Forum Moderators: open
However, for this particular site, we receive visits via 1,000's of kw's, and track over 500 kw phrases (typically we track less than 20 per site). The kw's we track for this site cover a representative sample of virtually all of the various sorts of ways we're found in the SE's.
We cannot find any kw's / pages that have dropped in the SERP's. Well, maybe 20 or so of the 500 have inched down, but an equal amount have inched up.
We've sampled a bunch of G servers, no clues there. We've reviewed logs, things seem to be down across the board as far as kw's go (% volume among words has not changed much).
If I can't SEE the problem, I can't figure out what the source of the problem might be. Our SERP's look fine.
I feel like I'm trying to find a *huge* boulder in a flat sandy desert, but all I can see in every direction is...sand. Never encountered this before. Any thoughts?
Caveman, did you check all the datacenters. You might have dropped in the SERP's in most of them but not the one that you are defaulting to when you do your SERP analysis.
Worth a try anyway, IMO.
Of course, your server might be on the blink, but I assume you've checked the hourly traffic to look for any anomalies....
GuinnessGuy
How we can see a relatively uniform drop across kw's, when all kw's seem unchanged in the SERP's, and every dc we look at seems comparable to the others (for these kw's, with minor variations) ... I don't know. caveman's brain hurts.
What other datacenters can I check?
Thanks in advance.
all of the www-? type datacenters were disabled *months ago*
Cabbie, could be something along those lines, but so far checks of our site reveal nothing except our pages in the results.
MHes, we've just been talking about that. We have not lost ground at all except as noted above (few up, few down)...but a strong competitor has recently show some strength just *below* us in some important SERP's. They must be on a link dev push lately. :)
I guess it's possible that a site just below us in most SERP's could cause some problems, but doesn't seem that they could be accounting for this.
Well, it still too early for cavejuice. Back to trying to figure this out...
My sept 11th memorial site is (and did in those years) showing HUGE increases in traffic. A hundred fold increase.
Richard
On further analysis it turns out the ranks for 2-3 word searches had remained (these were the obvious keywords and showed up on our reports as multiple people typed them) were they were. What had fallen off were the 4-5 word type searches that only one user a day had typed and we'd never logged those, because of err .. statistical insignificance.
Does this sound like what has happened to you?
Couple of things that have helped since this happened (late June / early July) was tightening of the link distribution. We had several hundred links which were spread between directory/ and directory/index.php this was sorted out. The home page was tightened up a bit too.
richlowe, all I can say is, your explanation is as good as any we've seen or come up with on our own. Anybody else see that? I have to say though that it's not across the board for our sites...
shri, your comment is similar to one above...we're doing a more extensive log analysis now to see if the longer strings are where we lost traffic. That would be something quite new, I think, since post-Florida, when the longer more specific strings seemed to perform best of all for us...as G was quick to point that out, when everyone was screaming about lost traffic.
I'm curious...how would your index links or homepage tightening affect your success with longer search strings, which I might generally have associated more with sharply focussed sub-pages?
Well first of all, (my untested theory is) longer and obscure strings usually have PR working for them. Given that they're unoptimized, never in any major onpage sections like title, Hx etc, about the only factor that can bring longer strings into play is PR.
When I looked at the pages in google I found two caches of the page ( directory/ and directory/index.php ) with different cache dates. Also the PR was different for the two variations. This lead me to belive that there is some sort of dampening going on due to mutiple views of one page and multiple PR values.
I'm in week three of this exercise and am seeing traffic rise back up very gradually.
We also reduced the number of internal links from the home page (highest PR) into the various subsections. More focus on internal anchor text over there.
My traffic was something like
Mid-Late June: 25K
July: 1.5K (cleaned up a bit of 301 code -- I know I got hit by something else here)
Early Mid August: 3-4K (cleaned up page variations)
September: 5-7K (seeing larger phrases come into play)
You might want to check one thing.. how old is the cache of your main page in Google. For a PR 6 site, I was very surprised to find a 7 day old cache in there when I was looking at the problem. The cache is still old and takes far longer to refresh that any of my other PR6-5 sites.
