Besides the fact that RSS subscribers can feel less connected or loyal to content creators (see The Wire Effect), content on a per source basis can also face ambiguity problems.
RSS feeds are published based on timestamps and not on relevancy or interest. That means that one of my better posts could get buried beneath a handful of other content. It can also result in content that is not of interest to readers. The second problem is typically addressed by providing readers with a wider selection of RSS feeds.
To my knowledge, however, there are less ways to fix the first. One solution is to totally ignore the relevancy issue on the content provider’s end and allow the consumer to use tools that account for attention data. Several RSS readers learn from reading habits and collect attention data, resulting in a smarter reading experience.
A different approach would be to use metrics to create a smarter RSS feed. FeedBurner already tracks views and clickthroughs. And since they are already resyndicating content, they could provide an alternate RSS feed based on their metrics (i.e., relevancy) and not solely timestamps.
Relevancy based RSS feeds would be particularly helpful to publishers that are really churning out content. From a consumer’s perspective, it becomes tiresome to have to wade through it all. Relevant based RSS feeds could serve both publishers and consumers better.