An Offset Content Penalty?
I want to do the right thing regarding accessibility and usability, but I don’t want to be viewed as some unethical search whore in the process.
Google handles search abuse reports on sites by compiling the submitted and discovered data, then adjusting their search algorithms to counter the identified abuse method next time around. This is an effective method of dealing with abuse and violations it seems — the most practical method, anyway. As an example, abuse such as using the style sheet display property “none” to hide a slew of links was reported to or discovered by Google, and now their algorithms can identify this type of index-inducing violation and respond accordingly.
But, like most things, it may be an imperfect model, especially when it comes to methods of enhancing accessibility or usability that may be in themselves a violation if viewed out of context. This could be the case for the offset class (or off-screen as some call it).
Now, I’ve written more than a couple of articles regarding what I feel are clever, practical uses of the “offset class,” for example:
- Hidden Focal Navigation (November 3rd, 2005)
- Offset Class Jump Links (August 13th, 2006)
- Enhancing Accessibility with Offsets, Cautiously (February 25th, 2007)
But do these methods carry risk? Risk of not being indexed by Google or other major search engines, that is? Google doesn’t like hidden links and will react accordingly to display:none, but how about using positioning to handle links — like my offset class jump links, for instance? If algorithms are adjusted to see this technique as a red flag and de-index a site, where’s that leave our honest efforts to enhance accessibility or usability? Since using offset jump links is controversial, it might be best to ensure they’re visible (but there are other uses that are more difficult to get around). This might be a sound move if search engines don’t take context or motivation into consideration before pulling the plug. If a search algorithm is adjusted to flag-n-bag sites that use this technique, how will they be able to tell the difference between legitimate uses and abusive ones?
I use this method, for jump links and more, depending on the situation. For instance, I’m working on a massive site right now — which is why I haven’t posted in a bit — and I have not only 4 jump links hidden by the offset class, but I have a snippet of text on the home page. This is what I wrote:
<p>Use the menu to the right<span class="offset">, or after the content if you don't support style sheets,</span> to get around this site.</p>
The idea here, is to provide this sentence to sighted, style sheet-supporting users:
“Use the menu to the right to get around this site.”
Other users, however, will get this sentence:
“Use the menu to the right, or after the content if you don’t support style sheets, to get around this site.”
Doing this ensures the sentence is accurate regardless of user capabilities. But can I get busted for doing this, treated as a “violator” by some algorithm that doesn’t (or can’t) consider the usage context? I’m concerned.
Even display:none can be beneficial in some cases. I had one site that had a style changer. I used display:none to address visual users that didn’t support styles. It basically said that if they can read this then they might want to try setting a style on page-X. I didn’t want that line available to screen reader users so I used display:none to hide it from them and visual users.
I want to do the right thing regarding accessibility and usability, but I don’t want to be viewed as some unethical search whore in the process. Anyone have more information? Stuff that might make me worry more, or hopefully put my mind to rest. Can a solid accessibility practice cross the line to the point that it might create indexing problems?
Your thoughts?
Joe Dolson responds:
Posted: June 14th, 2007 at 5:08 pm →
One thing I’ve heard fairly consistently from search engine reps is that the biggest test is whether what you’re doing is good for the user: “[build] for the user, not for the search engines.” They pretty much are concerned with intent. Of course, this won’t make you feel much better when you wake up and find yourself banned or penalized in Google!
However, it does mean that when you file a reconsideration report, they’ll probably honor your request, if you can show a sensible reason for your actions.
What it does, essentially, is bring you closer to a “gray area” in the study. It isn’t, in itself, a reason to be penalized: but, in combination with some kind of keyword stuffing, it’s a different story.
Interpreting that a bit more broadly, I’d hope that striking the word “visual” would still allow the statement to hold true.
John Faulds responds:
Posted: June 14th, 2007 at 8:50 pm →
It was my understanding that Google only read in-page styles and not external stylesheets. Additionally, you can always add your CSS directory to your robots.txt to disallow bots from viewing it.
atthijs responds:
Posted: June 15th, 2007 at 3:19 am →
As Joe says, if you build your site for your users, you should be fine. I cannot imagine Google (or other engines) penalizing sites with a hidden skip-link. What about hidden text from drop-down menu’s? Hidden text from collapsable items (question and answer lists for example)? Hidden text from image replacement techniques?