Added: In case it helps my site is several years old with a PR of 6, 1,500+ backlinks and the cache is 2 days old.
I believe this is the root of my current evil drop . . .
I have also seen "site map" type pages w/ internal links and descriptions of each page drop recently. They used to rank well this summer - and I was always a bit shocked to see it everyday -
A few days ago, for a site of about 1,000 pages with good traffic, G referrals dropped about 40%. Not much news there.
Could just be luck of the algorithm lottery this month. I've got several completely unique versions of my site just for such an occasion (all same topic just different names and pretty pictures ..... kinda like General Motors .....) anyway - each is optimized completely differently, some are so far apart that you wouldn't know the same SEO managed them. On the exact same day that traffic dropped on some sites, some stayed the same and it rose on others.
The drop in traffic was about the same as in your case (40%-50%) but the rise on the other sites was almost equal to that. So I figure Google has chosen a different flavour of the month as it suddenly liked some of my sites and suddenly disliked others whilst remaining impartial on the rest.
Even though it is bad that we have had multiple sites drop, the one good thing is that we can compare our sites to see what they have in common between each other and not in common with those that stayed ranking high) in order to determine what could have created this drop.
Unfortunately, we have been able to disprove nearly every theory we have come up with because it is not common among all of the sites that dropped.
So we are searching for basically anything that could have an affect. Here are a couple of things that the sites have in common (although I couldn't imagine they would be the reason, but I figued I might as well throw them out there):
1) All of the sites that dropped had multiple h1 tags - as well as multiple h2 tags (all of the sites ranking above us, along with our one site that is still ranking well, have 1 or less h1 tags).
2) Google is devaluing our links for some reason on these specific sites (we didn't lose any links and are still doing excellent in Yahoo and MSN). But I am not able to come up with any reason how thsese links differ from our competitors.
3) It is possible google manually penalized the sites since they are all owned by us (even though WHOIs info is different) and they did leave one site up which is doing very well in the SERPS, which would make sense if they did do something manually. This site does have more features than the others, so if someone were to penalize our sites and leave one, this one would be the most appropriate and the site that stayed up does not have as many links as some of the ones that dropped.
#3 holds up better than the others when doing our analysis, but since other people are having similar problems, I seriously doubt this is the reason.
For example, in one SERP, we were ranking 2,5, and 6.
Now we are appearing 24, 25, and 26.
On another we were 2, 3, 5.
Now we are appearing 30, 31, and 32.
This is repeated on many other SERPS as well.
That almost seems like too much of a coincidence to happen naturally. We do not interlink heavily (with no reciprocal links between our sites) and have many, many links from unique sites.
Again, the problem in our particular case is that we find *no* decline in the SERP's for our pages...and no vanished pages...only declines in traffic. If we saw concurrent drops in the SERP's, or newly missing pages, that would explain it.
The first thought in cases like this might be an overall decline in Web traffic. Not the case. Another thought might be some other external force like richlowe's comment about 9/11 (for example, maybe the traffic is just not there for some categories, even if it's hard to see a connection between category and event).
We're still looking at the possibility that the competitor that just came up *below* us is providing more competition for us than the site that used to occupy that spot. But I doubt it becasue while the site is nice, it's actuallly not so relevent to the searches as ours is. The site seems to have improved in the SERP's thanks to a growing network that spans many different categories.
Other than that we're still without answers, and will wait to see if traffic bounces back today, which might support the '9/11 effect' theory, rather than 'the guy below us' theory.
However, I have another update...traffic appears to be back this afternoon...still with no visible change in any SERP's at all across hundreds of kw's. And no apparent difference across G servers. We're talking 40% down, now 40% back up again...
So unless there are a large number of G servers we don't know about, it could be the 9/11 thing. I see a sort of distant connection between 9/11 and the contents of this site, but it's distant at best. Otherwise, I'm stumped.
Maybe it was vacations ..
Up and down, up and down, up and down...
