The date that decides everything,
is not your wedding day.

The wedding day is the date you know by heart, which is why every timeline gets written from it. But almost nothing in your invitation schedule is driven by it. The date that costs you money if you miss it sits in the catering contract: most venues and caterers ask for a final, chargeable number somewhere between a fortnight and a few days before the day. After that point, a guest who replies late is either a favour someone does you or a head you pay for. Find that date before you decide anything else; it is usually a line in a contract rather than something anyone says out loud.

Write it down and treat it as the fixed point. Everything else is built backwards: the RSVP deadline sits a few weeks before it, the invitations go out far enough ahead that people can answer, and any printing and posting happens before that. Work in the other direction, picking a nice date to send and hoping, and you find out in the last fortnight that eleven people have not replied and you have two days to reach them.

Two or three neighbouring dates in the same contract behave the same way: the seating plan the venue wants, the meal choices the kitchen wants, sometimes a separate dietary list. They are rarely the same day. Put them all in one calendar rather than in three old emails, and the collisions turn up while you can still move something.

For a wedding where most guests live within an hour or two, invitations usually go out six to eight weeks before the day. That is long enough for people to arrange a babysitter and reply, and short enough that the date still feels like a decision. Send eight months ahead and you get a warm message saying of course we are coming, and no actual answer until you ask again.

Four things push the window earlier. Guests who have to travel are booking transport and a night or two of accommodation, so three months is a fairer minimum and four is kinder. A date sitting next to a public holiday, or inside a school break in the countries your guests live in, competes with plans people make early. A destination wedding is a holiday for everyone attending: three to four months at the very least, and it wants a save the date long before that. And a short engagement collapses the whole thing. If you are marrying in ten weeks, the invitation goes out this week and the wording acknowledges it.

What does not move the window is how formal the wedding is. Sending a year ahead does not read as elegance, it reads as a date nobody has to think about yet. If you want people to hold the date early, that is what a save the date is for.

Put the reply deadline two to three weeks before final numbers are due. Not the day before, and not the same week. Those weeks are not slack, they are working time: chasing the twelve people who have said nothing, waiting for them to answer, reconciling the list and getting meal choices to the kitchen. Everyone who has set a deadline three days before final numbers has spent an evening on the phone.

There is a limit at the other end. A deadline months out gets ignored, because it is attached to nothing. A guest reading "please reply by 3 March", for a wedding in September, files it under later and means it. Aim for a date roughly four to six weeks before the wedding: close enough to feel real, far enough to leave you room. If that lands awkwardly near the send, the invitations are going out too late; the deadline is not the thing to move.

Say the date plainly on the invitation, and say it again where the guest actually replies. "Please let us know by 12 April" beats "RSVP by return of post", which describes a mechanism nobody uses. Our guide to the RSVP deadline goes through choosing the date and wording it.

A save the date changes what the invitation is for. The news has been delivered, the date is in people's calendars, and nobody is finding out for the first time. The invitation now carries the detail: times, place, how to get there, what to wear, how to reply. It can therefore arrive later than it otherwise would, because its job is confirmation rather than surprise. Six weeks is comfortable when a save the date went out eight months earlier.

The trap is the gap. A save the date followed by nine months of silence makes people quietly wonder whether they are still invited, especially anyone who was told in conversation rather than sent anything. If the gap is long, one short message in between keeps it alive without pretending to be the invitation: the hotel rooms you have held, or the ceremony time now that it is fixed.

One rule for the save the date itself, which our save the date guide goes into properly: it should contain nothing that can change. A date and a town. The moment it carries a start time, a venue you have not signed, or a hotel you might not end up using, you have printed something you will later have to correct in public.

In some places it is ordinary to invite one group to the whole day and another to the evening; in others the idea does not exist at all. Where it does, the evening invitation has to read as an invitation to that evening, not as a reduced version of somebody else's card. Same design, same wording, same warmth, different start time, and nothing on it that implies a longer day is happening down the road.

The second round, the people you invite once declines come in, is harder, and the giveaway is always timing. If the first batch went out in March and a second arrives in June, someone will do the arithmetic out loud at a family lunch. Two things keep it invisible: send the second batch within a few weeks of the first, and never print a date on the invitation itself. Decide in advance how many replacements you can absorb, because a batch that goes out after the reply deadline is not a second round, it is an apology.

If your day genuinely has parts, a ceremony some come to, a dinner for fewer, an after-party for whoever is still standing, it is cleaner to ask about the parts than to run two guest lists. In Vowly Event a guest can be asked to reply separately for each part of the programme, and a part can be invisible to the guests who were not invited to it, so nobody reads about a dinner they are not at.

The step nobody plans for eats the most time: you do not have everyone's address. Collecting forty postal addresses takes a fortnight of messages, and it happens after the cards are printed and before they can be posted. If you are posting, start collecting the day you order the printing. If you are sending a link, a phone number or an email will do, which most people can assemble in an afternoon.

One invitation goes to a household, but the count is people, not envelopes. That is where lists go wrong: an envelope addressed to two parents and three teenagers is one invitation and five heads, and five is what the caterer bills. Be explicit on the card about who is invited, naming the children or not, because the alternative is an awkward conversation later about a guest you never counted. The reply form asks for party size and companion names, so what comes back is a name for each seat rather than a number you have to interpret. If your list already lives in a spreadsheet, it imports from .xlsx or .csv with column mapping and duplicate detection, so the households survive the move.

Some guests need paper and always will. Send them a card. A printed card can carry a QR code that opens the same invitation page as everyone else's link, so an aunt with a phone gets the directions and the hotel list, and one without a phone gets a card and a number to ring. Both count into the same list, and you are not keeping a second spreadsheet for the people who are not online.

Posting costs days you have to pay for in advance. Design and proofing take a week or two if nothing goes wrong, printing another one or two, addressing and stuffing an evening you will not enjoy, and the post a few days domestically or a couple of weeks across borders. For cards to land eight weeks before the day, the order goes in around twelve. That is not an argument against paper, just the arithmetic, and it is why "we will do the invitations in the spring" turns into a rush.

A link costs nothing in days. You finish it and it is with everyone that evening, which is what makes it the answer to a short engagement or a guest list that grew. The bigger difference is that a link can be corrected after it has gone out: move the ceremony half an hour, swap the hotel recommendation, fix a misspelled surname, and every guest who opens the page, including the ones who opened it last month, sees the current version. A card that has been posted is finished. The only remedy is a second card, and half the recipients will remember the first one.

Most people end up doing both, which is fine as long as the two agree. Send the link to everyone who will use it, post a card to the ones who will not, keep the same information on both, and put the QR code on the card so the paper guests can reach the live page too. What does not work is posting cards and treating the page as decoration: the moment the two disagree, guests believe the paper. Our wedding invitations page shows what the live version looks like from the guest's side.

A share of your list will not answer, and it is never the people you expect. It is rarely rudeness: the invitation arrived on a Tuesday, they meant to reply at the weekend, and it left their head. Chasing is normal and it works, best when it is early and individual: a message to one person by name, with the date in it. Our article on chasing wedding RSVPs covers the timing and the wording, including what to send to the last few who still say nothing.

Keep one date in reserve that only you know about: your own cut-off, a few days after the published deadline, at which a non-reply becomes a no. Say that plainly to the last few and most of the remaining answers arrive within a day. Then hand the number to the kitchen and stop counting. A guest who surfaces after final numbers is a nice problem and often a solvable one; a number handed over late is neither.