postcard mailing snafu?
March 4, 2010 11:03 AM Subscribe
Is it possible the 14,000 postcards we sent were not delivered by the USPS? How would I find out?
We sent a promotional mailing of 14K cards several weeks ago. It is looking like a large number of these did not reach their destinations. We have checked with numerous names we know on our list, and they have not received them. These were sent through a reputable mail house, checked for weight and layout.
Does anyone have experience with this?
We sent a promotional mailing of 14K cards several weeks ago. It is looking like a large number of these did not reach their destinations. We have checked with numerous names we know on our list, and they have not received them. These were sent through a reputable mail house, checked for weight and layout.
Does anyone have experience with this?
Depends on what class they went out. Bulk mail will take a lot longer than first-class mail. Check with your mail house. We've had some bulk mail go out that took several weeks to reach everyone (depending on location).
And the USPS sucks. I sent a small package first-class to a Mefite from Tahoe to LA. It took SIX WEEKS to arrive.
Hint - always include yourself on the mailing list to verify that the mailing actually went out.
posted by HeyAllie at 11:12 AM on March 4, 2010
And the USPS sucks. I sent a small package first-class to a Mefite from Tahoe to LA. It took SIX WEEKS to arrive.
Hint - always include yourself on the mailing list to verify that the mailing actually went out.
posted by HeyAllie at 11:12 AM on March 4, 2010
Bulk mailers normally add various test addresses, some that don't exist, to determine if delivery was made. I'd be surprised if your "reputable mail house" didn't do this. Have you asked them?
posted by fydfyd at 11:34 AM on March 4, 2010
posted by fydfyd at 11:34 AM on March 4, 2010
Were they sent first-class or some other class (bulk, nonprofit, etC)
posted by radioamy at 11:45 AM on March 4, 2010
posted by radioamy at 11:45 AM on March 4, 2010
Response by poster: I'd be surprised if your "reputable mail house" didn't do this. Have you asked them?
Yeah... they did this. Claim they received them. We were in a rush and somehow did not get our own tests in the list (!). It just appears that a large chunk on these did not deliver. Of course, we're not expecting a landslide of responses, but we're getting nothing... nothing at all. Even at .5% we'd see 70. I've done a good number of mailers, and have not had this happen. Could the economy just be that sucky?!
posted by ecorrocio at 11:46 AM on March 4, 2010
Yeah... they did this. Claim they received them. We were in a rush and somehow did not get our own tests in the list (!). It just appears that a large chunk on these did not deliver. Of course, we're not expecting a landslide of responses, but we're getting nothing... nothing at all. Even at .5% we'd see 70. I've done a good number of mailers, and have not had this happen. Could the economy just be that sucky?!
posted by ecorrocio at 11:46 AM on March 4, 2010
Can the economy be as bad as your response rate would appear to be? Well, what are you selling? If it's a product or service for which demand has disappeared, well, then you have your answer.
posted by dfriedman at 11:52 AM on March 4, 2010
posted by dfriedman at 11:52 AM on March 4, 2010
Did you have one sent to yourself? That's a good way to check on mailings.
posted by Carol Anne at 11:59 AM on March 4, 2010
posted by Carol Anne at 11:59 AM on March 4, 2010
This might not be the answer you expected, but if you wanted, you could do a bit of probability to test this on subsequent deliveries.
You could use the Binomial distribution to test the success rate of delivery, but the numbers are large enough that the Normal approximation is better.
Let's say you expect that at least 95% of all your postcards are delivered successfully (*). We'll say the value p will represent the probability of success, p = 0.95.
We next calculate the mean and standard deviation based on the total number of postcards (n) and this probability (p).
μ = np = 14000 * 0.95 = 13300
σ = squareRoot[np(1 - p)] = 25.8
What this says is that, on average, we expect 13300 postcards to be delivered properly, plus or minus a small number.
Let's say that you find only 12000 postcards were delivered correctly, the last time around. Is this just bad luck or something fishy?
Next, we calculate a z score to find out how unexpected this 12000 number is:
Z = (x - μ) / σ = (12000.5 - 13300) / 25.8 = -50.4
That means you're roughly fifty standard deviations away from where you expected to be. A little bit more than two standard deviations is considered significant, in general. Fifty deviations strongly suggests that something is up with the delivery process.
The above calculations were done with the entire set of postcards (the "population"). It can be tough to ask all 14000 people if they got their postcards.
So you could instead do this test with a smaller sample, to test how things actually go in real life.
We'll keep the same p probability of delivery at 0.95.
Next, we ask ("sample") 100 postcard recipients at random if they got their cards. First, we calculate the mean and s.d.:
μ = np = 100 * 0.95 = 95
σ = squareRoot[np(1 - p)] = 2.2
Of those 100 recipients, only 80 report back that they got their cards. We apply a small correction ("continuity correction") and calculate a z-score:
Z = (x - μ) / σ = (80.5 - 95) / 2.2 = -6.8
We can look up the exact probability of seeing a z-score this low by looking at a z-score table. But anything more two standard deviations would generally imply something is wrong with the delivery process. It's unlikely that you are seven deviations away by bad luck. Not impossible, just unlikely.
