Rankings drop after changing the Meta Description

[benallos]

I have a situation that I have encountered on my site. One of the high PR pages on my site (I shall call it page4.aspx) has been ranked well in Google SERPs. The keywords that is pointing in page4.aspx is positioned in 3rd / 4th place of Google SERPs. I made a move to change the Meta Description of page4.aspx. I took out 1 word and changed the spelling of the other word of the Meta Description. After the revision of the Meta description, the rankings of the keywords that is pointing to page4.aspx has dropped to page 3 and page 4 of Google SERPs (from position 3 to position 23 and 34). Did it drop due to the revision of the meta description? Is the dropping of the keyword rankings will take a long time to recover or will it come back up again?

I change the description last September 28, 2010 and the dropping of keywords took place in October 5 or 6, 2010. Right now the keywords has not recovered yet.

[tedster]

The timing is suggestive, but it's very unlikely to be a cause & effect relationship. Google people have repeatedly explained that they don't use the meta description as a ranking factor.

[Lovejoy]

"The timing is suggestive, but it's very unlikely to be a cause & effect relationship. Google people have repeatedly explained that they don't use the meta description as a ranking factor".

That's what they say, but I've noticed the same thing as well. As a test I've made minor changes in mine within the last year and went off the first page. Changing it back to what it was put me back where I was by the next update.....

[tedster]

As Schultz would have said, "Verrry interrresting". I'm all ears for more reports.

It is difficult to isolate just one ranking factor these days, but I certainly don't rule out anything just because Google spokespeople said something in the past. If anyone has more anecdotal evidence, please share.

And if you've done some rigorous testing, we'd love to hear the details of how you tried to rule out other factors.

[Lovejoy]

I think without a lot of very detailed testing and a complete insider knowledge of Google, my testing would have to be considered anecdotal. I just know what I've experienced in my own sliver of the internet.

Its equivalent would be like " Old Wive Tales" that were once laughed at, such as "Blueberries being good for your eyes" or " Fish is brain food" but that we now know to be quite true.....

[scottsonline]

IMO it probably had little to do with the meta change but instead ANY change. Right now content changes are a killer. Try changing one of your lesser pages and watch it drop like a rock. Reaffirms the idea that "new" is bad right now.

My take on it is they revalued all our sites. Old content got old rankings. New content is ranked based on the new site valuation.

[HuskyPup]

Google people have repeatedly explained that they don't use the meta description as a ranking factor.

I'm not so sure about that:-)

I did a massive rebuild of a 10 year old site over the summer and one of the many things I concentrated upon was ensuring really good meta descriptions. All the pages which had slipped a little down the first page have now returned to their former top 1-3 positions.

Ok, I did do everything according to my bible therefore it could be the h1, it could be me cleaning up code, it could be adding synonyms, it could be adding further tech info, it could be more accurate image alt and title, it could be whatever, the point is that all these pages do now show that meta description in the Google SERPs which is repeated ad verbatim futher down the actual page.

I feel I'm right however cannot conclusively prove it!

[tedster]

But you are in a good position to TEST the idea, Husky - if you want to risk one of your ranking pages, that is. The OP made only two small tweaks: "I took out 1 word and changed the spelling of the other word."

[Lovejoy]

How about you try it Tedster? we already know it works ;~)))))))

[tedster]

Here's why not - I already know that Google will completely ignore the meta description and replace it with their own snippet under certain conditions and queries. I also know that URLs can rank #1 without any meta description at all.

So that, for me, means that the spokespeople at Google have been telling it straight. Now I'm not so foolish as to ignore good data if someone has it. But in this case, I'm using my resources in other directions.

[HuskyPup]

But you are in a good position to TEST the idea, Husky

Yep, ain't doing that after all my images on specific sites having been dropped from their image SERPs for some unknown reason!

@benallos - What were the two words if you can tell us?

I have seen significant changes when adding or deleting to the US Google.com SERPs specifically with American spelling V English spelling simply for ONE critical word.

For this one word I tend to dominate non-US Google.com and ccTLD results but for US Google.com the first page is usually like a yo yo for all the sites!

[scottsonline]

@Tedster

I think even a meta description change will return a "page has changed" 304 I think. Is that enough to trigger google into thinking it needs to revalue a page?

[tedster]

That sounds like a Traffic Shaping [webmasterworld.com] trigger, especially since geographic targeting is one of the big profiling classifications. Did you make those spelling changes only in the meta description?

[Lovejoy]

Just going from a singular to plural , eg "blue widget" to "blue widgets" can take you from #1 to # 8 for my main keywords....

[Lovejoy]

True enough tedster, I've seen sites that rank on page one that consist of an image map, but if someone makes no changes to their page than in the meta tags, and lost positions that return after reverting to the old tags, and none of their near competitors have made any changes, what other conclusion can you come to? Personally I take anything Google says under advisement, but with a large pinch of salt.

[tedster]

I don't know what to make of it now - there's not enough data. I certainly don't want anyone to assume a firm conclusion from one data point. That's how SEO myths get started, and then take years to die away.

[HuskyPup]

there's not enough data.

I agree and especially so if one has a proven method for constructing pages, I hesitate to say SEOing since I have constructed pages in the same way since the mid 90s however as more and more competition has come along each new site has been devloped using the most up-to-date proven techniques which is precisely why I re-constructed the 10 year old site.

I do still complete both keywords and description tags just "in case" they may have a little influence or if one of the majors decide to use them again...plus it is an excellent way to learn how to naturally descriptively phrase a specific widget, it beats machine code and you never end up with some of those hilarious ebay "Buy your dead bodies here" snippets.

[scottsonline]

I have one page that's indexed every few days. It's no longer sold so I don't care about it. I just changed the meta desriotion reversing the order of two words. Will report back in 1-2 days. I'm guessig it drops to the fifth spot.

[benallos]

"@benallos - What were the two words if you can tell us? "

@HuskyPup

I remove the word "good" and changed personalized to "personalised"