SearchPhilosophy.org

If you are looking for help jumpstarting your SEO, Paid Search, or other online marketing efforts, look no further than BAMN Marketing. This hotter than hot new agency includes yours truly, and some of the other finest minds in the game today. And our logo rules. Seriously. Not to be missed. BAMN!

I hate comment spammers. Truly hate them. They are the lowest of the low.

Thanks to these jerks we all have to moderate our comments and carefully review new user signups on user contributed sites to try to stop them polluting the community. Once you have a big community of users, they can help with the burden, but for a small site owner it really sucks to have to wade through all the spam posts.

There are some great tools out there to help stop comment spam such as captcha plugins or spam filters like Akismet which work great. However, there should be some better way to punish comment spammers.

It would be great if there was a blacklisting system where you could report spammers and have their sites get reviewed and possibly de-indexed by the major engines. The spam reporting features for Google for example are heavily focused on spotting and reporting spam results in the SERPs. If we could instead report the comment spammers as soon as they start posting, perhaps they could be eliminated before they get any SEO benefit at all.

I use Google analytics all the time, and find it very useful in tracking web user behavior.  Analytics isn’t always the easiest word to type though, and more times that I can count I’ve tried to go to www.google.com/analtyics (or somesuch similar misspelling of the url).  This takes you to one of the lamest things about Google, their terrible 404 page. 

Now Google is supposed to help us all find things we are looking for right?  So why can’t they figure out a simple thing like when I mistype the url?  It seems like they could at least do something like take the url and cycle it through their own search to see if something comes up.  Interestingly if you pass in “/analtyics google” to google you do at least get what you are looking for the in the adwords results.  (Maybe they could use this to generate some adwords revenue from themselves). 

In any case, a company with so much money and the market leadership in searching the internet should figure out some way to better facilitate user navigation of this type.  Google it seems needs to take some time to better find itself.

 

Tiltomo is a new image search site that allows you to search the Flickr database using keywords.  For results that are returned, you have links to find either more results that are similar by Theme, or results that are similar by Color / Texture.  While the text input describes itself as responding to flickr tags, the results you get are not the same results you would get by clicking on the corresponding tags in flickr.  The return order of the images seems to be somewhat randomized as submitting the same query repeatedly produces the same resultset but in a different order each time. 

Right now, the site has 2 databases you can search.  The first database is a 24 hour snapshot of the images recently uploaded to flickr, and is described as the ”General Images” database.  The second database comes from the Catchy Colors group on flickr.  The quality of the images in the Catchy colors database is much more consistent, and I fully expect the site to expand its coverage to further groups soon.

The related Color / Texture search works quite well in finding related photos with tonal similarities to a target image.  After finding a picture of day glow plastic horse toys for example, I submitted a request for related images and was pleased with the results.  Searches based on images with a strongly dominant color are especially effective.  In an open community where photos are free for the taking, this could be a fabulous tool for cobbling together borrowed images to produce a consistent look and feel in something you are building.   

The site is in a Beta period now and I can’t wait to see how it develops.  I’ve already bookmarked this site and will be watching it closely over the coming weeks.

In the Algorithm research session at Search Engine Strategies today several panelists mentioned that search engines might use the length of time one spends on a result page to downgrade its relevance in certain circumstances.  If everyone immediately clicks back to the search engine for example quickly after visiting a page, the engines may take this as an indication of user dissatisfaction with the result.

In a similar vein, it seems like one could track the time between when one clicks a SERP result, and when a corresponding ad is clicked on the destination page as a possible indicator of spam activity.  There are so many pages now that are nothing but lifted content designed to get in the SERPs to earn revenue through adwords or similar ad networks.  If very quickly after the average visitor is sent to page, another ad is clicked, this might identify these ad-centric pages and allow for them to be moderated down or out of the results as appropriate. 

This of course is somewhat limited as the engine must match the ads being served to track the time difference.  Also, there is a conflict of interest in these situations, as the engines may themselves be benefiting from the revenue these sites produce.  In the long run though, getting rid of these sites will improve the user experience, and thus improve the value of the ad channel to potential advertisers.   So even if it’s value is somewhat limited, it might be worth trying to put something like this in place to at least guarantee the integrity of the ad network within the serving engine itself.

There has never been an easier time to get your thoughts out there.  For like $10 a month you can have a website, spout your opinions, and share your wisdom with the world.  More and more people spend time reading things like blogs now, and these independent resources are playing a greater role in informing the culture.  At the same time, this means that there has never been an easier time to spread false information.  For example, put something totally false in your blog like mentioning the recent study linking cooking with Teflon to breast cancer.  Is there any such study?  Not that I know of, but you can say it.  Who really checks all the claims made in what they read?  It would be interesting to have instead of an SEO contest, a LIES contest (Loading Information Errors into Society) where people tried to get the most number of people to repeat their made up nonsense as truth.  This could be tracked by things like media mentions in mainstream journalistic broadcasts, or repetition as truth on unrelated websites.  The dark side of the urban legend is that it doesn’t have to be about something as silly as a dog choking on a finger, it can be a false notion that spreads and gains ground about any topic, serious or otherwise.  So get out there and start fibbing.  You have the power to shape reality.

