Special thanks to:
|
#21
|
|||
|
|||
|
Matt Cutts says that the tag only needs to go on the exact page or pages that are listed in the ODP and for which you do not want Google to use the ODP title and description.
|
|
#22
|
|||
|
|||
|
I think you may all find my experience with the use of the NOODP tag quite interesting.
I've always hated the DMOZ description that Google shows for our site. For quite some time they have been showing it for a search on our primary key phrase. So when I heard about this tag, I put it on my home page immediately. Once Google picked up on the tag, sure enough the ODP description was gone and they showed my meta description. Great right? Well except that simultaneously to that, my site disappeared for our primary key phrase. I don't mean just dropped off the first page, I mean gone. The addition of the NOODP tag was the only change I made. Now I thought that it was possible that it was a coincidence, so I took the tag off to see what would happen. Sure enough, as soon as Goog picked up on the new version, bingo! I'm back on page one. In fact I'm back on about 70% of the data centers, and the ones that I'm still gone from have the old version cached and the ones I'm back on have the new version. I would say that this is pretty strong evidence for the idea that some have already come up with that the ODP description being more trusted than something pulled from the site itself. It obviously factors into their ranking algorithm pretty heavily. I hate that ODP description, but not that much ![]() I guess the moral is: If you are going to implement this tag, have that be the only change you make and carefully monitor your ranking for your major phrases, especially the ones that are actually in your DMOZ description, and make sure your site doesn't get hammered in the serps. Last edited by phpmaven : 07-30-2006 at 02:15 AM. |
|
#23
|
|||
|
|||
|
I'm not disagreeing with you about the reason for the disappearance and reappearance of the ranking, but, to ensure that coincidence is not the reason, I'd like the same thing to occur several times, and in sync.
If it really is as you believe it to be, then the ranking disappearing from the top 1000 seems rather a big drop. Did you check all of the top 1000? Link text is attributed to the target page, so the DMOZ Titles are counted as being on the target pages, and it may be that their descriptions are given the same treatment. It may be that using the NOODP tag causes that text to be removed from the index, and in your case it was very influencial in acquiring the ranking. |
|
#24
|
|||
|
|||
|
No, I don't think I checked all of the first 1,000 results, but I did look through the first 300 or so. I guess the point I was trying to make was that I didn't just drop from #6 to #7.
Yes, it will be interesting to see in the weeks ahead if this happens on other sites. I'm not saying that this is conclusive irrefutable proof. I'm pretty confident that it was the adding/deleting of the tag that affected my ranking so dramatically. If I was a bit more of a risk taker, I would try the same thing a few more times to see if it's repeatable, but I'll leave that for someone else to try ![]() |
|
#25
|
||||
|
||||
|
Phil that is a good point: a larger sample would be nice. Any others have had similar experiences?
I will try to post some results when we get some... |
|
#26
|
|||
|
|||
|
Merged: Is it NOODP's Fault?
I have a client that recently added the NOODP tag to his pages. As soon as Google cached the pages, each of the data centers dropped their first page rankings for a very competitive phrase. A week or two later he added it back and again, as soon as the pages were cached the rankings returned.
He's convinced it was the NOODP tag that caused it, but I'm not convinced, but it does seem awfully coincidental. I'm going to recommend he does some more testing before drawing conclusions, but he's probably not willing to play with such a lucrative position. Anybody else see anything similar or have any thoughts on on this? Last edited by Chris Boggs : 07-31-2006 at 12:43 PM. Reason: merged thread... |
|
#27
|
|||
|
|||
|
Might be cause/effect. Matt Cutts today, in his video, explained when/why the odp description is used. The gist of it is that it is query-dependent, and will be used when the odp description better matches the specific query. Now, it stands to reason, that if the description better matches the query, then "maybe" it will also add an extra point or two to the score. Just thinking logically, if a "trusted source" says your site is about "xy and z" (the odp description), and the query is for xy and z, then you get extra points added to the rankings score. Taking away the description, and thus taking away the extra point or two, may cause it to have less "points" than the competition...thus dropping in rankings.
|
|
#28
|
||||
|
||||
|
Note: posts 26 and 27 merged from another thread...
|
|
#29
|
|||
|
|||
|
It seems, then, that the ODP description isn't just pulled and used in the search results, but it's also factored into the algo. And not just a little, but quite significantly. This makes using the NOODP tag a bit dangerous, don't it?
|
|
#30
|
|||
|
|||
|
Quote:
It could be argued, though, that if Google uses link context as a factor, then an ODP description might influence rankings. I've never been able to observe even the slightest effect of description vocabulary on rankings. I'd assume that, on anything very competitive, the effect is very small. Since Google is keeping the ODP descriptions in its database, ODP descriptions might conceivably be more influential than most. |
|
#31
|
|||
|
|||
|
As a PS to the above, I have a very strong hunch that using the NOODP tag does not affect rankings.
|
|
#32
|
|||
|
|||
|
>> does the noodp tag only need to be placed on direct urls that are in the ODP or do they need to be placed site-wide? <<
Matt Cutts has already confirmed that the attribute only needs to go on the exact URLs that the ODP has listed. |
|
#33
|
|||
|
|||
|
Quote:
|
|
#34
|
|||
|
|||
|
It most likely is a coincidence, but I wouldn't assume anything without testing it further. I realize it's a client, so that may not be possible, but in general, those kinds of possible cause/effect situations should be tested as much as possible. If it's just coincidence, great. But if not, it would be good to know.
I have seen some assumed coincidences turn out to be actual cause and effect relationships. Of course, it's more often the case that assumed cause/effects are really just coincidences. But not always! |
![]() |
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
|
|