Why Adding New IP Adresses Won’t Necessarily Solve Your Complaint Issues

Posted by Cherie Ansari on

I recently ran into a situation where a client was asking if adding additional IPs would help reduce the “carryover effect” of complaint rates and improve their sending reputation. Here is my response and supporting reasoning, which may be helpful if you find yourself in a similar situation.

First of all, let me explain what I mean by a “carryover effect.” A carryover effect in complaints occurs when high complaint rates are generated as a result of an email that was sent on a previous day(s). Let’s say you normally send 10K messages per day, however you decide to send an ad-hoc email message to 2MM subscribers on Wednesday. Your complaint rates on Wednesday may be within acceptable thresholds, but the problem starts to happen when subscribers press the spam/junk button on Thursday, Friday, Saturday, and so on. Suddenly, you have huge complaint rate spikes per day because the sending volume (or denominator) was only 10K per day after Wednesday.

Here was my client’s situation:

  • Unwilling to change frequency or cadence due to business decision made by CMO
  • Unable to make changes to the email creative
  • Must send to everyone in their list file (rather than using a segmentation plan)
  • Sends to B2C and B2B mailbox providers (e.g., dell.com, returnpath.com, etc.)

Does this sound familiar? If so, you’re not alone. I’ve seen several senders, including ESPs, in this same situation. They have limited control over the email program, but have full control over their infrastructure and are looking for solutions that they can apply and manage.

With their situation in mind, these are the recommendations I gave them:

  1. Start with the basics and look at the math
    • B2C senders need to avoid volume spikes, unless it is part of their normal mailing pattern, as that can negatively impact their inbox placement rates. Mailbox providers pay attention to spikes, which is anything that exceeds doubling your normal mailing pattern. The table below shows the percentage change based on the client’s current situation with 1 IP and what would happen if 2 and 10 additional IPs were added. Notice that the percentage change remains constant at 19900% regardless of the number of IPs used. An acceptable rate would be 100% (or below), which is equivalent to doubling the sending volume. Thus, adding new IPs would not help. By trying to use this work-around approach, the sender would merely spread the carryover effect in complaints to other IPs at the same percentage rate.
      table_1
    • B2B mailbox providers use their own algorithms when deciding which email messages to accept. Thus, they aren’t necessarily concerned with volume spikes.
  2. Consider an IP segmentation strategy – Applying an IP segmentation strategy could be a viable approach to this problem. If none of your email campaigns are going into the inbox at the mailbox providers you deem most important, then consider placing your most engaged subscribers on one set of IPs and the least engaged on another set. You could also have another group of IPs for subscribers that are moderately engaged. Having a group of IPs targeted to highly engaged subscribers will likely improve your complaint rates and help minimize the carryover effect. The group of IPs for disengaged subscribers will most likely have poor deliverability and high complaint rates, but with an IP segmentation strategy, at least your most important subscribers will receive their email.
  3. Address rate-limiting – Adding new IPs specifically to address rate-limiting concerns (e.g., exceeding the maximum rate) can help, but the impact will depend on the mailbox providers. If you’re exceeding the maximum number of connections per SMTP session, then an additional IP can help as you’re opening an additional connection. In the case with my client, they knew the maximum hourly rates for their most important B2B mailbox providers, however the maximum hourly rates were based on a total rate regardless of the number of IPs. So in their case, adding new IPs wouldn’t help.
  4. Proceed with caution when combine IPs – Combining IPs can work, but it’s tricky. One approach would be to combine existing IPs if you have multiple ones. If you belong to a parent company, you could consider combining business units. However, I cringe at that approach because the IP reputation will be combined which means the business with the good sending reputation can be negatively impacted by the business unit with the bad reputation. The same holds true for combining the IPs within the same line of business (e.g., combining Company X’s marketing and transactional IPs). I would only follow this approach if you’ve exhausted all other options. For the sake of this blog, let’s say that was the case. Here’s what I would do to determine which IPs to combine:

The key to this approach is to ensure that your send dates are consistent. If they vary, then this approach won’t work as you’ll never know when there is enough volume to minimize the carryover effect.

  • First, get at least 3 months of data showing the actual sending volume by day for the IPs in consideration.
  • Then, display the data in a line graph, placing one graph on top of the other.
  • Look for volume spikes and see if the 2 IPs combined would complement each other. For example, say you have a spike on Wednesday from IP #1. Does IP #2 have similar volume on Thursday and then slowly tappers down? Ideally, it should look like the image below. Why? Well, let’s examine complaint rates by date.

The table below shows IP #1’s complaint rate per day. Notice that the carryover effect causes the complaint rates to spike at 50%, 5%, and 2.50%.

 The table below shows the new complaint rates by day when IP #1 and IP #2 get combined. (For the sake of this exercise, I gave both IPs the same number of complaints.) Although, Thursday still has a complaint spike of .50%, it’s much better than what it used to be at 50%.

Adding additional IPs to minimize a carryover effect in complaints doesn’t always help. The bottom line is that you need to get to the root cause to understand why people are complaining and address it from that standpoint. If you’re unable to do that, I recommend that you follow the steps above. If you come to the conclusion that additional IPs won’t help reduce the complaint rates, use your findings to demonstrate that you’ve exhausted this alternative solution and make a case for directing attention and resources to determining the root cause of your complaint issues.

 


Popular this Month

 Video in Email: Is It Right For Your Business? (Part 1)

Video in Email: Is It Right For Your Business? (Part 1)

Video in email is nothing new. Marketers have been using some form of video...

Read More

 [New Research] Are These Hidden Metrics Harming Your Deliverability?

[New Research] Are These Hidden Metrics Harming Your Deliverability?

Reaching the inbox is not as simple as hitting send. Once a message is...

Read More

 What Job Is Your Subscriber Hiring Your Email To Do?

What Job Is Your Subscriber Hiring Your Email To Do?

Over the last 16 years, I’ve worked as a product manager, run product...

Read More

Author Image

About Cherie Ansari

Author Archive

Stay up to date

Enter your name and email address below to subscribe to our mailing list.

Your browser is out of date.
For a better Return Path experience, click a link below to get the latest version.