How Often Should You E-Mail People?

I was looking at the results of a number of permission email campaigns the other day and something jumped out at me. In our mailing program targeting consumers who specifically requested to receive product or service information, we’ve determined the optimal number of times to email someone is six.

I know six sounds like a lot, and you’re probably asking yourself, “How did they arrive at that number?” As with everything else we do, we conducted extensive testing. And now I’ll share this relevant information with you.

In surveys of email customers in our database, we uncovered many reasons why consumers may not respond to the first email they receive, including:

  • The email didn’t make it to its destination.

  • The message got lost in an email-packed mailbox.
  • The recipient deleted it by accident.
  • Someone else using the computer deleted it.
  • The person was too busy to respond.
  • He was on vacation.
  • She had to discuss it with someone else.
  • He wanted to compare with other offers.
  • It just didn’t “strike” the recipient in the right way.
  • She filed it for future review.
  • He simply forgot about it.
  • It was blocked by an ISP or user-installed software program.

Just knowing why someone didn’t respond to our initial email doesn’t tell us much about the number of additional mailings we can or should send as part of the optimal mailing cycle. To determine that, we needed to identify three key markers in that cycle:

  1. The point at which complaints would start coming in from recipients annoyed at getting too many email follow-ups

  2. The point at which response would fall off to the extent it makes no sense to continue mailing and incurring bandwidth and operational costs
  3. The point at which the number of unsubscribes would become too significant to ignore

So here’s what we did — and here’s how you can glean the same information about your own offer. (Of course, as you’ll soon see, not all offers behave the same way.)

We started our test by sending 10,000 emails to 10 groups of people and tracking the results. Group One got one email message, Group Two got two, Group Three got three, and so forth.

We then produced a spreadsheet that charted — by group — email sent, opens, clicks, revenue, unsubscribes, and complaints. (Though this information is pretty standard stuff, you may need to pull it from several sources.)

In this exercise, we found six mailings were optimal, because starting with Group Seven, there was a noticeable drop-off in revenue and a corresponding increase in unsubscribes and complaints.

That’s the simple process we used to determine an optimal mailing schedule. The entire test took less than three weeks, although we pretty much knew the bottom line after 14 days. The one variable we didn’t test — because we didn’t think it to be important — is how many days apart the mailings were sent. In our test, we sent them every other day.

Now, back to what jumped out at me when I was reviewing some statistics. On one of our offers that targets seniors, an interesting pattern emerged in terms of the percentage of revenue from each of the six mailings:

  • Mailing One = 58 percent

  • Mailing Two = 11 percent
  • Mailing Three = 7 percent
  • Mailing Four = 7 percent
  • Mailing Five = 5 percent
  • Mailing Six = 12 percent

Certainly you would expect a lot of activity to result from the first and second mailings, and it did. But what I found surprising was Mailing Six did more revenue than all the others, except the first.

From that, I concluded the following: Had we not tested the ideal number of mailings, we might have arbitrarily decided to send out only a couple. If we had done that, we would have lost a lot of revenue. Or we might have sent out 10, creating a painful number of unsubscribes and complaints.

If you’re worried about aggravating consumers who received your first email but chose not to respond, consider this: In the postal industry, advertisers send many consumers multiple packages for the same reasons I’ve explained here.

I’d also like to point out these mailings were sent to consumers who specifically requested to receive the information, and our incidence of complaints was virtually nonexistent. Also, by slightly modifying each email to provide new information, we kept them fresh and unique.

Without prudent testing of the numbers of mailings you’re sending for each offer, you could be leaving money on the table by doing too few or losing valuable email addresses by doing too many.

Keep reading.

Meet Paul at the Jupiter ClickZ Advertising Forum in New York City on July 30 and 31.

Related reading