How this site works
Every amount on this site is a starting figure someone published, plus a short stack of adjustments for your situation. Each line tells you which of those two it is.
This page is the whole model — the labels, the rounding, every figure we cite, and the date we will check each one again. Read it and you can rebuild any answer on the site with a pen.
Nobody publishes a formula for what to put in a wedding card. Every other page on the subject deals with that by asserting a range and never saying where it came from, which quietly blends two very different things: figures a real publication measured, and somebody’s opinion. We keep them apart, on every line, in public.
A named publication backs this exact figure. The line says which publication, and the methodology section of the tool links it. If a source publishes a range and we land inside it, that is still sourced — the figure is one they put their name to.
Our reasoning, and nothing more. No study measured it. The line says so in plain words, and the tool’s methodology gives the exact number so you know which figure to disagree with.
There is no third label, and there are no unlabelled lines. A line that fits neither does not ship. That sounds like a small rule and it is the whole product: admitting that an adjustment is our opinion is the credibility signal, and hiding one behind an implied study is the only thing that could destroy this site.
The tag describes the figure, not the data behind it. Where we apply a real published index to a question nobody has actually studied — adjusting a gift for how expensive your state is, say — the index is cited and the line is still tagged house judgment, because the inference is ours however good the data underneath it. A sourced input does not launder a house inference into a sourced line.
This is not a mock-up. It is a live result from the wedding gift calculator, computed on this page by the same code the tool runs, so it cannot drift from what the tool would tell you.
Worked example
A close friend's wedding. You're bringing a plus-one, and you're paying to fly there.
Give $250. It starts at $200 — what The Knot suggests for a close friend — and moves from there because you’re bringing someone and you’re paying to get there.
| Why that much | Amount |
|---|---|
| Baseline for a close friendSourcedThe Knot puts this tier at $150–$250 and recommends $200. | $200 |
| You’re bringing a plus-oneHouse judgmentOne gift covers you both, so it goes up — half again rather than double, because the couple invited you, not your date. | +$100 |
| Getting there is on youHouse judgment8% of the $600 you’re spending to be there, capped at $50. This is not "cover your plate" — Emily Post rejects that idea and so do we. It is that showing up already cost you something. | −$48 |
Four things are always true of that card. The plain sentence comes first, above the arithmetic. The breakdown is visible by default and never behind a “show details” toggle. Every line carries one of the two labels. And the number at the top is an amount you could actually hand someone — 2×$100 + $50 — not a range.
Every other site in this category stops at a range. We stop at a number you can hand over, which needs one step no formula gives you for free: rounding to an increment people really give in. Below $50 a five is the smallest unit anyone thinks in; from there to $200 it is a ten; past $200 people round to quarters of a hundred.
Those bands are house judgment — nobody publishes a rounding convention for gifts. They come from how US cash denominations work, and they are stated here rather than buried so you can see the model’s last step.
| The model produces | You are told to give | Which is |
|---|---|---|
| $43 | $45 | 2×$20 + $5 |
| $47 | $45 | 2×$20 + $5 |
| $188 | $190 | $100 + $50 + 2×$20 |
| $213 | $225 | 2×$100 + $20 + $5 |
The method is version 1.0 — the shape an answer takes, described above. Each tool then has its own model version, covering the actual figures it uses. They are versioned separately because they move for different reasons: a model changes when a publication reprints a number, the method changes when we change our mind about how an answer should be built.
| Tool | Shape of its answer | Model | Reviewed |
|---|---|---|---|
| Wedding gift | One cited baseline, then signed adjustments. | v1.0 | August 2026 |
| Holiday tipping | One row per person. No single baseline. | v1.0 | August 2026 |
| Wedding vendor tips | One row per vendor. The contract decides the row’s shape. | v1.0 | August 2026 |
This is the file the site actually keeps, published rather than summarised. Only sourced figures appear here — house-judgment factors have no source to drift from, so they are reviewed when we change our minds, which is a version bump.
The last line of each card is the part we have not seen anyone else publish: not just where a number came from, but what would make it wrong, and when we expect to notice.
75 / 100 / 150 / 175 / 200 / 250
The Knot's recommended wedding gift amount by relationship to the couple: distant relative $75, coworker/acquaintance $100, extended family $150, good friend $175, close friend $200, best friend or immediate family $250
What would change it. The Knot revises this page when a new Guest Study lands, typically alongside the Real Weddings Study each winter/early spring. The recommended figures and the ranges can move independently — check both.
$150
The Knot Guest Study's average wedding gift across all guests, quoted as context in the methodology section and FAQ
What would change it. Same publication cycle as the tiers above. This figure is quoted, not computed with, so drift here is a copy fix rather than a math fix.
$130 (most guests $100–$150)
Zola's average wedding gift spend, quoted as a second published average for context
What would change it. Zola publishes a First Look Report for the coming year each autumn (the 2026 edition landed late 2025). When the next one ships, the report URL changes with the year — update both the figure and the link.
high 110 (CA 110.7, HI 110.0, NJ 108.8); low 87 (AR 86.9, MS 87.0, IA 87.8); 2024 data
BEA Regional Price Parities (all items) by state — the high and low bands used to adjust a gift for where the wedding is. US average is 100 by construction.
