The Seminar Confirmation Sequence: How 106 Registrations Became 59 Households at Dinner

In my practice's last two-night campaign, 106 households registered, 64 confirmed, and 59 sat down to dinner. Those three numbers tell you the most important thing about seminar attendance: registration is a maybe, confirmation is a plan, and the work that happens between the two is what decides how full your room is. This article is that work, step by step.

Registration is not a commitment

Someone saw a mailer or an ad, clicked, and typed their name three weeks before your event. That cost them nothing and committed them to nothing. Treating that registration as a reservation is the single most common cause of the half-empty rooms that make advisors quit this channel.

The 106-to-64 drop in my funnel is not a failure of the sequence. It is the sequence doing its job: sorting the definitely-comings from the probably-nots while there is still time to act on the difference. The alternative is not that all 106 show up. The alternative is finding out who was real on dinner night, when the information is worthless. I wrote about how this sorting interacts with show rate math in what a good show rate actually is.

Confirmation is bilateral

The core principle of the whole sequence: a confirmation only exists when the household does something. A delivered reminder is not a confirmation. A reply, a tapped link, a yes on the phone: those are confirmations. You are not broadcasting at people, you are collecting commitments, and the collection requires an ask they can respond to.

This is also why the sequence needs to run over channels that can receive replies. A text from a number that cannot take a response is a flyer, not a conversation.

The sequence

Instantly at registration: welcome text and email. Confirms their spot exists, states the date, time, and venue, and sets the expectation that you will be in touch. This message does double duty as a data check: bounces and failed sends tell you immediately which registrations came in with junk contact info.

Seven days out: the confirmation ask. The pivotal send. A short text asking them to reply to confirm they are coming. Everyone who replies moves to confirmed, and everything downstream treats confirmed and unconfirmed households differently. The exact copy and timing patterns are in the companion piece, seminar reminder text templates.

Three to four days out: pursue the silent ones. The unconfirmed households get a second ask, and the highest-value ones are worth a live phone call. A two-minute call recovers a surprising number of households who meant to reply and did not, and it surfaces the polite cancellations while their seats can still be given away.

Day before: reminder to the confirmed. Different message for a different audience. The confirmed are coming; this send removes friction. Venue, time, parking, who to ask for. No selling, no confirm ask, just logistics.

Day of: the short one. Time, address, see you tonight. Sent early enough to be useful.

What the confirmed count is for

By three days out, the confirmed number is the only attendance number I look at, and it runs the real decisions:

• The restaurant guarantee. You commit to a headcount based on confirmations plus a modest allowance, not on registrations. This is direct money: the gap between guaranteeing for 106 and guaranteeing for the real room is a large bar tab you did not owe.
• Seat reallocation. Every household the pursuit calls surface as not-coming is a seat you can release, which matters when a campaign overfills.
• The registration review. Before the confirmation asks fire, the registration list gets a review pass, which is where repeat free-dinner collectors get caught. That workflow is the plate licker check, and it belongs before confirmations, not after, because you should not be collecting commitments from people you do not want to seat.

What it produced

The funnel from the May campaign: 106 registered, 64 confirmed after the full sequence, 59 households at dinner across the two nights. Attendance landed just under the confirmed count, which is what the sequence is supposed to produce: a confirmed number solid enough to plan a room around, with the drift measured in single digits instead of dozens.

Run the same three numbers on your own last campaign. If you cannot, that is worth sitting with. If the gap between your registered and attended counts arrived as a dinner-night surprise, the sequence above is the fix, and it is almost entirely automatable.