Engagement based delivery makes testing tricky

Yesterday I wrote about how important recipients are to achieving good delivery. The short version of yesterday’s post is that delivery is all about engagement, and how the ISPs were really focusing on engagement and proving custom user experiences.
This is great, for the user. Take the common example where a commercial list has some highly engaged recipients and a bunch of recipients that can take or leave the mail. The ISP delivers the newsletter into the inbox of the highly engaged recipients and leaves it in the bulk folder of less engaged recipients.
With user focused delivery people get the mail they are interested in where they can read it and interact it. People who have demonstrated a lack of interest for a topic or a sender don’t see that mail.
This can get complicated for those of us trying to troubleshoot deliver problems, though. I have a couple mail accounts I use for testing at various ISPs. Even though I do very little to try and personalize the account I am seeing behaviour that leads me to wonder if ISP personalizing the inbox experience is going to make it that much more difficult to troubleshoot delivery issues.
I have to wonder, too, where this leaves delivery monitoring services in the future. If delivery is personalized, how can you know that the delivery monitoring addresses are representative any longer? Is there even a “representative” mailbox any longer?