What would change it. BEA releases Real Personal Income and RPPs by state each winter, roughly a 14-month lag on the reference year — expect the 2025 figures around February 2027. The band values here are the rounded top and bottom of the state table, so a new release can change both the numbers and which states are named.
18 roles — see fact text for each range or ceiling
The Emily Post Institute's published holiday tipping amounts by role. Dollar ranges: day care staff $25-70 each, doorman $15-80 (at least $15 each where there are several), superintendent $20-80, garage attendant $10-30, trash/recycling collector $10-30 each, yard or garden worker $20-50 each, newspaper delivery $10-30. Pegged to what you already pay: live-in nanny or au pair up to one week's pay, regular babysitter up to one evening's pay, housekeeper up to one week's pay, dog walker up to one week's pay, pet groomer up to one session, pool cleaner up to one cleaning split among the crew, barber up to one haircut, salon staff up to one visit divided among staff, personal trainer up to one session. Package deliverer: a small gift in the $20 range, cash usually prohibited.
What would change it. Emily Post revises this guide ahead of each holiday season. Both the dollar ranges and the list of roles can change; a role appearing or disappearing matters as much as a figure moving, because each role is a ledger row. The `published` string on each role is quoted verbatim on the tool page, so a wording change is also a copy fix.
$20 per occasion / $50 per customer per calendar year / no cash ever
USPS policy on gifts to letter carriers from customers: a gift worth $20 or less per occasion, no more than $50 worth from any one customer in a calendar year, and cash and cash equivalents (checks, money orders, gift cards exchangeable for cash) may never be accepted in any amount.
What would change it. These limits derive from the Standards of Ethical Conduct for Employees of the Executive Branch (the $20/$50 thresholds are the government-wide de minimis rule), so a change would come from OGE rulemaking rather than from USPS. The verbatim quotes appear on the tool page — re-read the bulletin, not a summary of it.
20 vendors — see fact text for each range or percentage
The Knot's published tip per wedding vendor. Dollar ranges: officiant $50-100 (plus a $100-500 donation to the religious institution), ceremony musicians $15-50 each, maitre d' $200-300, cake baker $50-200, photo booth attendant $50-100, day-of coordinator $50-150, delivery and setup crew $10-50 per person, alterations $20-30, photographer $50-200 per team member, videographer $50-200 per team member, DJ $50-150, band $15-50 per member, florist $50-200, valet $1-5 per car split between attendants. Percentages: catering and bar staff 15-20% of the food and drink fee split among waitstaff and bartenders, wedding planner 15-20% of the total fee, hair 15-25% of the bill, makeup 15-25% of the bill, transport 15-20% of the total bill. Also: "tips are unnecessary if gratuities are already listed on your contract or invoice."
What would change it. The Knot revises this guide periodically and the vendor list moves as much as the figures — a vendor appearing or disappearing matters as much as a dollar change, because each vendor is a ledger row and a checkbox. The `published` string on each vendor is quoted verbatim on the tool page, so a wording change is also a copy fix. Check ahead of spring wedding season.
voluntary + customer-determined = tip; required or business-set = service charge
The IRS test separating a tip from a service charge: "A payment is a tip only if the customer voluntarily decides to pay it and determines the amount." The IRS names "required gratuities included in a contract or invoice" and "automatic charges added to a bill (for example, a mandatory 18% charge for large parties)" as service charges rather than tips — meaning they are the business's revenue and reach staff only if the business distributes them.
What would change it. A change would come from new IRS rulemaking or a superseding revenue ruling, not from the page being refreshed. The page's wording is quoted verbatim on the tool page and in its FAQ, so re-read the page itself rather than a summary of it. Note also that state wage law can be stricter than the federal test (New York's Hospitality Industry Wage Order bars calling a service charge a gratuity) — the tool does not model state law and says so.
A visitor who screenshotted a number and sent it to the person they were deciding with deserves to know if the model moved afterwards. So every change gets a dated entry, including the ones that turned out not to be changes at all.
| Date | Method | What happened |
|---|---|---|
| 21 August 2026 | v1.0 | First published, with the wedding gift calculator. Two provenance labels and no third; the itemised breakdown shown by default rather than behind a toggle; every factor adjustable by the reader; amounts rounded to something you can actually hand over. |
| 21 August 2026 | v1.0 | Extended, not changed, by the holiday tipping calculator: an answer may be a list of people rather than a baseline with corrections. No principle moved, so no version bump — recorded here because a reader who screenshotted a number deserves to see what did and did not change. |
| 22 August 2026 | v1.0 | Extended again, by the wedding vendor tip calculator: an input may change what kind of line an answer carries, not only how large it is — a vendor whose contract already includes gratuity renders at $0 and stays on the ledger. Also the first line whose reasoning cites a source and then departs from it in the open: the IRS distinction between a tip and a service charge is theirs, the decision to keep budgeting for a vague service charge is ours, and the line says so. Two labels still sufficed, so no version bump. |
Good — that is what publishing them is for. Every house-judgment figure on this site is written down as an exact number in its tool’s methodology section, so there is always a specific thing to argue with rather than a vibe. If you have a source we have missed, or a figure has moved since we last checked, tell us at hello@kindhow.com and we will re-check it and log the outcome here either way.
Informational only, not professional advice. The Kindhow Scale is a published model, not a standard. Method version 1.0. Last reviewed: August 2026.