# Thread: % of people who cancel from attending an event

1. ## % of people who cancel from attending an event

Hi all,

So I am trying to calculate the % of the audience that cancelled at an event a year ago to predict how many will cancel this year. I have two formulas and am not sure which one is most accurate.

I had 696 total people registered to attend and 161 of them cancelled. Is the proper way to calculate the % of the audience that cancelled to divide 161 from the 696 total? Or do I subtract 161 from 696 and then divide 161 from 696 (535)?

Thanks in advance to anyone who can help with this!

2. ## Re: % of people who cancel from attending an event

Originally Posted by DjJazzyDeath
Hi all,

So I am trying to calculate the % of the audience that cancelled at an event a year ago to predict how many will cancel this year. I have two formulas and am not sure which one is most accurate.

I had 696 total people registered to attend and 161 of them cancelled. Is the proper way to calculate the % of the audience that cancelled to divide 161 from the 696 total? Or do I subtract 161 from 696 and then divide 161 from 696 (535)?
The percentage of people who canceled their participation compared to total people is 161/696.
The percentage of people who canceled their participation compared to people who didn't cancel is 161/(696-161).

I guess you should be interested in the first value of 161/696 ~= 23.1 %

3. ## Re: % of people who cancel from attending an event

Yes, I have been going off 23% for making predictions for the next year's event. But here's my problem - after 161 cancelled their seats I was left with 535 registered and then only 410 showed up. So we could say that 23% also "No Showed". So if I want to predict how many "confirmed" people I need after 23% cancel to reach 500 attendees I need to:

Find what 23% of 500 is. (115)

This tells me that after 23% of my audience Cancel, I need 615 still "confirmed" to reach my 500 goal because another 23% are going to "No Show". So does it make sense to apply 23% to 615?

23% of 615 is 141.25 (round up to 142 since we're talkin' about people here)

615 + 142 = 757 for my total audience in order to achieve 500 attendees

Seems like it makes sense, right? My issue is if this is correct then this tells me I need to grow my audience YoY. Shouldn't that mean I will have a higher number of total cancellations and no shows YoY because I'm dealing with the same percentages but a larger total audience?

2012 No Shows = 125

2013 Predicted No Shows = 115

2012 Cancellations = 161

2013 Predicted Cancellations = 142

This is weird, right?

4. ## Re: % of people who cancel from attending an event

Originally Posted by DjJazzyDeath
Yes, I have been going off 23% for making predictions for the next year's event. But here's my problem - after 161 cancelled their seats I was left with 535 registered and then only 410 showed up. So we could say that 23% also "No Showed". So if I want to predict how many "confirmed" people I need after 23% cancel to reach 500 attendees I need to:

Find what 23% of 500 is. (115)

This tells me that after 23% of my audience Cancel, I need 615 still "confirmed" to reach my 500 goal because another 23% are going to "No Show". So does it make sense to apply 23% to 615?

23% of 615 is 141.25 (round up to 142 since we're talkin' about people here)

615 + 142 = 757 for my total audience in order to achieve 500 attendees

Seems like it makes sense, right? My issue is if this is correct then this tells me I need to grow my audience YoY. Shouldn't that mean I will have a higher number of total cancellations and no shows YoY because I'm dealing with the same percentages but a larger total audience?

2012 No Shows = 125

2013 Predicted No Shows = 115

2012 Cancellations = 161

2013 Predicted Cancellations = 142

This is weird, right?
You made the whole problem a lot harder for yourself than need be. All you really care about is the proportion of people who actually turn up, compared to those you had initially registered. Since last year you had 696 people registered, and, after cancellations and "no shows" only 410 showed up, you can deduce that you have a (410/696)=0.589... registration to showing up turn-over rate. So if you actually want to have 500 people show up on the day, your best guess (which won't be that accurate since you have only 1 years data to go by, but it is a decently large sample nonetheless) is that you need 500/0.589...=848.78... (849) people registered to come to the event.

You can easily check that your calculation was incorrect from observing that despite that fact you want 90 more people to turn up (500-410), your calculation only predicts you need 61 more people registered than last year (757-696) which clearly doesn't make sense! Hope I've been of some help!