Forum Moderators: Robert Charlton & goodroi
I use the following Supplemental test:
-site:www.example.com/* site:www.example.com
At least for the site in question, this test closely matches supplemental results from the past when supplementals were reported. The page I discovered the gotcha on is no longer reported by this test. Unfortunately the shortcut search site:www.example.com/& no longer works. It's probably likely the trick above will stop working too, in the near future.
One thing not mentioned much as a possible reason for a supplemental placement may be keyword frequency in searches, perhaps even seasonal keyword frequency. If a page should rank well for a search like KeywordA KeywordB KeywordC, if these Keywords are not frequently searched for perhaps this page becomes supplemental.
The page in question was designed (to a limited extent) to have three specific keywords produce a high ranking in the SERPS, and this page does now rank very well for these three keywords (Hooray).
This page had very substantial unique content and had similar page rank to other pages from this site that do rank well in the "normal" index.
The problem! There was a fourth, hidden keyword
In fact this keyword had more occurrences on the page than any other. What was the keyword in question "bullet". When I discovered this I learned very quickly that several of this site's pages that were supplemental actually ranked in the SERPS for their designed keywords and the word "bullet". But I never put the word "bullet" anywhere on the pages! But what did the page have, massive quantities of "bullets". This page had large itemized lists, each paragraph having a bullet. And guess what Frontpage themes do, they embed the Alternate Text of "bullet" anywhere on the page where there was a "bullet" graphic.
When you searched with Google for the keyword "bullet", Google will show your Alternate Text in the SERPS. So for a year, as far as Google was concerned, the primary keyword for this page was bullet, followed by the designed keywords KeywordA, KeywordB, KeywordC.
So how frequently would anybody search for:
Bullet, KeywordA, KeywordB, KeywordC? Of course virtually never so guess what SUPPLEMENTAL INDEX.
Don't get me wrong I do have pages with "bullets" that are not supplementals. I'm sure there are numerous other factors weighed by Google. I have pages with "bullets" that rank number one in the SERPS for their keywords. What I can say is these are very popular pages, frequently searched for, and I believe my visitors stay at these pages for a long time. (I think they actually read them!). The page in question did have many bullets from top to bottom, probably 50 or 60!
So on this page I got rid of the bullet graphics altogether, thereby eliminating the "hidden" keyword. A little more than a month later, the page now ranks #2 in the SERPS for it's proper keywords. For a few days it still showed up as supplemental, but now it is no longer indicated as a Supplemental page.
I can't knock Frontpage too much. It would be just as easy to make this mistake, perhaps massively, with CSS especially if you have design teams building multiple Cascading Style Sheets. So watch you alternate text and even titles on those style sheets. (For now embedded styles only would cause problems, but who knows when Google will start reading Style Sheets. They are easier to understand than a web page's actual content!)
Frontpage was particularly nasty about hiding the Alternate Text "bullet". Believe it or not "bullet" does not show up in the source code in Frontpage even if you Preview the page and View Source. Bullet does not show up in the source even if you Preview the page in a browser. Only the source published to the server has the ALT text "bullet" in it!
BUT, bullet does show up in the source once you publish the website
So the Alternate Text "bullet" is truly well hidden from the website designer! ( I know many hate Front Page, but for me it's benefits truly outweigh the flaws).
So now I'm done experimenting and I'm pulling all the bullet graphics out of my pages. Coincidentally I was going to do this anyway with CSS. Wow, had I not experimented I might have thought that CSS itself improved my web page's ranking drastically! I do expect some of the pages to remain supplemental even after this fix, there are clearly many reasons for supplementals.
So, webmasters, start checking for all those hidden keywords in those many, many, supplemental pages. This could certainly be a reason for most of a site becoming supplemental.
One thing not mentioned much as a possible reason for a supplemental placement may be keyword frequency in searches, perhaps even seasonal keyword frequency. If a page should rank well for a search like KeywordA KeywordB KeywordC, if these Keywords are not frequently searched for perhaps this page becomes supplemental.
Yes, I'm beginning to think that the search frequency of a page's primary keywords is a factor in whether that page is supplemental or not. Just before they removed the supplemental index indicator from the search results, searches for more obscure keywords (like some scientific names) were beginning to show mainly supplemental results on the first page. Even pages from large, high authority government or university sites for these obscure keywords were being listed as supplemental, which was something that I had not really seen before.
What's annoying is that pages in the supplemental index now often have their images removed from Google's image search, so an image search for obscure or highly specialized keywords sometimes only returns random, off-topic images from index or directory-type pages that have more popular (and generic) primary keywords and just happen to list the obscure keywords once somewhere on the page. Many obscure keyword images that I remember seeing in the image search results in the past are no longer there, but when I navigate to the sites where I remembered seeing those images, there they still are.
When you searched with Google for the keyword "bullet", Google will show your Alternate Text in the SERPS.
Since keyword-containing alt text for very small images may look suspiciously like a spam technique to search engines, I searched my old sites for any of these images and replaced "keyword" with "", which still validates (I didn't know that when I created these pages years ago).
Could the 2 be related?
Frontpage webmasters and perhaps even embedded CSS should search there site for "bullet".
site:example.com bullet. I haven't found a search yet the singles out just Frontpage sites.
And along that line of thought, does anyone know of a search engine that truly indexes the text or source code of web pages. This would be a truly useful search engine, in many ways. I've seen javascript search engines, but none that seem to successfully index all the source of a web sites pages.
[webmasterworld.com...]
You should have Alternate text for images...Conforming to the standards strictly you really should add this bloat to every bullet graphic.
This is not the case - only informational images should have alt text, decorative images should have alt="", as you suggest above. IOW, every image requires an alt attribute, but not every image requires alt text.
The WAI guidelines make this point quite clearly: provide a text equivalent for visual content.
The use of the appropriate structural markup (i.e. unordered lists), with the bullet image specified in CSS, would be an even better solution.