There are so many cases around. All the web standard guru’s have written tons of articles about these things and how they are beneficial.
If Google would automatically start penalizing for that, 90% of the modern websites build with web standards would disappear from the search results. That wouldn’t be so smart for the users of google itself wouldn’t it?
Harmen Janssen responds:
Posted: June 15th, 2007 at 4:29 am →
I can see a situation in which Google has to penalize these kind of techniques, but my first reaction is “wow, harsh”.
As you describe above, there are perfectly valid uses for
display: noneand it feels just plain rude that Google interferes with you trying to do the right thing.Robert Wellock responds:
Posted: June 15th, 2007 at 7:16 am →
As long as it only considers the inline CSS rather than a *.css file then that it is not too much of an issue really to be honest.
Stevie D responds:
Posted: June 15th, 2007 at 7:46 am →
Is it the case that Google penalises pages that have hidden text, or that it ignores any text that’s set to
display:none?I think it would be reasonable to expect that Google looks at, and indexes on the base of, any content that is visible under the default stylesheet(s). Anything that is, by default, hidden or offset could/should be ignored. It’s not without problems, but it seems to be the most sensible solution. Unfortunately, we have to accept some compromises in the battle against spammers if we want efficient, automatic indexing agents.
Skip links would be ignored by indexing agents, but that won’t be a problem. Nested “hover” menus, if they’re written using
display:none, wouldn’t be indexed either, which would reduce the indexing of your site.Outright penalising anyone who uses these authoring techniques seems an unfair and unreasonable response, given that there are many legitimate uses for them.
Mike Cherim responds:
Posted: June 15th, 2007 at 9:29 am →
You guys are right, as I would think myself. I would fully expect to see Google apply common-sense before taking any de-indexing action. This all came about from a discussion I was having with John Pozadzides, of HTMLHelp fame. John felt that any such action would bring down the razor sharp ax of Google fast and hard. I argued that Google wouldn’t do anything if the intention wasn’t “to spam,” but John felt it was risky nonetheless. Even the offset “honeypot” input on my contact form brought about concern over my methods. I headed over to Google to read their policy on the matter where they state they do regard intent and purpose. But I’ve also read at Google (can’t find it now) where they deal with a lot of this stuff not on an individual basis citing they don’t have time. In that bit of text they stated that if they get complaints about specific abuses they take action only by tweaking the search algorithm. The onus, apparently, if the algorithm fails a site, is on the the site owner to argue the action. That did leave me wondering how they deal with this one. The term “Guilty before proven innocent” came to mind.
John and Robert are probably right in that they won’t, or can’t, correlate styles (my styles directory is listed on my robots.txt file) applied externally and can only key in on inline CSS. That makes a lot of sense. My discussion with John (P.) about it did leave me feeling uncomfortable and wondering.
Anthony Brewitt responds:
Posted: June 22nd, 2007 at 8:13 am →
- seems to be a word that has recently escaped a lot of authorities in our domain Google included. Search Engines need to employ some common sense into there complicate algorithms and start thinking in human terms. So I guess I have nothing new to add - oops!
John Lascurettes responds:
Posted: June 28th, 2007 at 10:35 pm →
As you pointed out, once Google sees abuse cases it adds them to the algorithm. I’m sure the actual offset abuse cases are stuffed full of fluff and bloat (the same way keywords and same-colored text on a background were back in the day). There’s a huge difference between hundreds of fluff and repeated words and a few skip links or a couple of phrases for descriptions. That must come into play and I think your techniques would be relatively safe.
Of course, I am just guessing.
On top of that, in the age of AJAXy “goodness”, there are many items that are hidden until the user interacts with them. I’m sure Google isn’t penalizing those just because items are hidden.
webecho responds:
Posted: June 30th, 2007 at 12:21 am →
It would seem that it’s not an issue Mike.
http://www.seroundtable.com/archives/013790.html
See Matt Cutts’ comment.
Mike Cherim responds:
Posted: June 30th, 2007 at 12:45 am →
Awesome. I was pretty sure common-sense would be employed when it came to this stuff.