Forum Moderators: open
Anyway, the site just took a major hit across all pages, and it's a rare time when we've wondered if we were affected by the algo element(s) that some refer to sandboxing.
Any experience along these lines? We were not aware of an existing site being hurt *just* by virture of adding 30% more pages. (No dup issues or anything obviously like that at play here.)
There are text blurbs for each entry, and we write every one of the entries ourselves; 10 or 20 entries per page...thousands of entries across the site. And there are also contributed articles (hundreds) from enthusiasts with interest/expertise. ...Also, I used widgets to illustrate, but the site is not about things for sale. Mainly information. It's closer to a site about butterflies than it is to a site selling books or bbq's. ...
IF the site were about butterflies, which it's not, and we added a new section on European varieties, when the site had mainly covered N. American varieties, one might expect the new pages to be structured similarly to the old, as is the case here. ...
...all we did was to expand a previously meager section of the site (single page) to be comparable in quality and structure to the other sections. Doing it in pieces would have made the new expanded section look odd/incomplete, so we waited until it worked well as a stand alone entity.
All original content, as noted. Uploaded when the new section was complete and ready for prime time.
Then we saw a dramatic drop on Sept 9. The pages that were dropped out were deep pages that did not have much PR, so I don't think we were hurt because of a PR loss, we have some great links coming in to our important pages.
We also noticed we stopped getting fresh tags about 3-4 weeks after the # of pages were cut in half.
My best guess is that we suddenly appeared 'abnormal' ... but in this case, a 'best guess' and a 'good guess' may not be the same thing.
And no, there were not common external links to the new pages, but that was a great question. :-)
I'm pretty conviced that G has been tweaking their dup filters on any number of levels: internal and external links, anchor text, body text, page structure/templates, etc.
We're trying to sort out now what aspect of the new pages may have caused this. We're looking closely at links.
However, I can't prove that the new pages had anything to do with the problems.
I'd be curious if many other members could cite recent situations where they added a substantial number of new pages to a small/mid-sized site and saw their entire site drop in rankings.
Also, more important, are there many who can say they've added a large percentage of pages with no adverse affect? I frankly still keep thinking it has to be something else. If *new* sites have trouble ranking, and *old* sites can't add pages, then G would essentially be strangling the Web. :-(
The trouble was, not only had it been around for years, but it hadn't really been updated for a few years except for adding the extra bit of content to pages here and there. We wanted to start getting more out of the site in terms of revenue so it needed an overhaul.
We knew of the suspicions that changing a site could lead to problems so we were careful. Every single page (read exact URI) from the original site was maintained including main category, sub category, content and site map pages. They were left there and the linking structure was kept identical.
Regarding on page factors the template is basically the same with some small adaptations but the bulk of the facelift was done by tweaking the external CSS.
The biggest changes were:
1) to add a significant amount of new content to each page
2) we launched 2 or 3 new categories, about 180 pages per category
So, where are we now? Well basically Google dumped all of the old pages in mid September, despite our efforts to keep them looking pretty much the same in terms of their underlying code.
Whats is funny though, the new categories actually rank prety well, in fact really well in some cases - where I would have expected them to rank if they had existed on the old version before the changes.
Somethig else, we thought we might be on the come back trail last week. One of the search term where we used to be #1 got dumped to about #90, but every now and then we were seeing it it pop back up to #1 (and #2 if you count the indented result). But this was really intermittent, even when querying just the one data center where it was showing. Sadly though, this week we don't seem to be figuring at all. I am pretty confident that these pages will surface again though!
I must admit I don't really know why this could be but its probably something along these lines:
* many pages updated at once
* google somehow thinks their is an old version and a new version which are duplicates, but I can't see how as all the URLs are the same
* the site was going to get hit anyway and its just a coincidence
* because the pages are now new they must meet some additional criteria (but then why would the new pages that didn't previously exist rank well)
* because the pages have been updated G thinks the links on those pages are also new now and must mature before target pages get benefit (again though, what about the new pages)
* etc etc...
Basically more questions than answers at this point!
What we have found is that any new page comes out with a PR of 0 causing the site to now be linking out to 10% of its total, to pages that have a PR of 0. In this situation we did not drop off the map, but dipped considerable about 20 places. What complicates the matter is the sandbox, now those new sites are stuck with PR 0 for a long time. The recent PR update did get the bars to go green on about 20 new pages we had, and alas most sites moved back up.
What troubles us is we want to expand, and provide new content for our visitors, but in this climate the only way we can do it and maintain our traffic, is to add way slower than we would like to. If we added 50% new pages at once, I am convinced we would fall at least 100 spots in the SERP’s for our main key words. I know this doesn’t help explain the problem, but at least you’re not alone!
