Why an event should have one source of truth
Programme data gets rebuilt in a spreadsheet, a website, a set of speaker documents, a showflow and a script. Here is what that actually costs, measured in the only currency an event has: time on the day.
The rebuild tax
Every copy of the programme has to be built, and then rebuilt every time the programme changes. Nobody budgets for this, and everybody pays it.
The first build is visible work. Someone sits down and makes the speaker pack, and everyone can see that it took an afternoon. The rebuilds are invisible: fifteen minutes here, half an hour there, spread across four people and three weeks, none of whom are in the same meeting.
On a two-day conference with forty sessions, the programme will change somewhere between twenty and sixty times between first draft and doors. Each change costs a small multiple of five, because there are five places it has to land.
Where the copies drift
Copies do not drift evenly. They drift in a predictable order, and knowing the order tells you where the day will hurt.
The spreadsheet is always right, because it is where the change is made. The showflow is usually right, because a production person is looking at it daily and is professionally paranoid. The website is right within about a day. The speaker pack is right as of whenever it was sent, which may be three weeks ago. The script is right if someone remembered it exists.
So the failure surfaces in a specific place: the speaker who was briefed correctly for an event that no longer exists, standing in the wrong room at the right time.
Nobody is doing it wrong
It is worth being clear about this, because the instinct is to treat drift as a discipline problem and solve it with a stricter process. That does not work, and it is unfair to the people it is aimed at.
Every copy exists for a legitimate reason and is maintained by someone competent. The drift is structural. Ask five people to keep five documents in step by hand, through sixty changes, while also delivering an event, and they will fail — not because they are careless but because you have asked for something that does not survive contact with a live week.
What a single record changes
Hold the event once and generate the views, and the rebuild tax goes to zero — not because people work faster but because the work no longer exists.
A session has a time, a room, a set of speakers and a duration. The portal renders that for attendees. The speaker page renders the parts that speaker needs. The showflow renders it as cues with department tags. The signage renders the next three. Change the session and all four are already correct.
One write
The change is made once, in the place it was decided.
Many reads
Every surface renders the same record, shaped for its audience.
No reconciliation
There is no second copy to compare against, so nothing can be out of step.
The reasonable objections
Two come up every time, and both are fair.
"The public site needs different words to the showflow." Yes — and that is a view, not a copy. The portal can carry marketing copy, images and a friendlier session title while still reading its time and room from the same record. A view is allowed to show less, or to show it differently. It is not allowed to hold its own version of the facts.
"We need to publish a version and then keep planning." Also yes. That is what a published state is for: the programme keeps moving privately, and you choose when the public view catches up. The point is that catching up is a decision, not a re-typing exercise.

Plan the event. Publish the programme. Run the show.
One event spine, many views. Build the programme once, then give every person the right live view of it.