This is part 2 of a two-part special on web-based feed aggregators. See part 1.

The previous post on Bloglines and Google Reader brings me up to a another topic, that of feed reader lock-in. Both Google Reader and Bloglines support the export and import of feed subscriptions via the OPML standard (that’s how I switched from Bloglines to Google Reader in the first place). This means that both aggregators do not lock-in customers. Nominally.

Unfortunately, one big stumbling block against migration is the attached user metadata on the feed item level. One example is the starred items in Google Reader. While I can migrate my subscriptions to other aggregators, I can’t export which feed items are starred. This would be eased somewhat if Google Reader had the option to export the starred items into a bookmark file so that I can import them into my browser’s bookmarks/favorites or upload them into del.icio.us or similar sites.

Another stumbling block, especially among web-based aggregators, is the way they “aggregate” feed items. Aggregators were named as such because they aggregate multiple feed sources for you to read in just one location—no more visiting numerous blogs and newspaper sites just to keep up with updates. But web-based aggregators also aggregate in another sense: they store a feed’s items in their servers and serve them to all users that are subscribed to that particular feed.

Unlike desktop-based aggregators that only store a feed’s items from the time you subscribed to that feed, web-based aggregators store a feed’s items from the first time any user subscribed to that feed. If you subscribe to a feed that another user has already subscribed to, you get the benefit of free feed “back issues,” feed items that were already “seen” by the other user. This, admittedly minor though useful, benefit provides the web-based aggregator a hidden sort-of network effect feature, which contributes to lock-in.

I guess that there is probably a need to extend the feed syndication mechanism such that you can request the last N items of this feed or request the items published in a particular time period for that feed. Some websites already offer this feature, but I haven’t seen of any standard way to do it. Such a mechanism, if successful, will reduce lock-in and will make even desktop-based aggregators much more useful, but this is made complicated by the presence of feed services like FeedBurner.