What does a single guest cost us?

That distinction is the whole article. Use the wrong one and you will make the wrong call — believing that cutting ten guests saves $2,900 when it saves $1,400, or assuming that late declines reduce your catering bill when the food and beverage minimum means they do not.

The reference figures are from The Knot Real Weddings Study 2026 for weddings held in 2025, with catering benchmarks from the same source. All amounts are US dollars.

The marginal number is what actually changes when one guest is added or removed: their meal, their share of the bar, the place setting, the chair, the invitation, the escort card, the favor. In the US in 2025 this typically ran $110 to $180.

The average number is the total budget divided by heads: $292 nationally. Inside that figure sit the dress, the rings, the photographer and the venue fee — none of which change when someone cannot come.

Both are correct. Only the first one is any use for deciding whether to invite your colleagues.

The mid-range column lines up with the national catering benchmark of $80 per head all-in, plus the items caterers do not include. Regionally that food figure runs from $62 in the Midwest to $123 in the Mid-Atlantic, so shift the whole column accordingly.

What moves a guest from the left column to the right is four decisions rather than the location. First, the service format: a plated dinner typically runs $15 to $40 per head above a buffet, because it binds more staff for longer. Second, the length of bar service: each additional hour on a per-person package commonly adds $8 to $15 per guest. Third, what the bar includes — a beer, wine and one signature cocktail bar sits well below a full premium bar for the same crowd. Fourth, the number of separate eating moments: passed appetizers, a late-night snack and dessert beyond the cake are three individually priced items rather than parts of the dinner price. None of the four is a matter of quality. All four are a matter of format, which is why they are the levers to pull first.

Three groups are routinely mis-costed:

• Children. Most caterers charge a reduced rate, commonly 50 to 60% of the adult price, and no bar. A dozen children can shift a catering total by several hundred dollars in your favor — but only if you tell the caterer.
• Vendor meals. Photographer, videographer, DJ, planner and their assistants all eat. Vendors are typically billed at 50 to 75% of the guest rate. At four vendors and a $70 plate that is $140 to $210 that appears in no quote you were shown.
• Plus-ones. They enter planning as a courtesy and leave as a full-price guest.

Ask the caterer how vendor meals are billed before you sign. It is the single most common reason a final catering invoice exceeds the headcount you planned for.

The guaranteed count: when a decline stops saving you money

Your marginal cost per guest is only marginal for as long as the headcount can still move. Caterers and venues manage this with a guaranteed count: a binding number you submit a set number of days before the event, which is then billed regardless of who actually walks through the door. In the US the standard window is 7 to 10 days, and some venues ask for 14.

Most contracts also cap how far the number may fall. A common term allows a reduction of no more than 5 to 10% below the count in the contract. Book for 120, stand at 98 acceptances ten days out, and you may still be billed for 108. At $140 all-in per head that is $1,400 for ten place settings nobody uses.

Three questions therefore belong in the contract conversation, and the answers live in the contract rather than the proposal. When is the final count due? How far below the contracted number may it go? And are late additions accepted after the deadline, at what price? The last one is regularly overlooked and typically carries a surcharge per person added — which is the mirror image of the first two and just as expensive.

Fixed costs explain all of it. A $3,000 photographer is $83 per head at 36 guests and $23 per head at 130. So a small wedding lowers the total and raises the per-head figure at the same time — worth knowing before you promise yourselves that going small will make things cheap per person.

One consequence is worth stating plainly: a small wedding is cheaper in total and more expensive per head, and both statements are true at once. If someone tells you a small wedding is not worth it because the per-guest figure is so high, they are quoting the number that does not govern any decision you are making.

You need exactly one figure. Here is how to get it:

1. Take the per-person food price and the per-person bar price from the venue or caterer proposal.
2. Add the service charge percentage to both.
3. Add rentals, cake, stationery and favor — usually $12 to $50.

That sum is your number. Multiply every guest list change by it.

Where it stops working: the minimum

Most full-service venues impose a food and beverage minimum — a contractual floor on catering and bar spend regardless of headcount, normally calculated before tax and service charge. Below that floor, declines save you nothing at all: you pay the minimum either way.

So work out the guest count at which you reach your minimum. That number, not the capacity written in the contract, is your real lower bound. And ask specifically whether bar, rentals and room fee count toward the minimum — at a venue that credits food only, the headcount you need is considerably higher than it first appears.

One more thing worth budgeting for: at roughly 20 to 30 guests per server for a buffet, the twenty-first guest can trigger an entire additional shift.

Two numbers for two points in time

Because of that deadline, your marginal cost means two different things across the planning period. Before the guarantee date, $140 per guest is a genuine lever: every decline saves it, every acceptance costs it. After the guarantee date, the same $140 has become a fixed cost — the decline saves nothing, while the acceptance still costs.

The practical consequence is that the guarantee deadline, not the date printed on the invitation, is the real end of your guest list. Set the RSVP date well ahead of it: two to three weeks is a workable buffer, because a predictable share of replies arrives after the deadline and has to be chased by phone.

A second effect applies to the bar. Under a per-person bar package the same mechanic holds — the number is locked with the guarantee. Under consumption billing, the bar cost keeps falling when guests drop out late, because you pay for what is poured. Which of the two your contract specifies therefore determines how expensive a late decline actually is, and it is worth knowing before you set the RSVP date rather than after.