(*) : How do you pick p = 0.95? This is up to you, since this is the success rate that you expect for postcard delivery. In reality, 0.95 may be too stringent and you may have to derive this from observation, instead. The post office might tell you from their own experience that their delivery success rate is 0.86. Or, for example, you send out small batches of cards at random, and on average you observe that only 65 out of a 100 make it, so your p of choice would be 65/100 = 0.65.
posted by Blazecock Pileon at 12:00 PM on March 4, 2010 [18 favorites]
You could use the Binomial distribution to test the success rate of delivery, but the numbers are large enough that the Normal approximation is better.
Let's say you expect that at least 95% of all your postcards are delivered successfully (*). We'll say the value p will represent the probability of success, p = 0.95.
We next calculate the mean and standard deviation based on the total number of postcards (n) and this probability (p).
μ = np = 14000 * 0.95 = 13300
σ = squareRoot[np(1 - p)] = 25.8
What this says is that, on average, we expect 13300 postcards to be delivered properly, plus or minus a small number.
Let's say that you find only 12000 postcards were delivered correctly, the last time around. Is this just bad luck or something fishy?
Next, we calculate a z score to find out how unexpected this 12000 number is:
Z = (x - μ) / σ = (12000.5 - 13300) / 25.8 = -50.4
That means you're roughly fifty standard deviations away from where you expected to be. A little bit more than two standard deviations is considered significant, in general. Fifty deviations strongly suggests that something is up with the delivery process.
The above calculations were done with the entire set of postcards (the "population"). It can be tough to ask all 14000 people if they got their postcards.
So you could instead do this test with a smaller sample, to test how things actually go in real life.
We'll keep the same p probability of delivery at 0.95.
Next, we ask ("sample") 100 postcard recipients at random if they got their cards. First, we calculate the mean and s.d.:
μ = np = 100 * 0.95 = 95
σ = squareRoot[np(1 - p)] = 2.2
Of those 100 recipients, only 80 report back that they got their cards. We apply a small correction ("continuity correction") and calculate a z-score:
Z = (x - μ) / σ = (80.5 - 95) / 2.2 = -6.8
We can look up the exact probability of seeing a z-score this low by looking at a z-score table. But anything more two standard deviations would generally imply something is wrong with the delivery process. It's unlikely that you are seven deviations away by bad luck. Not impossible, just unlikely.
(*) : How do you pick p = 0.95? This is up to you, since this is the success rate that you expect for postcard delivery. In reality, 0.95 may be too stringent and you may have to derive this from observation, instead. The post office might tell you from their own experience that their delivery success rate is 0.86. Or, for example, you send out small batches of cards at random, and on average you observe that only 65 out of a 100 make it, so your p of choice would be 65/100 = 0.65.
posted by Blazecock Pileon at 12:00 PM on March 4, 2010 [18 favorites]
Find out which USPS location your mailing house uses. Contact that location and ask the manager about your mailing. They can and will look to see if there is a delivery from your mail house which is still sitting in the back. They may say "nothing here" or "Yes, we received that and it was shipped out on XXXXXX date." You may find your postcards being delivered a few days later. I've had to do this more than once.
posted by onhazier at 12:10 PM on March 4, 2010 [1 favorite]
posted by onhazier at 12:10 PM on March 4, 2010 [1 favorite]
Response by poster: Blazecock... wow. Way over my right brained head, but thanks for your time.
onhazier... we're looking into it thanks.
Thanks folks. I'll update if I find out anything.
posted by ecorrocio at 12:13 PM on March 4, 2010
onhazier... we're looking into it thanks.
Thanks folks. I'll update if I find out anything.
posted by ecorrocio at 12:13 PM on March 4, 2010
Your reputable mail house undoubtedly has an account with the USPS.
Ask them to send you a link to the USPS page (not a page they have any ability to edit, in other words) describing the receipt and delivery of your order, complete with date, rate, weight and number of pieces. That ought to be enough for you to recognize your mailing.
If they resist or tell you it's impossible, I think you may have trouble with your mail house.
posted by jamjam at 12:22 PM on March 4, 2010
Ask them to send you a link to the USPS page (not a page they have any ability to edit, in other words) describing the receipt and delivery of your order, complete with date, rate, weight and number of pieces. That ought to be enough for you to recognize your mailing.
