With complaints driving the majority of deliverability issues, it is critical to know whether you have complaint problems-- and how to fix them if you do. There are some simple ways to find out whether complaints are hurting your email reputation and response rates. One way is to focus your efforts to find "pockets" of complaints inside your program by performing a complaint analysis. Here are some things to aid that research:

Customer Support: Your own customer support email boxes and any auto-reply addresses can show you which emails are offending customers. Do not forget to monitor these internal resources. Also, you should register your abuse addresses with http://www.abuse.net/.

ISP feedback loops: When people complain about your email to their ISPs, that data is often captured in a Feedback Loop. AOL currently provides that information back to marketers, and several other ISPs will start doing so soon in the new Abuse Reporting Format, or ARF, which will standardize the data elements in feedback loops, making it easier to machine parse.

Reputation Reporting Systems: Reputation monitoring systems are coming to market in the near future.

The information from feedback loops and your customer service emails can show you who is complaining, what content drives complaints, and much more. In running a complaint analysis for a client recently, we found the following actionable items:

• The best co-registration data partner had complaint rates that were 20 percent of the worst co-registration data partner. The implication here was to drop the worst data partner(s).
• Subscribers that were in their first two weeks on the list had 10x higher complaint rates than clients that had been on the list longer. However, this held only for third-party marketing messages and not for the company newsletter. Mailing new clients from separate IP addresses made sense here. In addition a more detailed review of the client's lifecycle mailing approach was in order.
• Particular subject lines and creative drove much higher complaint rates (in some cases up to 20x the average). This indicated a need for more pre-mail testing.

This is the sort of information that can be gleaned from feedback loops. Best of all, it's free.