 200, saving another 22 per buyer.
So scaling from 200 to 300 takes the direct cost of 1,060 and adds staffing at 227 and platform at 43, giving 1,330 against the original 1,400. A 50 per cent increase in buyers bought you a 5 per cent reduction in unit cost, and that is the honest ceiling on the efficiency argument.
The completion rate is the lever that actually moves the number. Moving completion from 80 to 90 per cent takes the cost per completing buyer from 1,750 to 1,556, an 11.1 per cent improvement, without hosting a single extra person or spending anything. That is why the measurement of who actually turned up is worth more attention than the procurement conversation about airfares.
The paid tier changes the denominator, not the cost
Several organisers have moved part of the hosted population into a paid buyer tier, and the accounting needs care.
IBTM World announced in July 2025 that its 2025 edition would offer visitor buyer tickets priced at 399 euros, rising to 499 euros after 16 September, alongside its traditional hosted programme. Visitor buyers choose how many meetings they schedule and book directly with exhibitors, and they receive lounge access, daily lunch vouchers and the education programme. The hosted programme continued, with a personalised appointment diary and a contribution towards flights, transfers and a four or five star hotel stay in exchange for meeting commitments.
That structure does two things to your unit cost. It removes buyers from the hosted denominator, so the programme gets smaller and its unit cost stays roughly where it was. And it introduces a second population whose cost is concierge time and platform capacity, with revenue attached instead of spend.
The mistake to avoid is blending them. A combined figure of total buyer cost over total buyers falls every time you sell a paid ticket, which makes the programme look more efficient while nothing about the hosted operation changed. Report hosted cost per hosted buyer, and report paid tier yield separately, and let the two numbers move independently.
There is a real analytical prize in this structure, though it needs an edition or two of data. Paid buyers chose to be there and paid for the privilege, so their completion rate is the natural control group for your hosted buyers' completion rate. If paid buyers complete at 91 per cent and hosted buyers at 78, the 13 point gap is telling you something specific about the incentives in your hosted terms, and the cancellation policy is where that gets addressed.
Where this stops
Cost per hosted buyer is an input cost with no view on value, and it can be improved by making the programme worse.
Cut the hotel to three star and the number falls. Cut the concierge team and it falls further, right up until schedules stop getting built and completion collapses. Any unit cost that can be improved by degrading the thing it measures has to be read next to an output measure, and the output measure here is meetings held and what came of them.
The second limit is attribution. A hosted buyer who holds 12 meetings and signs nothing is expensive. One who holds 3 and places an order worth six figures paid for the whole programme. Neither outcome is visible in a cost per buyer figure, and the honest position is that this metric governs efficiency and somebody else's metric has to govern effectiveness.
The third is that the denominator is a policy choice you can quietly move. Define completion as 8 meetings instead of 12 and your completion rate jumps, your cost per completing buyer falls, and nothing changed. Freeze the definition, publish it next to the figure, and treat any change to it as a restatement of the whole series rather than an improvement.
Start by taking last edition's total programme spend, including the staffing line nobody attributes, and dividing it three ways: by buyers hosted, by buyers who met the commitment, and by meetings actually held. If the first and second differ by more than about 15 per cent, non-completion is the largest single cost in your matchmaking programme and no amount of negotiating on hotel rates will touch it.
Questions people ask about cost per hosted buyer
- What goes into the cost of a hosted buyer?
- Flights or a travel contribution, hotel nights, airport transfers, on-site catering and lounge access, the concierge staffing that qualifies and schedules them, and the platform cost of running the meeting programme. Staffing is the line most often left out, and on a well-run programme it is a substantial share of the total.
- Should cost per hosted buyer use buyers hosted or buyers who completed?
- Report both and manage on the second. Buyers hosted is what you spent money on. Buyers who completed their meeting commitment is what you received. The gap between the two figures is the cost of non-completion, and it is invisible in the headline number that most programmes report to their boards.
- How does a paid buyer tier change the cost calculation?
- It moves buyers out of the hosted denominator entirely and adds ticket revenue. A buyer paying their own way costs the programme concierge time and platform capacity but no travel spend, so blending them into one average unit cost hides both the hosted spend and the yield from the paid tier.
Related reading
- Measuring the meeting no show rate without flattering yourself
- A meeting cancellation policy that changes behaviour before the show opens