Forum Moderators: Robert Charlton & goodroi

Message Too Old, No Replies

Possible Explanation for Dec/Jan SERP loss

A simplistic, but accurate observation

         

doughayman

11:54 pm on Jan 14, 2007 (gmt 0)

10+ Year Member



Hi all,

I decided to start a new thread on this, due to my discoveries. I seem to have isolated the problem/issue that has caused my obliteration and subsequent restoration of my keyword rankings, and wanted to share this with all, with the hope that it may help some of you guys. Please keep in mind that this may be unique to me, but I thought it was worth sharing.

First of all, I run about 9 websites, that are all under a single domain, that has been in existence since 1996. I host these sites locally on my home server, via a local broadband provider. Each of my sites is located in a subdirectory, under my domain. For example,

www.domain.com/sub1/home.htm
www.domain.com/sub2/home.htm
.
.
etc.

Timeline of Problems and Resurrection
-------------------------------------
Prior to December 3rd, 2006, I had ranked highly in Google (1st page) for many years for many keyword phrases. Starting on December 3rd, I started seeing a degradation in my rankings -- I saw a gradual trickle downward, although nothing catastrophic. This timeframe seems to be consistent with many others who saw something similar, and I will surmise that this is the data that Google began implementation of an algorithm/filter tweak of some sort. Subsequent to this date, here is a timeframe of events for me:

12/20/2006 - Obliteration from the SERPs
12/30/2006 - Resumption of SERPs to pre-December 20th levels
01/12/2007 - Obliteration from the SERPs again
01/14/2007 - Resumption of SERPs to pre-December 20th levels

Observations
------------
I made minimal, very controlled changes to my websites since 12/20, and believe that I have isolated what caused my problems and "resumption of service". As it turns out, each "outage" was consistent with the highest ranking PR page on my domain being content-changed. Each time my Sitemap was fetched (my sitemap always reflects the latest changes to my sites), and the sitemap reflected this high-ranking PR page change, I was almost instantly obliterated from the SERPs within 12 hours. Each time my SERPs were "restored" it was within 12 hours of this page being spidered by Googlebot. This cycle has now repeated itself in full, twice. Although Googlebot fetches my sitemap at least once a day, my pages are fetched/indexed by Googlebot at indiscriminate times. My site has always been treated by Googlebot like this, and there seems to be no way to regulate how often a given page on my sites is fetched (I've attempted to control this with the "frequency" attribute in Sitemap, to no avail).

From a respected poster on the HighRankings website, I am excerpting the following:

Remember too that Google only has to spider your site, but also spider all of the sites linking to you in order to come up with a score for your ranking potential. So even if they've already re-spidered your site, if you have nothing but lots of lower quality incoming links it may well take days or weeks for them to spider all of those sites and apply them to your score.

The implication of the above quote, is that pages of one's sites that rank high (due to links to it), may involve a period of time, for it to "re-rank". I believe that Google has altered their ways in regard to spidering/indexing/caching as well as Data Refreshing. The combination of all, and the mechanisms that they use to achieve this may be the culprit to the constant thrashing that many of us have been seeing over the last couple of months. Here today, gone tomorrow, back again the next day, etc. This could be the result of high-ranking pages having their content being updated.

Ramifications
-------------
If my hypothesis above is correct, then my dilemma becomes whether or not to content-update my high-ranking pages or not. Not updating will result in stale-content, but frequent updating can certainly mean that I may be out of the SERPs quite a bit.

Homework
--------
To prove the simplistic view that I have posed above, I emplore everyone to try and time-synchronize the events of outage with events from your weblogs. I'd be interested in hearing what others are seeing.

Regards,

Doug

[edited by: tedster at 12:00 am (utc) on Jan. 15, 2007]

mirrornl

12:24 am on Jan 15, 2007 (gmt 0)

10+ Year Member



interesting and very good post
might not update in a while to get some good nights sleep