The short answer: fewer ads than you think, earlier than you think
A typical event in this data started advertising around a month before the date, used roughly seven different ads across the whole run, and had about two of them live on a normal day. Busy days peaked around five. Nobody was running twenty ads at once, and the events that sold well were not the ones with the biggest ad library.
So a sensible starting plan for a single-date show looks like this: go live about thirty days out, keep two or three ideas running at a time, and have six to eight ready to rotate in as the weeks pass. Check in at fourteen days, seven days and three days before the door opens. That is a plan to test against your own event, not a law of physics - festivals and last-minute parties behave differently.
Waiting for the last two weeks costs you money
The most common instinct is to save the budget for the final sprint. The numbers do not support it. Across all the events we looked at, only about four in ten purchases happened in the last two weeks. Half arrived earlier than that, and the rest landed on the day itself. The stretch between one month and two weeks out was quietly one of the most efficient of the whole campaign.
What each stretch of the countdown tends to give back
| When | Back per euro | What we read into it |
|---|---|---|
| More than a month out | 2.0× - 2.9× | Early interest, usually cheap to reach |
| 30 to 15 days out | 3.5× | Often the quiet workhorse of the campaign |
| 14 to 8 days out | 3.1× | Closer does not automatically mean cheaper |
| 7 to 4 days out | 3.1× | Steady, rarely spectacular on its own |
| Final 3 days | 4.5× | Strongest pre-event stretch we see |
| Event day | Very high | Mostly demand already created earlier |
The final three days are real - the two weeks before them are not magic
There is a genuine last-minute effect, but it is narrower than the folklore suggests. The final three days were the strongest pre-event stretch, at around four and a half euros back for every euro spent. The two weeks leading into them were not better than the month before - in several cases they were worse.
We also compared the same events with themselves: the fortnight from four weeks out against the fortnight right before the date. Ten events improved in the final stretch and nine got worse. Spending rose by about seventy percent, while purchases rose by about a third. More money later bought more tickets, but each one cost noticeably more.
More ads live ≠ better results
How much came back, grouped by how many ads were running that day
Notice the dip in the middle. Days with three or four ads live did worse than days with one or two. The busiest days look good, but they come from a small handful of unusually strong events, so we would not read them as a recipe. The honest summary is that ad count is not the lever - what the ads say, and who sees them, is.
What this means for your next event
1. Start about a month out, and mean it. Early weeks are usually the calmest and cheapest place to gather buyers, and they also give the platform time to learn who your audience is. If you switch everything on ten days before the show, you are paying for that learning during the most expensive part of the campaign.
2. Keep the ad pool small and refresh it. Two or three ideas live at a time is plenty. Prepare more, but bring them in gradually as older ones tire. Piling on ads mostly splits the same budget into smaller pieces and makes it harder to tell which message is working.
3. Hold something back for the last three days - not everything. There is a real spike at the end, and you want room to lean into it. But it is a top-up on demand you built earlier, not a replacement for it. An event that arrives at the final weekend with no warm audience has nothing to push on.
4. Judge the budget by results, not by the countdown. Spending more simply because the date is close is the fastest way to raise your cost per ticket. Check in roughly two weeks, one week and three days out, and let what the campaign is actually returning decide how hard you push.
Based on our own advertising records for Romanian music events and festivals between late 2023 and the end of 2025, rounded and anonymised. Purchases are checkout events recorded by the advertising platform, so one purchase can include several tickets, and returns are platform figures rather than a ticketing ledger. These are observed patterns across many music events, not a controlled experiment, and they should not be read across to conferences or other business events.