Robert_Charlton - 10:18 am on Jul 25, 2010 (gmt 0)
I do have my own viewpoint on this, but it's pretty much based on a feeling built up over time, rather than coming from a disciplined test.
Gut feeling here too, and lots of mental note taking. I think that A-B tests are almost impossible for hard-to-obtain "fine" links, particularly in real life competitive situations. The best I've been able to do in these circumstances is to try not to muddy up the waters when I've gotten a link worth evaluating.
As I see it, different links, like different wines, age differently... but the differences may be due to specific factors inherent in the types of links and sites. With wines, chemistry and storage are key factors. Toss in a few ingredients like acidity, sugar, tannin/phenols, alcohol content, etc... add the effects of heat and oxygen, and you may or may not have a link worth drinking.
Some links, I've observed, appear to build slowly and get better over time. Very hard to isolate this from other factors, of course, including additional upstream inbounds, particularly with big sites over a long time period. It may be that these are simply stable links on good, stable sites, and the neighborhoods they're in tend to attract positive attention in the long term.
Blog links, in contrast, have a quick rise and then diminish relatively quickly too... either because they move down and eventually rotate off the front page of the linking site, and/or because there may be a QDF type factor built into blog links which gives them a short life as well. Secondary links to a blog or news article can help keep them from fading... but if these secondary links are also from blogs or forums, they too will fade.
Other links, and these have been the most curious for me, seem to have a honeymoon period. They react more slowly than blog links, both up and down. I'd put some directory links in this group... and it may be that Google evaluates them to see if they receive other confirmation. Or it may be that they rotate off the "new listings" pages in the directory.
All linking pages are subject to changing conditions upstream of them... anything from code changes to business changes to algorithmic changes, all of which can positively or negatively affect inbound link juice. Some pages have good enough content to attract inbounds once they become sufficiently visible, which can mask other upstream changes. In some market areas, links will be affected by upstream link buying and selling as well. Large corporations do some notoriously dumb things on their sites and servers that can suddenly cut off link juice to their subsidiary companies. For the most part I'm surprised how long some inbounds seem to stay viable.