Google wants to be GUARDIAN, who knew?

Try 48273426.com on your cell phone/pda or on a wap emulator such as wapsilon.com

48273426 is Guardian in numbers (look at your phone keypad).


Technorati tags:
technology, Web/Tech, mobile, google, wireless

Google Doodle - the future of mobile search

Google Doodle. No, it's not for kids. It's for everyone to make the mobile search experience the most enjoyable and efficient. Here's the proposal:

The first step is to transform the cell phone keypad into a touch pad. Take the keys/buttons off, and make it a flat surface, like a touch pad for the laptop. Keep the numbers and letters visible, printed behind a clear surface.

The next step is to keep the Google Number Search system but instead of the search input being numbers, it will be a doodle (touch pad input) by the end-user. That's right. The end-user just needs to move their finger along the route that they would follow when typing in the numbers for a GNS term, tracing the path as a doodle. So instead of tapping 32632646 (FANDANGO), the user just needs to connect-the-numbers, using the visible keypad as a reference. This picture is how the doodle will be made for FANDANGO:








Google will get input as a set of touch pad coordinates that, when graphed, would look like this:








Note that somehow - and I have absolutely no expertise in this - Google would determine where the doodle began (on the 3) and ended (on the 6).

If you think about it, or try GNS enough, you'll find that there are a gazillion distinct patterns that one follows along the keypad for just about all the retail brands, company names, etc.... Try Microsoft (642767638), Google (466453), Amazon (262966), etc... Each route that you are following when connecting those numbers is unique. All Google needs to work out is how to interpret double- or triple- taps when the finger just stays in one place for (what ordinarily would be) two or more taps on the same key, as in the case of the two sixes in 466453. That shouldn't be a big problem for Google to figure out.

The tremedous value of this mobile search proposal is that it is FAST. Darn fast. I can doodle a FANDANGO in 1.3 seconds. What other mobile search technology can beat that? Heck, I can't even say Fandango faster than doodling it! There's no limit to how fast you can doodle entries as long as it is legible (to Google). If you don't believe this is within the faculty of human predisposition, watch an expert signer (sign language) or an expert in drawing Chinese characters or a floor trader's hand signing on the NYSE.

Any takers?
mobile search, google, doodle, fandango, internet, mobile technology

Sponsored results: I shall have nothing to think of then said the basswood with a smile but my Dictionary. You asked Gonzalo said charlottesville tomatoeses by wholesale. It was curious and interesting nevertheless Mr. Faulkner said not one word though the old lady looked to him as if for his commentary on and putting them through all kinds of contortions in his small pantry.

Why most emerging mobile search technologies will fail

Microsoft recently announced a new search technology called Photo2Search, which would allow a user to search web databases using cameraphone photos - the search would return relevant results as would a text query.  Another mobile search idea is to install barcode scanning technology on cellular phones so users can 'scan' products while shopping to get instant access to product info. 

 

It is clear that mobile search technology is moving, however it is not moving forward.   Mobile users will soon agree on one golden rule for mobile search - there should be one portal which gives me access to all the information I need.  As of now, Google is the one portal that lives up to that rule on the home computer.  However, searching on a mobile device using Google, or any other portal, is problematic because multi-tapping on a phone keypad is too cumbersome.  There is no one mobile portal or solution that meets the golden rule.   To overcome the keypad problem, Google's competitors are trying to come up with solutions to substitute text input with barcodes, photos, voice, etc...

 

The mobile market has not only developed a bad case of amnesia of Google Number Search, but is headed down the wrong road.  More and more subscriptions are being offered to users by carriers that limit users to a small selection of content. When users are offered search capabilities, the results and efficiency of use are far inferior to what they can get on their computer at home. Shrinking a user's web experience or search capabilities will not work in the long run.  Even predictive text technology, which is becoming ubiqitous as the staple technology of mobile search solutions firms like Jumptap (not related to this blog or its editor), will not survive because it is more so much more demanding than GNS on the user. Although both methods require roughly the same number of keystrokes, T9 requires twice the users' attention and faculties than GNS because it requires simultaneous attention on keypressing AND repetitive (visual) scanning of possible results for each keypress, whereas GNS requires attention only to inputing the query.