I think one of the more difficult issues those of us that work in the search marketing field have to deal with is the issue of to what extent our promotion efforts are harming the internet as an open information resource.  Osx86 project wrote a nice article on how hacking contests can lead to a reduction in information sharing for example in a way that impedes technical progress.  The SERP land grab is undoubtedly reducing result quality in many semantic fields as people make more and more crosslinking ad-words sites filled with other peoples RSS info and not much more.  There is money to be made, so ethics aren’t always taken that seriously, however at some point there has to be a kharmic comeupance for spamming your way to the top.  In the end, I worry that all the spam sites out there can lead to a kind of web 3.0 which resembles the gated communities suburban america created in the 80’s and 90’s.  Imagine a total social search opt in web where you have to meet stringent qualifications to enter, and the result set is only a series of approved sites granted access by members.  In a sense this would be great as an effective way to clean up the garbage.  I suppose a safe web - free web hybrid would be possible where you can limit yourself to a trusted set or explore the vast sea of potential junk.  But the more that you scare people into living within their social result sets, the more you impede the rate at which divergent points of view and bodies of knowledge are shared.  So, if you are polluting, all I would say is don’t expect it to work forever, and try to think about what kind of social possibilites you may be undoing in the quest for short term gains. 

Whenever I search for domain names, I am amazed by how many are being held for ransom by speculators.  Is there information anywhere about the number of domain names that are owned versus the number that are active sites?  I would define active as those having some content live that are updated at least once per year, excluding parked pages at domain registraton vendors showing only ad content.

Out of curiousity, I went to the Regex Dictionary to see if there were any words matching the patterns com$, us$, de$, at$ etc. that would give you a del.icio.us type domain name.  Amazingly, almost every domain matching these patterns, like Anxio.us, MorseCo.de, and Obnoxio.us were already registered.    I haven’t started looking for the variations like GreatMorsCo.de and WorldofMorseCo.de yet, but I’m sure there are some people buying that garbage up as well.  Thankfully, some of the sites like Obnoxio.us have real content, and some of the ones like Unico.de are functional sites matching the country the domain extension was intended for.  However, there were tons of unhosted domains, or ones like MorseCo.de, which clearly seems like a speculator site. 

I would love to see regulations that limited the number of unused domain names one could register.  Hopefully people will just stop buying from speculators.  Until then, I’ll be sitting pretty on  BuyingABunchOfDomainNamesAndHoldingThem
WithTheHopeThatSomeChumpWillPayYouBigBucksBlo.ws. (Thanks so much Western Samoa!)

 

In a continuing trend of more and more entities embracing the potential of social search, Kratia.com has launched what they call a ‘democratic’ search engine. From the greek word ‘to rule’ Kratia attempts to put the power to control the SERPs in the hands of its users. 

You start off with the MSN search results which you can vote on.    You choices are ‘I Like It’ and ‘I Don’t Like It’, and you have the option to also leave a comment with every result.  It is really new, so in total only 7172 votes have been cast so far. 

While I like social search, I am disappointed with this implementation.  The main problem is the weak, manichean voting system.  While it’s useful on some level to understand if people like something or not, it would be far more useful to understand why they liked it, and what they were using it for.  For example, choices like ‘Great for Scholars’, ‘Great for Shoppers’, ‘Entertaining’, and ‘Spam Site / Adwords Nightmare’ would more usefully divide sites up for me.  If people put in comments, one could go to the comments to try to determine what the like or dislike was about, but the comments are not searchable at the moment and the extra time to read through them would lead to too slow a search process for my taste.  They could make the comments more useful by giving partial snippets under each SERP Result.  Another thing Kratia.com really needs is advanced search.  You should be able to limit your query to results that have been voted on, or only results that have comments. 

To be fair though, Kratia is brand new, and perhaps they will be able to make a better go of it in the social search arena given more time.  Right now though, it feels less like Greek self-determination than passive, couch-based Nielsen ratings style voting.  I can feel myself giving them valuable information but don’t feel like I’m getting the same in return.

One of the problems with trying to fight content redundancy over multiple sites is that if you don’t properly identify the original source of the content, you might penalize the author by wrongly labeling them as a redunandant instance. This has been pointed out by various sources as a possible problem with the current Google rules around duplicate content. 

Here is one idea for how this problem might be solved. 

1. A site owner creates an account at a search engine where they establish their identity and receive a unique site id.
2. The site owner then logs in and provides an article, which the search engine in turn gives an article id.  The author embeds this id in the cotent via some kind of meta tag when the article is published to the web.
3. When content is analyzed by the engine during its normal indexing process, if it sees an article id, the search engine cross-references the id with the site owner id and profile to see if it is registered to that site owner. If the article is registered, the site owner is considered the source whenever there is a question of duplicate content where some site must be hidden.
4. To deal with instances where content is fraudulently registered, you would need some kind of challenge process where someone could submit evidence that they were the rightful author. In turn, site owners found to be repeat offenders in terms of false content registrations could have their accounts revoked and could be removed from the SERPs entirely in extreme cases.

There would seem to be benefits all around from this kind of system.  First of all, site owners who work hard to create original content would have more time to spend being creative and could spend less time on copyscape and similar services hunting down the people stealing their articles.  It might promote greater use of RSS, as you could build into feed systems something where it is not syndicated until each entry is tagged with a valid article id, giving possible syndicators less to worry about in terms of duplicate content penalties.  The engines themselves would benefit as well as they would have a good way to have content pumped into them directly without having to be as dependant on going out to find content.  Also, the engines would have yet another possible tool to use to round up and discipline spammers.

There are of course a ton of logistical issues to overcome, and figuring out what to do with all the existing content out there would be an enormous hassle, with huge potential for fraudulent claims.  However, ultimately we need something better than the copyright system to identify the source of content.   

By the way, if someone does create such a system, I claim my right to article ID #1.  First registered post!