ROI on Postcard Mailing

[Newby]

Re: Has Anyone Ever Figured....

Anyone who is interested in doing post cards should see the ones that Newby is using.

Xnay on the ard-ca! Those are not for public consumption.

 

[xrac]

Re: Has Anyone Ever Figured....

Xnay on the ard-ca! Those are not for public consumption.

Yes! Master Yoda Newby!

 

[padthaiforlunch]

Re: Has Anyone Ever Figured....

owhay anymay erehay onknay igpay atinlay?

 

[BNTRS]

Re: Has Anyone Ever Figured....

Nope, you are wrong. 1% response means 1 response per 100 mailed. I also took a couple of stat classes at NC State many years ago. We're not talking about the probability of a 1% response, we're talking about absolutes.

What do you have a mouse in your pocket or something? I'm reading questions about what people will expect to receive back for a response from sending out a number of mailers. How can this not be a probability consideration?

I'll assume those were intro stat courses at NC state before I go making unkind comments about how badly they under-served you with respect to imparting an understanding of stats and probability.

In an assumed standard normal distribution (and that's what we are assuming in most mailing campaigns). We're simply solving the cumulative probability function given the parameters of a 1% success rate and 100 trials. Mathematically this comes out to .63 look at that I was wrong the probability of at least 1 is 63%.

Now, there is a slight omission of information I made that you could have jumped on me about, if your knowledge of stats was really as you attempted to suggest. 50% probability of at least 1 would have been the more appropriate way of putting it. Of course, you should have been able to infer what I was talking about when I mentioned binomial distribution solve.

All of this is said to explain why sometimes you could mail 100 pieces and get no responses. Some people fault the response rate (and this could be the case) but response rates might not be as low as we sometimes assume, it's a probability issue.

The 50% probability you mention above, for a sample size of 100 would be a different value if 100,000 were mailed.

So you do sort of understand what I'm talking about...maybe.

I believe you are saying there's a 50% chance that out of a 100 pc. mailing that 50% of the time one is mailed back and 50% of the time zero is mailed back. If that is what you are saying then you've already used up the whole universe pie, meaning there's a zero chance of 2 or more being mailed back.

Wrong. I'm saying there is a 50% (actually 63% after I actually did the math) chance of at least 1. We get this by solving the probability density function of exactly 1 response given the 100 trials and .01 response rate, (which is 36% for those who are interested) and we subtract this probability from 1 to get the probability of getting at least 1.

Remember, 1% means one out of one hundred. Nice try though.

But 1% response rate and 1% probability do not mean the same thing and a 1% response rate does not guarantee that you will get 1 out of 100 once probability is introduced to the situation, which is vital to any useful application of this conversation.

Simply telling someone a 1% response rate means you'll get response out of 100 back is technically correct from a very literal stand point, but again it's of extremely limited applicable use when actually performing a mailing campaign.

 

[Newby]

You guys are making my brain hurt.

 

[arnguy]

You guys are making my brain hurt.

x2.....I hated the statistics courses I had to take. However, I enjoyed calculus.

 

[BNTRS]

x2.....I hated the statistics courses I had to take. However, I enjoyed calculus.

If you loved calculus this should right up your alley as a solve is simply the integral of the p dist curve given the parameters.

 

[tins]

Re: Has Anyone Ever Figured....

owhay anymay erehay onknay igpay atinlay?

I don't

............

 

[theinsuranceman]

Re: Has Anyone Ever Figured....

What do you have a mouse in your pocket or something? I'm reading questions about what people will expect to receive back for a response from sending out a number of mailers. How can this not be a probability consideration?

I'll assume those were intro stat courses at NC state before I go making unkind comments about how badly they under-served you with respect to imparting an understanding of stats and probability.

In an assumed standard normal distribution (and that's what we are assuming in most mailing campaigns). We're simply solving the cumulative probability function given the parameters of a 1% success rate and 100 trials. Mathematically this comes out to .63 look at that I was wrong the probability of at least 1 is 63%.

Now, there is a slight omission of information I made that you could have jumped on me about, if your knowledge of stats was really as you attempted to suggest. 50% probability of at least 1 would have been the more appropriate way of putting it. Of course, you should have been able to infer what I was talking about when I mentioned binomial distribution solve.

All of this is said to explain why sometimes you could mail 100 pieces and get no responses. Some people fault the response rate (and this could be the case) but response rates might not be as low as we sometimes assume, it's a probability issue.



So you do sort of understand what I'm talking about...maybe.



Wrong. I'm saying there is a 50% (actually 63% after I actually did the math) chance of at least 1. We get this by solving the probability density function of exactly 1 response given the 100 trials and .01 response rate, (which is 36% for those who are interested) and we subtract this probability from 1 to get the probability of getting at least 1.



But 1% response rate and 1% probability do not mean the same thing and a 1% response rate does not guarantee that you will get 1 out of 100 once probability is introduced to the situation, which is vital to any useful application of this conversation.

Simply telling someone a 1% response rate means you'll get response out of 100 back is technically correct from a very literal stand point, but again it's of extremely limited applicable use when actually performing a mailing campaign.

I took the stats classes about 35 yrs. ago. I'll defer to your expertise.

 

[Newby]

If you loved calculus this should right up your alley as a solve is simply the integral of the p dist curve given the parameters.

Yeah. That's what I was going to tell him too. You beat me to it.