 

Google Number Search has the optimal flexibility for mobile search.  It's strength over other mobile search technologies is that it enables the user to locate ANY results on the World Wide Web (Google's index) using a simple interface and set of rules.  GNS allows the user to incorporate very basic and fundamental tools into their mobile search such as the wildcard (asterisk), quotes (number one), spacing (zero) and manual search/override (NUM).

 

One only needs to experiment for oneself the power of number search using allnum: or num: on Google search.

 

Asterix reloaded

As a follow-up to the last post: the asterisk works both ways!   You can not only search Noam* to get the spelling of 'Noam Chomsky' but also *Chomsky to get the spelling of his first name.

This feature works in Google Number Search as well, however the placement of the asterisk is a bit peculiar.  The asterisk doesn't go just before the word-number but before the NUM operator.

So you would type (2466759 is Chomsky): *num:2466759
http://www.google.com/search?hl=en&lr=&q=*num%3A2466759Placing the asterisk BEFORE the search term has somewhat different applications than placing it after the search term.  What are these applications?  One use is to find out the name of company post-merger, as in BNP Paribas or ABM AMRO.  Another use is to find people who have a certain surname, like Rockefeller.   Or perhaps you are a chemist and want to know all of the compounds ending in Hydroxide.    Same case for meteorology - types of clouds

My question is how would this work on a cell phone.   I can understand that two of the portals on 466453.com would allow you to NUM the WEB and also ALLNUM the WEB (or similar terminology) without asking you to insert a 'num:' or 'allnum:'.  In this manner, I can see how the user can append an asterisk to a query on the NUM search.  The user, however, couldn't place an asterisk before the NUM function in this scenario.  I can only guess that somehow the NUM the WEB function will somehow seek for an asterisk if placed directly before a search term.  That way the user can benefit from the uses of the asterisk just like in normal Google (except the asterisk doesn't - everyone join in - work with ALLNUM).

Say asterisks 10 times fast

"I know you are not," said I, "because if you had been, you would have told me or
at least have been better away."

Jamptap has learned that the asterisk works with Google Number Search.  Advanced google users already know that the asterisk acts as a wildcard for a missing word in a normal Google search phrase, as in

 

Search: Today is the first * of my life:

 

http://www.google.com/search?hl=en&lr=&q=today+is+the+first+*+of+my+life

 

If you're not familiar with the 'power' of the asterisk, try the same search without using using it:

 

Search: Today is the first of my life

 

http://www.google.com/search?hl=en&lr=&q=today+is+the+first+of+my+life

 

In normal google search, the asterisk can also be useful when appended to the end of a single word if you want to see the most popular word (or in a few cases words) that follows the provided search term.  This also works in Google Number Search, except that you have to append the asterisk to the end of the word-number; it only works for a single search term.  For example, if I am searching in number mode and feel fairly confident that if I search Paula* then I'd get Paula Abdul, then I can search 'num:72852*'

 

http://www.google.com/search?hl=en&lr=&q=num%3A72852*

 

This can save users a lot of time on a lot of mobile searches.  

 

You can also use the appended asterisk on GNS to confirm spelling of a person's last name, like Noam Chomsky:   Num:6626*

 

http://www.google.com/search?hl=en&lr=&q=num%3A6626*

 

The asterisk also works in a way that reminds me of Google Suggest.    Let's say I want to find the most common word that follows the word 'Mobile.'  I'd search: num:662453* and get 'mobile phones, mobile Treo, mobile espn, ...' in that order. 

 

The asterisk doesn't work well either within or appended to a word-number phrase (using ALLNUM).  Oh, well. 

