We’ll assume the plaintiffs in this WebmasterWorld thread know the difference between art and pornography and have enough judgment to understand what’s PG-13 and what’s R. That’s a smart-aleck way of saying we’ll give them the benefit of the doubt that their images are suitable for sensitive eyes and that Google Images is, actually, in the wrong.

Google Images is set to moderate by default, but users can be more Nazarene about it if they like or can let it all hang out by removing the filters altogether. If you’re a parent—or even just not all that pervy—then you know filters are good sometimes, just like on Camels.

The filters aren’t perfect. Sometimes they block legitimate images for a search query. The BBC’s recent linkbait headline, for example, informs the uncultured that great tits aren’t always what you think they are. Sometimes they’re birds, but don’t ask Google to show you. Safe search omits that one word entirely and unfiltered search brings back, well, what everybody would guess it brings back. To see the birds, you may have to go through Wikipedia.

There may be no hope of a nest of great tits (the birds, dammit) getting into Google’s image index at all. But for others who have misunderstandings with Google, there are things you can try.

SearchEngineRoundtable’s Barry Schwartz summarized the advice in the WebmasterWorld thread, which includes making separate folders, editing your text so that there are no naughty or could-be-construed as naughty words, checking the neighborhoods of outbound links, and other guilt-by-association aspects.  

 Google Korea Sex Filter

Users must provide their name and the Korean version of a social security number in order to see unfiltered results for some searches.  Otherwise, they’re liable to see a list of results, but also a yellow information box and the number 19 – that’s the age cutoff – circled in red.

The potential for breaches of privacy here seems huge, but Ionut Alex Chitu reports, "Apparently, this information is not saved on Google’s servers and the result of this verification can be saved to a Google account."  Also, "Most other sites that operate in South Korea use age verification . . ."

We’re still not comfortable with the idea, but to avoid belaboring the subject, we’ll at least grant that the filters seem somewhat effective.  No yellow boxes appeared due to a search for "breast cancer," for example, and although one did interrupt Chitu’s test search for "sexual harassment," the relevant Wikipedia entry was still plainly visible.

Google seems willing to go to great lengths to accommodate the Korean market.  In addition to the age verification, Google Korea’s homepage has already been redesigned with cartoon-like moving images, and more significant changes are supposedly in store for YouTube Korea.

The filtering project was part of the governments $189 million NetAlert program launched in August 2007 to protect minors from online sexual predators and to block adult content.

It was projected that 2.5 million households would use the free porn blocking filters within 12 months but only 144,088 have been downloaded or ordered on CD-ROM since August of last year. The Department of Broadband, Communications and the Digital Economy estimated 29,000 filters were being used, less than 2 percent of the set goal.

 Federal Communications Minister Stephen Conroy said," The program has clearly failed, despite over $15 million being spent in advertising to support it. Labor has always said that PC filtering is not a stand-alone solution to protecting children from online dangers."

"The Government has a comprehensive cyber-safety plan that includes the
implementation of mandatory ISP-based filtering to deliver a filtered feed to all homes, schools and public Internet points. Education for parents and teachers as well as children is a priority."

Opposition communications spokesman, Bruce Billson said the government was too quick to criticize  the NetAlert program ." NetAlert is a program which is relatively new, as is the minister in his role, and I’m sure he would like a little more than six months or so before the public decide if he has been a failure or not," he said.

"Proper supervision should be front and centre of any efforts to protect children from inappropriate material on the Internet; supported by additional tools such as content filters, not some mandatory and ill-conceived ‘clean feed’ measure by a government that believes only it has the authority to decide what’s appropriate or inappropriate content for computer users," Billson told the Sun-Herald.

They damage rankings and ultimately may result in banishment from the search engines. However, before panic sets in, let us guide you through the basics of penalties and filters.

A penalty is caused by significant violations of a search engine’s website guidelines, such as:

Penalties can be issued after a person reviews a website or after it has been crawled and processed by search engines. They result in a website being heavily held back in the rankings, or removed from the search engine’s index entirely. A penalty is called a ban when a website is completely removed from a search engine.

To remove a penalty, a site needs to first correct the problem that caused it, then contact the search engine and request a reinclusion. It is important to note penalties are not common for business websites, particularly in Google.

1. In Google, this is done through a Webmaster Tools account, which can be set up for free.
    
2. MSN has just created a Webmaster Tools system like Google’s.
    
3. Yahoo does not have a defined reinclusion process.
    
4. A filter is caused by passing a search engine’s threshold setting for one or more optimization/link building elements, such as:
    
   1. Too many keyword mentions on a page’s body content (over-optimization).
       
   2. Too much keyword blurring between a site’s pages.
       
   3. Too many links with the same anchor text.
       
   4. Too much keyword-rich internal linking (can cause 950 filter).
       
   5. Too much link building in a short period of time.
       
   6. Too much link building using the same anchor text.
       
   7. Link building in bad neighborhoods.

Filters are common. They are issued automatically after the site is crawled and processed by search engines and result in a site being held back in the rankings.

1. Filters can be keyword-based or site-based.
    
2. Filters can be mild (held back a few positions) or heavy (held back hundreds of positions).
    
3. Filters can have a time element (like the normal Google Sandbox process, where a site is initially held back many positions and over time gets held back less and less until eventually it ranks near its allinanchor rankings).

In general, to remove a filter a site needs to first correct the problem that caused it, then wait for the search engine (s) to crawl the site again and find the corrections. The next time the search engine updates its rankings with the corrected data, the filter will be lifted automatically.