If they resist or tell you it's impossible, I think you may have trouble with your mail house.
posted by jamjam at 12:22 PM on March 4, 2010
I once received a stack of maybe 1,000 direct mail postcards bound together with plastic straps. Our postcard was on the top, natch. I wasn't sure what to do with them, so I cut the binding and threw them all back in the mail. No idea what happened after that.
posted by JoanArkham at 12:41 PM on March 4, 2010 [1 favorite]
posted by JoanArkham at 12:41 PM on March 4, 2010 [1 favorite]
The direct mail person in my department sent out a huge mailing and eventually they were all returned because someone had messed up the database and the street rows were one off from the state and zip rows. Not saying something like that happened, but it's a possibility.
posted by Kimberly at 12:52 PM on March 4, 2010
posted by Kimberly at 12:52 PM on March 4, 2010
Have you ever done a mailing like this before through your vendor?
Was there a unique telephone number on this mailer - one that would have been sufficiently tested "out of use" in the past?
Is the telephone number on the postcard the telephone number you expect it to be?
Is there another form of traceable contact on the postcard (unique URL, email address, foot traffic up at a store, etc.) that has seen lift?
Is this a postcard or an insert? The two are very very different. Postcard has its own unique mail code on it, whereas an insert may be bundled with other companys' inserts and sent in bulk out to customer's in batches. Some insert vendors also send things out on a graduated scale, meaning that if they were going out with say a cable bill, that it takes a full 30 days to unload all mailers.
Ultimately, its the vendor that needs to provide you the information, not the USPS. You need to hound them, and if you are too small for their business, then you need to find someone else who has some integrity in the business - there are companies that can handle a drop this small and ensure you get a proper delivery.
posted by Nanukthedog at 1:38 PM on March 4, 2010 [1 favorite]
Was there a unique telephone number on this mailer - one that would have been sufficiently tested "out of use" in the past?
Is the telephone number on the postcard the telephone number you expect it to be?
Is there another form of traceable contact on the postcard (unique URL, email address, foot traffic up at a store, etc.) that has seen lift?
Is this a postcard or an insert? The two are very very different. Postcard has its own unique mail code on it, whereas an insert may be bundled with other companys' inserts and sent in bulk out to customer's in batches. Some insert vendors also send things out on a graduated scale, meaning that if they were going out with say a cable bill, that it takes a full 30 days to unload all mailers.
Ultimately, its the vendor that needs to provide you the information, not the USPS. You need to hound them, and if you are too small for their business, then you need to find someone else who has some integrity in the business - there are companies that can handle a drop this small and ensure you get a proper delivery.
posted by Nanukthedog at 1:38 PM on March 4, 2010 [1 favorite]
>Yeah... they did this. Claim they received them.
Which is why you want your address (discreetly) inserted on the list. Not just theirs. Putting themselves in is nice for them, but not terribly informative for you, since you have only their word for it.
Putting yourself or your business name in is a step in the right direction, but any dishonest mailing house is probably going to be smart enough to check for the client. Next time (with whatever mail house you go with), choose someone who's name and address they won't recognize as your guinea pig. When I was doing solicitations for a university, we always had at least one department member's home address in the pile as a "contro" and kept our own record of who that was. Plus we had one sent to our business address. It's worth the cost of the stamps & postcards.
ALso: pay attention to whether it's first class or bulk: That's a HUGE pricing difference! That's not a datapoint to be fuzzy about!
Ask to see the cards they got (originals, electronic copies. They can FedEx them; don't accept any delays). If they went through the mail, they'll be postmarked. And you may find that the postmark is later than you expected. This happened to us when we were just figuring out the Gotchas with a new company.
posted by Ys at 6:41 PM on March 4, 2010
Which is why you want your address (discreetly) inserted on the list. Not just theirs. Putting themselves in is nice for them, but not terribly informative for you, since you have only their word for it.
Putting yourself or your business name in is a step in the right direction, but any dishonest mailing house is probably going to be smart enough to check for the client. Next time (with whatever mail house you go with), choose someone who's name and address they won't recognize as your guinea pig. When I was doing solicitations for a university, we always had at least one department member's home address in the pile as a "contro" and kept our own record of who that was. Plus we had one sent to our business address. It's worth the cost of the stamps & postcards.
ALso: pay attention to whether it's first class or bulk: That's a HUGE pricing difference! That's not a datapoint to be fuzzy about!
Ask to see the cards they got (originals, electronic copies. They can FedEx them; don't accept any delays). If they went through the mail, they'll be postmarked. And you may find that the postmark is later than you expected. This happened to us when we were just figuring out the Gotchas with a new company.
posted by Ys at 6:41 PM on March 4, 2010
Post Office is going through hell right now. Plus, you might want to consider all of the bad weather that has been nationwide as of late.
posted by bkeene12 at 8:01 PM on March 4, 2010
posted by bkeene12 at 8:01 PM on March 4, 2010
oops: "control" not "contro" and "original NOT electronic copies".
posted by Ys at 8:21 PM on March 4, 2010
posted by Ys at 8:21 PM on March 4, 2010
This thread is closed to new comments.
posted by elle.jeezy at 11:12 AM on March 4, 2010