|Directory first impressions: Can you tell a "book" from it's cover?|
Is a certain look unavoidable?
I've visited a few directories in my day.
Lots of overlap with DMOZ. Not big surprise.
Lots of websites running the same scripts. More overlap.
Is it the hierarchical nature of "directory" that mashes all directory websites into the look-alike phenomena?
Is there such a thing as a directory website that doesn't look like a directory but either manages to "work (for the visitor) like one" or, even if it doesn't "work like a traditional directory" it still manages to "work like a directory"?
Is the plethora of directory look-alike websites conditioning visitors to run away? "I've seen this before, haven't I?"
Just a thought: Is a directory a thing that conforms to certain structural standards or expectations?
I somewhat think that tagging - which creates something akin to a representation of a global structure - is at least a model of what I'm driving at in raising the questions.
Here's "the issue" in simpler terms: Is your directory's presentation "of itself" a conditioned stimulus for people to flee?
Again, in different terms, can you directory outside the box? Or, is the box the ingrained expectation, and therefore directorying outside the box jams the gears and is no better than "the expectation"?
Can a directory be playful and work? Can a directory misbehave intelligently?
Or, is it just the same old, same old, boring directory?
I've contemplated this a lot, webwork.
IMO there are many obvious flaws in how most directories function. The perfect directory should present the user with all the important relevant options in proper context, and very few irrelevant options when not in context.
As an example: It seems to me that a web directory by its very nature should nearly always have an individual listing available under all relevant categories, and I've never seen a directory that allowed this... DMOZ certainly doesn't.
If you have software that stores pictures of your dog on a web site, then shouldn't your listing be found under software, pets, image hosting, and web sites? This would make the listings add up much faster, but it would be more accurate.
My directories allow a webmaster to place his or her directory by submitting only the URL, category, and zip code if applicable, and then the listing may appear under as many as 11 different headings, but after some time of growth it is obvious to me that even this is short sighted.
I've also played around with a directory site that changes the focus of it's content based on what people are looking for. It was an interesting experiment as the most obscure things began to surface, the search engines picked up on those topics and sent visitors looking for that information so the site evolved even further toward those topics. Eventually the site started ranking very well for subjects that were so obscure that you would think that no one in their right mind would use it. Guess what? The majority of my AdSense income comes from that site, which is more than enough for my entire salary.
Very interesting dataguy. Very forward looking IMHO.
I routinely scan the posts back here. The paramount or repetitive concerns are:
- PageRank - having it, passing it, paying for a page.
- Search engine love, that is, some directory dropped out of SERPs, etc.
- DMOZ grievances.
- Is "this link" worth $x.
I'm not sure if most who post also run directories but IMHO the intelligence to be gained from the subject matter of the repetitive threads is unlikely to advance any directory operator's chance of success.
Why do I say so?
- PageRank is ephemera. It comes and goes. It is subject to manipulation. It is "page" rank, not sitewide directory rank. It is overrated in the directory realm. Push the issue as a sales point and you might just blow your credibility.
- Search engine love is the poorest of model for building a viable directory. I'll be even more blunt: Depending on a search engine to make your directory a viable thing is just plain dumb. Oh, yeah, it's nice to have traffic, but it's just plain dumb to make it part of your plans. If it happens it happens. Next!
- Most DMOZ threads are not productive of anything: The DMOZ is good, it's bad, they're not fair, blah blah blah. What, of any value, has come of thread after thread, other than perhaps saying "Make a quality site and there's some chance your site will be listed." End of meaningful discussion. Either you get listed or not.
- Is a link in directory M worth $X? Is the $25 in issue worth the cumulative time invested in address the same issue over and over? How about a definitive guide to analyzing the issue: What's the value of a sales lead the converts? How well filtered is the directory's traffic? How big is your budget? Etc.
I'd like to drive the directory dialogue a bit into subject matter this is likely to be more productive of value and long term success. However, I'm increasingly convinced that most people in the game look no further than their AdSense or affiliate stats for the last 4 hours. Not a bellweather of success.
