| 9:03 pm on Jun 2, 2009 (gmt 0)|
I'm really surprised that your editing of the page could have caused a rankings change "literally within the hour". For that to happen, the page would have to be spidered and the new information "digested" and then the rankings updated in at least one data center. Maybe it's possible, but I'm skeptical.
| 9:05 pm on Jun 2, 2009 (gmt 0)|
If you changed your meta description and an hour later your rankings changed, it's very likley you are hitting a different data center, not observing the effect of the change.
| 9:08 pm on Jun 2, 2009 (gmt 0)|
It was the meta description. The meta description is always updated very very quickly.
I just checked, and the latest cached version was from today right after I made the change. This might be a coincidence, but you are right, whenever you make a change on your site, google does not update your rankings right as soon as it re-crawls your site (or does it?)
| 9:08 pm on Jun 2, 2009 (gmt 0)|
I certainly agree that Google looks for keyword in different areas of the page - certainly expected. And if the keyword is rare or non-existent on the page, it becomes very difficult for the page to rank on that keyword - but not impossible.
None of that is the same as establishing that "keyword density" - a specific calculation - is a factor in use in the algorithm.
You wrote "one more instance of the word 'widgets' in the meta description tag". Did the keyword already appear in the meta description and you added another occurence?
| 9:14 pm on Jun 2, 2009 (gmt 0)|
"free" and "widgets" already appeared in the meta description, but not as "free widgets", they appear as "free widgets" in the title tag.
I added "free widgets" to the additional meta description, changing nothing else.
| 9:51 pm on Jun 2, 2009 (gmt 0)|
What was the date shown in the cache, compared to when the change showed up?
| 11:30 pm on Jun 2, 2009 (gmt 0)|
Though I'm not saying you're wrong, it sounds doubtful that you'd be able to test the effects of adding keywords to metas that quickly & easily. A system that was that responsive/testable would be very vulnerable to exploitation. Though the SERPS are being exploited, that level of directness just doesn't seem to fit with Gs overall black box approach
| 4:39 am on Jun 3, 2009 (gmt 0)|
Were you logged into your Google account while making the search?
| 8:25 am on Jun 3, 2009 (gmt 0)|
|And if the keyword is rare or non-existent on the page, it becomes very difficult for the page to rank on that keyword - but not impossible. |
This may be true but Dell UK ranks very high on a search for the word "computers" but this word does not appear anywhere on their home page. They seem to be ranking from the weight of their inbound links containing the word computers.
I would be very sceptical about any claim that a site's rank changed within an hour of making a change to the meta description. Bear in mind that Matt Cutts has gone on record to say that the meta description is not used in ranking.
| 8:32 am on Jun 3, 2009 (gmt 0)|
Yes, Dell is a classic example that it is "possible" to rank for a keyword that is not on the page.
When a search result does not contain the searched for term on the page, if you click on the Cached link, you will see the explanation, up toward the top:
|These terms only appear in links pointing to this page: [keyword list] |
| 6:28 pm on Jun 3, 2009 (gmt 0)|
I understand the weirdness of it. It could very well be a coincidence that I gained ranking right after I made the change to the meta description.
Last year for 3 months I ranked #1 for this search term, after a site design and some changes, I dropped to the 7-10 area. I had added more content, and added no more instances of the words "free" or "widgets" on the page, thus reducing my keyword density to less than 1% for these 2 words.
I have been doing SEO for over 10 years now, I make a living off of it, I study my site and competitors every day, and no matter how much I learn or think I know, google always makes adjustments.
I know for a FACT a page can rank for a word that does not exist on a page, I have a few sites like this.
Needless to say, I am now #3 on my computer at work on both FF and IE (logged in gmail and logged out). I am also #3 on my home computer, my iphone (3g connected), and the 3 friends I asked.
I made no on site changes the last month, no linkbuilding nothing, so I am left to assume that this change caused my boost.
| 7:32 am on Jun 5, 2009 (gmt 0)|
I've seen the same thing happen that is being reported by Brinked. However, I find that the ranking increase is usually short-lived and you'll see the site fall back to at or near where it was ranked before.
| 8:39 am on Jun 5, 2009 (gmt 0)|
You tweak, they tweak, we all are a tweaking... and chasing tails. Long or short... (grin)
Seriously, most of these changes are short-lived. Not to say one shouldn't make the effort, but there is a point of diminishing returns and SEO and GOOGLE are each head-butting toward the point.
| 4:36 pm on Jun 8, 2009 (gmt 0)|
maddawg is correct,
a few days later, i dropped to the 3rd page for these 2 terms. I undid these changes and now i am going back and forth from the first page to the 3rd page for these 2 search terms daily. Half the day I seem to be on the first page, and half the day I am on the 3rd page.
I also notice, on these 2 search terms, my site doesnt have the "cached" link in google, but for all other searches I have the "cached" link.
The experement was worth it, hopefully my rankings will be restored fully. If nothing else, I have learned something new.
This is why I only make a very small amount of changes per month..in case of anything I can go back and have a better understanding of what I did.
| 5:00 pm on Jun 8, 2009 (gmt 0)|
brinked, did you by any chance run any press releases for the site? Did some blogs run posts about your site recently?
The bounce you describe is similar to what'd happen if you acquired a few links. The bouncing around is similar to what happens when one obtain links. What is happening is you're connecting to different DC's and the DC's are showing you a different ranking calculation.
| 6:36 pm on Jun 8, 2009 (gmt 0)|
No press releases, no spike in backlink activity.
This is not a case of me connecting to different DC's. I can see it in my traffic stats.
Whats weird is when using thee query as such: [google.com...]
I come up at #3 across the board on google.com and other datacenters. However when I use: [google.com...]
I come up on the 4th page