1. In some cases, like the Google Sandbox, you can simply outwait a filter. Websites commonly spend anywhere between a few months and a year or so in the Google Sandbox. The time websites spend in the Sandbox has significantly decreased over the last few years.
    
2. In some cases, you can remove a filter on a site by doing things that search engines like (such as getting quality links to the site from other respected websites in the same sector) to outweigh the things about the site or optimization they don’t like.

Filters are common, especially in Google.

1. The Google Sandbox is technically a filter. Google closely examines new websites for over-optimization to try and minimize spammy websites filling its SERPS. As a result, newer websites often trip filters when they start an optimization campaign using traditional SEO (lots of keywords on the page, keyword-heavy titles and description, keyword heavy anchor text in incoming links). This pattern of new websites getting filtered and eventually getting released from the filter is called the Google Sandbox. Websites can speed up their release date from the Sandbox by getting quality internal links and not going overboard with on-site optimization.
    
2. New websites are not the only targets. Older websites can trip filters when they go overboard with over-optimization

Being aware of how search engines assess penalties and filters is essential to avoiding them. Do not try to cheat the system, over-optimize or trick search engines. They are savvy to these tactics and punish those who attempt to take advantage of them. What may help in the short term will only end up hurting in the long term.

Being competitive in the SERP’s is important, and sitting idly is not a good strategy. If you must build links or optimize your website try to stick to the guidelines and use common sense.

Comment

On the Official Gmail Blog, Robby Stein outlines a way to have colored labels cooperate with filters.  The colored labels were released about two weeks ago, so the explanation represents reasonably quick work on his part.

Stein begins, "To set up a filter with a colored label, simply click the ‘Create a filter’ link next to the search box.  Add senders or certain words you want to keep a better eye on, click next, and assign a label by checking ‘apply the label’ and choosing an appropriate one.  Then just pick a label color by clicking the color swatch next to the label title in the left-hand navigation menu."

As a writer, I’m inclined to point out that a quick scan of your inbox should have the same end result.  Still, there’s no doubt that color recognition occurs quickly, so as long as the associations make sense to each user, this could become a handy time-saver.

It will, at least, make Gmail more bright, and just in time for Christmas, too.

• The Google advanced search page allows you to search for pages that were recently indexed, letting you filter through days, weeks, months, and years. Here are pages from SeoBook.com indexed in the last week.
• In the URL they place qdr=w qdr=d (day) qdr=w (week) qdr=m (month) qdr=y (year). You can also search for multiples of these units, like search for pages indexed in the last 2 weeks by placing qdr=w2 in the URL string.
• If you change your content management system or add new sections to your site you can see how quickly they are getting indexed, and look for any duplicate content issues as the new pages are getting indexed by looking for pages indexed under multiple URLs.
• If you have never checked your site for duplicate content issues, but recently published content, that might also show any content duplication issues or Google indexed pages that you do not want in Google’s index.
• In addition to using date based filters to find how well your site is getting indexed, you can search to see who is mentioning an idea with a footprint, or use date based filters for doing link research.

Comments

Tag:

Flickr has one of the most vocal user bases I’ve ever seen – a single thread discussing the site’s filters/censorship received almost 5,000 replies before it was locked.  (The c-word was not at work here; the conversation was just redirected to a fresh thread.)   Flickr’s staff is also good at communicating its thoughts, and, after a tip from the DPA, it’s one of those employees who we have to thank for the following explanation.

“As of a short time ago, we changed the way the content filter setting works for German members, allowing them to turn SafeSearch off to allow photos flagged as ‘moderate,’” writes “Stewart.”

Yet, “In Germany, local law (Jugendmedien-Staatsvertrag JMStV) requires stringent age verification in order to display online content that could be considered harmful to minors,” so, “[w]e are still limiting access for users in Germany to the ‘restricted’ category on Flickr, which applies to pictures not considered appropriate for kids and teenagers according to local law.”

The end result of all this, according to Stewart?  “At this stage, pictures rated as ‘restricted’ can only be uploaded and viewed privately, but not displayed publicly.”

That’s an improvement, but you’ll note that there is no mention of Singapore, Hong Kong, or Korea – the other places in which Flickr’s filters have been installed.  Flickr and Yahoo still have a ways to go before that c-word is out of sight.

Flickr Says No To German Photos

Flickr, which is owned by Yahoo, is using the same defense its parent company used regarding China: Don’t blame us, we’re just obeying local laws.  And, as you might have guessed, Flickr’s blocks are in effect in China.  Korea, too.  Yet, according to the LAist’s Malingering, some unexpected examples of filtering have occurred.

“Flickr’s filtering system has now prevented users from Singapore, Germany, Hong Kong or Korea from viewing any photos considered ‘unsafe,’” she writes.  “Compounding this problem is the fact that the filter system has so many glitches in it, that even photos of dogs and kittens can be considered ‘restricted content.’”

A bit of humorous commentary then follows: “Oh Germany.  Here we tried so hard to get past the whole ‘Hitler’ thing and then we were so proud of you for tearing down your big ol’ wall, but now you come up with this and we have to be worried all over again.  You can’t spit in public in Singapore, so they’re obviously more easily offended, and we won’t bug Korea because they might nuke us, but come on, Germany.  Really.”

Those countries’ governments do indeed deserve part of the blame.  As for how whether Flickr should be held responsible for this mess . . . opinions vary.  One thread in the Flickr Forum has over 2,000 responses, some of which are very energetic.  But Flickr representatives have assured users that the company is working to make its censorship and/or filtering less intrusive.