I suspect that the subject matter of this thread will receive little uptake, not because it doesn't matter, but because so many people don't even know how to think this stuff through: "What the heck is he talking about?" "Um, can't matter 'cause he ain't mentioned pagerank."
It's not that they can't think. It's that they can't see outside a certain box.
Provocative? Hopefully. There's issues to discuss that matter and those that don't. Those that don't just seem to keep coming up again and again. Does that mean that they matter? No, IMHO.
Well stated Webwork.... I think you've nailed it.
I would add one more: "Which directories are the best?"
Best at what? Which looks the best? Which has the most traffic? Which makes the most money? Many of the directories listed in these discussions I have never even heard of. For sure, my sites have never been mentioned in any of these lists.
Passion or profit? My passion is for the technology. I'm glad we live in an age where this kind of passion can be profitable. Which directories will survive? The ones that I see discussed aren't really surviving now...
What I do think is interesting is that Google takes such an interest in directories. I believe their engineers have a passion for the technology as well. Remember, their mission is "to organize the Worlds data", and I think that is why directories are favored in their search results where none of the other majors care. Is your directory something you would be proud to show a Google engineer? I was when I showed them mine!
Make no mistake. Directories are important to the Internet Ecosystem. But you've got to see the big picture.
Bottom line is, does your directory provide a unique and useful experience to the end user? Anything short of this is not even worthy to be considered hobby, and certainly not a serious web business.
So how can a directory provide a unique and useful experience to the end user? Answer the question, win a million bucks.
Well, there may not be many takers on this thread, but I for one am interested in the input from experienced folks.
I have three very "niche" directory type sites. One includes the membership of a small non-profit (503(c)5 I believe). Inclusion in the directory part of the site is open to members, who are also able to include themselves in any of aobut a dozen categories, so there is a members list plus these additional cats.
Most of the site's traffic comes in via the sub-cats.
On another site I added a directory specifically to target localized niche terms. It is for a very specific area and is simply broken out into US States. It has been very effective in generating SE traffic on those localized searches, ranking top 5 in G across several hundred search terms.
Both the above use home brewed PHP scripts.
The third site in question is too new to tell. It is a rewrite of an older site, and it uses an open source CMS that has been modified a tad ;)
Now, reading this thread, I have to wonder what what would happen if a more dynamic approach were to be used. Self generated cats based in parsing out the query string from SE referrals. That would be interesting indeed, although the presentation of the heirarchy visually to visitors might be a bit of a test.
|Is there such a thing as a directory website that doesn't look like a directory but either manages to "work (for the visitor) like one" or, even if it doesn't "work like a traditional directory" it still manages to "work like a directory"? |
Perhaps we need to start by defining 'directory'. My definition (only good for the WWW - but that's what we are talking about anyway) is:
Directory is a site that is dedicated to the classification of sites according to a predefined criterion. The classification is human driven and not automatic / mechanical or software driven.
To me the real 'intersting' bit is the classification. The reason that the vast majority of the directories look like (somewhat distorted( Carbon Copis of DMOZ, is that they all us the same classification. In fact, even DMOZ was not original - they just took the Yahoo classification. To see for yourself - check the first entry of Yahoo in the waybackmachine.org here [web.archive.org]. In fact, even Yahoo did not invent the classification model - they followed quite closely the system that was devided centuries ago by Dewey - see here [en.wikipedia.org].
Few attempts were made to break that mold. The most noticeable is Gimpsy, which classifies sites according to online activity - not subject matter. You are asked to enter what action would you like to DO - not which subject you are interested in.
It is still a directory as defined above (you can browse and traverse the category tree as you would in any directory), yet it is distinctively different. Unlike the Dewey system, it has been around for just three years - so it certainly doesn't have the time-proven quality of the other system, but at least it is different!
I've been giving a lot of thought to willybfriendly's suggestion of self-generated categories. Still not sure how it could be implemented. I could imagine running the search string through a dictionary filter to separate nouns and verbs and clean up the data a bit, but it still would need human intervention. What I've done was self-generated navigational menus, but not categories. very interesting.
