Forgive me for this but I'm about to jump around a bit and take a quantum leap or two, in a somewhat round-about way of answering "the directory question". I'm taking this approach, in part, because it's difficult to address questions like yours without getting deeply into the profiles of the directories you have been targeting and the websites you have been listing - amongst dozens of other factors. So I'm going to be a bit sweeping in my analysis and discussion.
What makes my reply even more problematic is my lousy PageRank attitude. At least in the context of directories, and probably in every other context, my "PageRank advice" is to cease making PR a consideration OR make it the lowest of all considerations.
It's my view that IF search engines are ever to "get things right" - have a long term value proposition - then their ranking algos will have to deliver SERPs driven by factors heavily weighting user experience. I wish PageRank equated with user experience but, in my experience, it isn't, especially in the PR2 to PR4+ PR realm. There can be great webpages lacking PR and there are crummy websites/webpages that tout their "PR-4". (Frankly, I'm not sure I see a good reason to grant PR to internal directory pages unless they are a thing of unique brilliance=utility. How rare is that?)
What does this "snooty PR thinking" have to do with how search engines treat - OR ought to treat - directories, themselves, OR links from a directory? In other words, how might your company best contextualize your company's engagement with directories?
User experience, in the context of directories and/or links to directories and/or links from directories has a lot to do with the satisfaction of search intent - since, at least in theory, directories, like search engines, "are about search" - so when it comes to choosing directory links it may be of value to ask whether a directory uniquely satisfies user search intent.
A million general directories can provide a link to sites like Business.com, which is of some value, but what's the unique value? Not much. Worse, maybe all the directories use the same blah-blah-blah description of Business.com. That's my observation. Business.com "feeds the search engine" a description of itself. How helpful is to user experience to clone that description?
Maybe better descriptions connected to the outbound links would help boost the link value and user experience? How uniquely helpful are the descriptions associated with the links in the directories you have been linking to? Did you fully consider that before submitting a link request to the directory?
Maybe some of the best directories don't even call themselves a directory, yet operate like one? Have you noticed a downfall of link juice from this type of "non-directory directory"?
Maybe some of the best directories are actually sub-directories of topical websites? Maybe some of the best directories are found on .edu or .gov websites, directories crafted by folks with deep expertise or respect for their reputation or the user's time? Have you noticed a drop in link juice from "directories" in this category?
The idea of getting links from "niche directories" is popular but is "niche alone" a sufficient driver of value when it comes to satisfying user search experience in a manner that might either aid or be superior to what a search engine, alone, might offer.
Maybe what you're seeing, i.e. issues of directories either not passing PR or some listings resulting in a PR drop, might have something to do with the editorial quality of the directories you are targeting getting an algo tweak OR maybe there's been a dampening of the determination that the directories add value, maybe there's been a determination (algorithmic) that a paid link is a paid link OR that directory links as a whole are a bit more suspect of attempts to boost PR, etc.
At to the latter, look at your own comment. One would have to infer that you may have been "selling the service" based upon some reference to PR, ranking, etc. Whose to say that one of your "customers" wasn't . . someone with search engine connections? Expressed differently, hasn't Google offered the advice that, in all ranking fairness, paid links shouldn't pass PR? So, should links in a paid directory pass PR? OTOH, links from a "free directory" - in many, many cases, might not pass PR simply because the editorial controls are consistent with "free".
Maybe the "drop" you're seeing relates to the rate of link aggregation? Maybe it's just too easy to get links from the directories you're targeting? Maybe it's not a matter of "directories are bad" but that their links are just not that good a signal - as a whole - and maybe they're a signal of a website attempting to game link juice? Ergo, too many of 1 category of links, too fast, and your client's site gets nailed? Maybe you'd be better off getting a few links from a few directories that send signals of quality? Maybe that signal of quality is, in part, how many "fresh links" the directory is sending out, especially to "new websites"? (It's such fun to peel the algo onion, isn't it? :-P )
I'd say unthink PageRank altogether and think user/vistor experience AND, in that context, I wouldn't waste another second on the questions about directories and PR and, instead, I'd ask myself "In looking at this or that directory can I tell where the operator's head is at and where the directory's development is headed and is that consistent with this directory, now or in the future, likely to drive traffic and drive the type of traffic that would engage my listed website in a way that all parties view beneficial"?
That's a lot to think about but it's my view on where the WWW is headed. Yes, there will be no end to spam and attempts to manipulate SERPS, but at the same time the effort to identify, find or produce signals of value and quality - especially signals that "your time won't be wasted" by engaging this or that website - will be on the rise.
So, if you are going to offer a service that includes seeking or delivering links to directories I might suggest that you develop your own list based upon criteria I've discussed and others that you might develop. AND forget PR - anywhere and everywhere in the criteria.
Maybe your best criteria might be "Would I bookmark this directory"? I think, given Google's propensity for aggregating user data, that might not be a bad "decision point" when it comes to choosing where to seek to place links for clients, especially when it comes to directories.
Hold this final thought: Do you remember the scene in "Miracle on 34th Street" where Macy's trumps Bamberger's by doing the right thing and "making the referral"? When a search engine chooses to send "a searcher to a search competitor" is it doing the Macy's v. Bamberger's thing? Would it be a good thing to only make that referral IF "the other guy did this particular search better"? When you choose a directory to seek a link from it for a client's website are you asking yourself the question: "Is this just another link (Google looks askance) OR is this website doing something just a bit better than Google in this instance?" There are myriad ways in which a variety of directories that I'm familiar with "do it better than Google", for now, but Google is always working to "do it better" themselves. (Note the bulked up local listings atop Google, for example, and ask yourself what challenge "that design change" was intended to meet or counter in the directory space.)
Alrighty, I'm sure that worked, right? I mean, as always, I've answered a direct question directly, made things perfectly clean, and delivered new insights, right? Argh. Well, hopefully, I at least haven't muddied things or wasted your time. As I often say "Food for thought" . . and indigestion. :-P
All other opinions welcome as I wouldn't want to lead UKSEO off the cliff that I'm likely headed towards myself. :)
[edited by: Webwork at 5:47 pm (utc) on June 13, 2008]