We split off four highly specific topics to their own domains roughly four months ago. (All of the domains are still showing PR0 for every page.)
After that, we added quite a few new pages (probably 15% or so in terms of total pages) to the remaining topics. We don't do anything weird, and the vast majority of the pages are content pages with only Adsense on them. We've basically disappeared from Google, but interestingly enough, almost all of the other search engines think that the content is useful and rank us fairly decently.
The only thing that we think could possibly explain the disappearance is that new pages are allocated a PR0 and a substantial increase in PR0 pages must trigger some sort of lowering in importance of the index page, even though the Public Page Rank shows as PR5/6 on all of the disappeared pages.
I don't have any proof of this, of course, but it would explain the phenomenon perfectly.
This might actually be an improvement to the algo. It would reduce the effectiveness of link selling, blog spamming and throw-away domains, as well as many affiliate sites. The BBC's of the world would be largely unaffected because their growth is usually not extreme in relation to their site size.
God/Google help anyone who makes a major change or addition to a site though.
What if the new pages were to be added with the meta directive robots - NOINDEX. I know this defeats the purpose for adding pages... but would that defeat the problem? Speculating that would actually work, the directive could then be selectively changed over a period of time.
The NOINDEX seems like a good idea for some situations.
BTW, my site grew considerably, but much of the growth was months before I got hit on Sept 22. I suspect that it might take quite a few months for a page/link to mature.
Now I think I've done it again. I had a site I hadn't touched in over a year because of some issues that have nothing to do with SEO or websites for that matter. It was still ranking decently for some keywords in Google, but after reading all the sandboxing discussions I realized this site was not in the sandbox and thought I'd give it a push and see what happens. The site was 40 to 50 pages and I added about 80 dynamic pages to it, all should be spiderable and indexable. I expect that will push it into the sandbox. Here's why I think that:
I'm not going to go into the numbers or the various theories themselves, but I assume the main index is full and that's why there is a supplemental index. It is not pages that are put in the main index but sites. New sites go into the supplemental index. Sites in the main index are going to grow. The main index can only hold a fixed number of pages. Thus, as sites in the main index grow, some must be moved to the supplemental index, since both the main and supplemental indexes hold whole sites, that is, you won't find half a site in the supplemental and half in the main, at least not for more than a few weeks. As sites in the main index grow, which ones are selected for retention and which for migration to the supplemental? Sites in the main index that have the most listings in the SERPs will be selected to remain the in the main index over those that do not have as many SERPs listings. Sites that do not grow in pages are not likely to be subjected for consideration for removal to the supplemental. Sites that do grow in size become subject to consiseration for removal to the supplemental index with those sites adding the least number of pages being prererred to remain in the main index.
Thus, by just tripling this site's size, it just became a likely candidate for being sandboxed.
The reason why I say that the main index is a set of whole sites is that for the purpose of matrix calculations it is much more efficient and the defining points are easier to establish if whole domains are included in a single index.
I can't yet find another explanation for this, but I'm just having trouble believing that adding pages like this really hurts sites badly. There are millions of reason why a site might grow in 10% or 30% increments, rather than arithmatcially and evenly.
I'm hoping some more members will jump in and tell us this is a non-starter...
Maybe it's just a severely tightened dup filter.
Is 'tightened' the same as 'retarded'?
More likely is something to do with the glitch/readjustments on aug 5, aug 25, sept 25 - where sites plunge down the rankings for just one month
some say yes, some say not but there wasn't proof this IS NOT an issue.
also strange GG dont answer many questions as often as before.
saying nothing can 'tell' a lot also ;)
Someone else said this and I will echo it. Whether google is the cause or not, it breeds fear. Too many links, not enough links, link placement, anchor text wording, and now....the fear of adding content? Insane.
Now you think twice about adding new pages to sites that are 4 years old!
also strange GG dont answer many questions as often as before
Not strange at all, since google is now a publicly traded company.
How would you like to be the one blamed for a 10% drop in stock price for a comment you made on a public forum? Thta might not soudn like a big deal, but to people (including many google employees) who own millions of share, that is big bucks.
No...now that the google (most) of the google employees have a vested interest in the highest stock price possible, I doubt we'll be hearing much "inside" information anymore (not that we got much in the first place).
Google simply has too much interest in keeping as much as possible about it's systems secrets, becuase:
1. The want a high stock price
2. Their "enemies" are gunning for them
3. The spam problem has grown out of their control, but to admit it would be harmful to their interests (see #1)
4. [theory] : Their database has reached it's limit, and they are frantically trying to develop one that can scale further (when was the last time the "Searching 4,285,199,774 web pages" was changed?)