In defining directory I prefer to go to the definition in it's most basic form. A directory gives people directions, or people go to it to be directed.
I think it is useful to ask yourself then, how can you best give someone directions? My answer is that you nearly need to read their minds to find what they really want to know, and then present a short list of relevent options. Short of actually being able to read their minds, you need to lead them down a short path to determine what they are looking for. The key is keeping the path as short as possible.
If someone stops along the road to ask for directions, do they come to you totally devoid of any information before they open their mouth and start speaking? No. I would guess that most people would realize that the car pulled over the side of the road with the man hanging out the window probably contains people who are lost. What other information can you determine before the driver begins to speak? The direction the car is heading, the fact that it's Saturday afternoon, the fact that the entire family is loaded in the car, the fact that the car has out of state license plates, the fact that the driver is wearing a tuxedo and the rest of the family is dressed up?
You may be able to lead them down a shorter path if you determine that this lost driver is likely headed for a wedding and needs directions to the large church around the corner. Point him in the right direction before he opens his mouth and he will be amazed, thinking you've read his mind. Do this for your web visitors, and they will be amazed too.
Now how do we do this programmatically?
One thing that I haven't seen used by directories is visual emphasis of hotter categories. "Hot" could mean those with more listings, more viewing activity, more new listings, etc.
One popular way of doing this is by making font size proportional to the activity level. Take a look at [43things.com...] for example.
This is a fairly intuitive way to present the data, although it does give a "ransom note" appearance to the listing.
But, if we're thinking outside the box, every directory need not look like DMOX or Yahoo.
[edited by: Woz at 10:39 pm (utc) on Nov. 13, 2005]
[edit reason] fixed link :-) [/edit]
>> In fact, even DMOZ was not original - they just took the Yahoo classification. <<
Hmmm. I thought that it was loosely based on the USENET top level heirarchy, no?
43things is pretty cool. I had never heard of them before. Rogerd is right about the "ransom note" effect too, but it's very interesting.
What would happen if for categories, I allowed every webmaster to submit their own category names for their web site with the only limitation that the name be 3 English words.
My rough numbers show that the category names would likely be repeated once by the time I had 100 category names submitted. I could group category names when they are an exact match, re-arranged match, 2/3 match, or 1/3 match. I think this could work because there is a very finite number of English words, minuscule when compared to the number of web sites in existence. I can get a few thousand good submissions per day and I would expect that this could add up pretty quickly. Submissions would be hand-weeded, I am doing this already.
Would this provide more organic, non-linear directory listings?
These category names could also be used to provide search results, as 95% of all keyword searches are 2 words or less.
I think this might go a long way in phrasing categories in a manner that mimics common thought.
|I allowed every webmaster to submit their own category names for their web site with the only limitation that the name be 3 English words |
IMO, that would present you with a big problem of "overlap"; spelling mistakes, people trying to game the system by creating a unique category, etc. Could be solved with a lot of dictionary work, and if it took off could be very interesting, as it would build up a community interest structure.
43things.com - what the heck is it? A directory? A community? A wacked out version of h2g2? Good for fun browsing, but suffers from the same userbased problems as Wikipedia - only as good as the information that the volunteers put in. As much fun as it is to read about somebody trying to "frack a cyclon", it's not really useful.
If you used the "ransom note" effect on a standard directory, I can see that popular topics would soon weight the whole page, as people would start clicking on the bigger links, thus making them even bigger.
The reason that so many directories use the Dewey classification system is because it neatly orders all of the worlds subject matters into compact little data trees. If we want a logical directory structure, then we need to go down the Dewey (or similar) path. And while we can start to classify sites under multiple listings, that soon produces a horrendous amount of clutter that bogs down the directory.
So, instead of putting a "Pictures of my pet" site under software, pets, image hosting, and web sites, the Dewey system would put it in a tree consisting of Animals ¦ Domestic Common Pets ¦ Hobby Sites ¦ Pictures thereof. One entry, in a clear position, where anybody with half a brain can find it. If I'm in a directory, and come across an entry for a pet site under "software" or "web sites", I'm going to conclude that the system is wonky, and loose trust in it.
