Reading your invitation statistics

A digital invitation tells you things a printed one never could: how many people have opened it, when, and roughly how they arrived. Useful, and also easy to over read.

Here is what each number is actually saying.

Opens#

The number of times your invitation has been looked at. Not the number of people, because somebody who opens it twice counts twice, which is correct: a guest reading it again in August is a real signal.

What to expect: a large spike in the first two days, roughly two thirds of your eventual total. Then a long quiet stretch. Then a second, smaller rise in the fortnight before the day, when people are checking the time and the address.

That second rise is the most underrated number on the page. It tells you the invitation is still doing its job months later, which a card in a drawer cannot.

Reach#

Roughly how many different people have seen it, counted once each per day.

The important thing about this number is that it is deliberately imprecise, and that is a feature. It is worked out from a one way fingerprint that is worthless the following day, so nothing about a guest is stored and nobody can be identified from it. It is a count, not a list.

Compare it to your guest list rather than to your opens. If you sent to ninety households and forty different people have looked, half your list has not seen it yet.

Replies#

The only number that changes a decision. Everything else is context.

Expect roughly half of your eventual replies in the first seventy two hours, then a flat middle, then a burst at your reminder. If the shape looks like that, nothing is wrong, however uncomfortable the middle feels. Setting an RSVP deadline covers where to put the reminder.

Opens against replies#

The useful comparison, and the one that tells you which problem you have.

• High opens, low replies. People have seen it and not acted. That is a wording or a friction problem. Make replying easier and more explicit
• Low opens, low replies. A delivery problem. The message did not arrive, or it arrived somewhere nobody looks. Resend through a different channel
• Opens climbing weeks later. Normal and good. Guests are checking details

Do not divide one by the other and treat it as a percentage. They are different kinds of thing, and one guest opening it four times will happily push the figure past a hundred.

Sources#

Where a visit came from. Most of your traffic will be labelled direct, and that is not a fault in the counting: a tap from inside a messaging app usually arrives with nothing attached to say where it came from.

The exception is the invitations you send by email from your own page, which are tagged, so they show up separately. That makes them the one channel you can actually measure. See how to send your invitations by email.

What you cannot know, and should be suspicious of#

Whether an email was read in somebody's inbox. Anybody claiming to tell you that is inferring it from a tracking pixel, which is blocked by most modern mail apps and is a small privacy intrusion on your guests either way.

Whether a specific person opened your invitation. Deliberately not knowable here, and it should not be. Your guests are entitled to read an invitation without being logged.

What to do with all of it#

Three decisions, and nothing else.

1. Resend to households with no activity at all, through a different channel. This catches the typo in an address and the message that landed in spam
2. Release the next wave once declines make room, if you are running one. See how many guests to invite
3. Chase the people who opened it and did not answer, individually, close to the deadline. The guests who never reply covers the wording

Then stop looking#

The numbers are a tool for three decisions and they are not a live measure of how excited everybody is. Check them twice in the first week, once at the reminder, and once before final numbers. That is genuinely all the value there is in them.