My gut feeling on the matter is that it depends on the ultimate purpose of our site. If we want to classify information, than a Dewey type data tree model is the way to go. If we want a stew pot of information, into which the user can dip at random to receive mad pieces of information, then the H2G2 or 43things ways can be best, as they present the user with topics they never would have thought of themselves.
Well, that depends what you're looking for ;)
|"Which directories are the best?" |
Of course, part of the problem as I see it is that until your reputation takes off, all directories are basically acting as an intermediary step between the SE and the end site. So, the user types in "Cat care site", and your directory pops up as one of the choices. So, the end user can either click on one of the other listings in the SE and go straight to the cat care site of their choice, or they can click on my page and see a list of cat care sites. How do we make it worth the users while to add 2 clicks to their surfing experience? By providing a clear, concise choice of sites for them to choose from. Check the referrer keywords, run it against your own database, and dynamically add new choices to the page. Suggest "related searches". Highlight the most popular pages. Etc!
There is an abundance of technology available now that can be used in directory development. I personally like to experiment and trial and it gets very expensive at times.
Efficient submission processing on one of the directories Iím affiliated with that offers free submission was a long project but has resulted in one of the best submission processing systems I have ever worked with. Itís nothing the user will see but enables us to offer services at well below competitive market values due to major labor savings. We have also had some very good results in the cross-referencing of submissions. I canít speak for other directories but the ones Iím affiliated with have continual development projects to improve services.
In my opinion any directory that sticks with the 90ís directory format and the ODP submission processing system will not be around in the future.
On the other hand directories have to be kept simple and easy to use due the nature of the majority of the directory users. Combining simplicity and advanced technologies available can be a challenge.
Just a quick question to quench my curiousity: Are there directory owners who are making money by using off-the-shelf directory software for their directories?
Part B: is there anyone making money from ODP clone directories these days?
I haven't seen many directories become very popular that weren't custom designed... I'm wondering if this is a mis-perception on my part.
I started out with an off the shelf script on one directory but it took many thousands of dollars in script development to make it efficiently usable. Script development seems to be a never ending project.
It did make a good base to start from however. Operating it as it was purchased I canít see as feasible for any large volume of business.
I developed my directory in php from scratch a few years ago. I didn't look into canned scripts because I enjoy programming and am very good at it. I figured no script could do the job that I had in mind. The site has evolved considerably from my initial vision of what it would be. Hard to see how that could have happened with a canned script solution.
The site looks quite different from other directories. Some people think that the home page is too busy and contains too much information. I designed it that way in order to rank better in SERPs. I wonder if now that G considers it an authority toning down the amount of info on the home page would hurt? I'm reluctant to risk it, but would be interested in hearing anyone's thoughts.
The idea that the "goodness" of a directory is a function of how quickly a visitor can get in and get out to be sent along to another site that has what they're really looking for concerns me. I have also read on this thread that depending on search engines for traffic to one's directory is a bad thing.
My directory depends on search engines for traffic. I work at developing other traffic, but the results of my efforts are small in comparison. It's nice to imagine getting by without the traffic from the major search engines, but in a regional, industry-specific site, that may be an especially difficult goal. My directory does have a loyal and growing following of regular users. The high ranking that my site enjoys for many keywords gets me traffic and it also helps for selling advertising.
I add content to give people reasons to stick around and browse, come back again later, and tell their friends. It's working. If it takes a little longer to find what you're looking for on my site, but the experience getting there is interesting and enjoyable, that's a good thing in my opinion. It's a balancing act between usability and stickiness for me, and I'm trying to target the sweet spot between them.
"As much fun as it is to read about somebody trying to "frack a cyclon", it's not really useful."
I strongly disagree. I can think of few things online that are more useful than information of this nature (for example, a great page I recently found titled: T-Rex versus M1 Abrams battletank - which would win?)