 

 

 

Google isn't that dumb

The posts on Jamptap thusfar have been based on the assumption that Google will re-launch its Google Number Search without updating the outdated lexicon that it has used since 2000. But Google isn't that dumb. One only needs to ask why do the ALLNUM and NUM operators work only with words that were around during 2000 or before. Quite simple. Google Number Search was announced as a new search technology in 2000, but quickly dropped from the spotlight of the press by Christmas. What was to become a new search mode for WAP users became a 'discontinued' product in a few short months after Google disabled the updating function for the lexicon of common words for GNS. Did Google do this to throw everyone off? Or did Google simply want to showcase technology rather than actually draw in users? Or was there not enough user demand? Whatever the case, Google allowed access to GNS for use (or experimentation) for a number of years at 466453.com although the lexicon it used was never updated. It appears Google wanted to still draw attention to GNS as a technological invention but wasn't ready to make it 'live.'

As a novice, I had to ask: how is that that GNS uses a different lexicon than normal Google search? The reason for this is that when GNS was launched it was fitted in 2000 with its own lexicon that mirrored the normal Google lexicon at the time. However at some point in late 2000, when demand dried up or whatever the reason, GNS's lexicon was no longer updated. Proof of this is that the present version of GNS incorporates a lexicon of search terms that lacks terms that were introduced after the year 2000, like iTunes or Cingular (the wordnumbers for these brands don't resolve with GNS because these weren't part of our lexicon in 2000). Having said that, GNS search results are the same as regular Google search results. GNS uses the same index servers as normal Google search and thus retrieves search results from freshly stored documents. The unique quality of GNS is that it limits the user to search terms current as of late 2000.

Why hasn't Google taken one of its best technological discoveries to the masses? It appears Google is waiting for the progress of mobile technology and adaption to that technology. My feeling is that NUM and ALLNUM, being new operators that Jamptap stumbled upon, are indicators that GNS will be re-launched and heavily marketed at some time in the near future. Google is readying itself for a launch.

Most of the lessons learned from GNS with an outdated lexicon are still relevant to a GNS with an updated lexicon. For example, if GNS's lexicon is updated less frequently than the normal Google lexicon, then how can a new startup, like Obopay, gain attention in the marketplace; this is assuming most people will access the internet on their cell phones and use number search. More importantly, over time there will be an increasing frequency of competition for wordnumbers such as 222 (abc, bbc) or 3776 (esso, espn). Also, the problem of the dash will remain; T-Mobile will lose on every search to the University of Mobile. Finally, there will be the problem of autostemming and its effect on search engine ranking.

So when GNS is launched (with its updated lexicon), there will still be hurdles to jump through. Google is obviously not unaware of these problems and likely has optimized GNS for re-launch. We'll just have to wait and see.

Two Threats to Web Publishers

Webmasters and web publishers should be planning now for two of the greatest threats of GNS to searching engine ranking.

 

The first threat has to do with auto-stemming.  Several years ago, Google automatically began returning search results for the singular and plural forms of search terms.   Auto-stemming also applied to grammatical variants of words as well. The problem begins here: the difference between normal Google search and Google Number Search is that the latter does not apply auto-stemming.   Whereas a normal Google search of 'search engine' will return results on 'search engine' and 'search engines,' the GNS for '732724 364463' returns the results that mostly match the numeric term - ‘search engine’ - but does not autostem and search for 'search engines.' 

 

This may be a terrible thing as it was in late 2003 when optimized webpages dropped in search engine ranking and, as a result, webmasters needed to readjust their SEO strategy.   Now, webmasters need to make sure that non-auto-stemming doesn't pose a problem to the optimization they did in late 2003.

 

The second threat to web publishers from GNS is the problem of unresolved word-numbers. (A word-number is the set of numbers that correspond to the spelling of a word, company name, etc... on a phone keypad. i.e. the word-number for ESPN is 3776.)   Newer tech companies, like Cingular and T-Mobile, will be suddenly forced to deal with the fact that Google Number Search will not bring consumers to their sites any longer.  This is because newer tech companies’ names in word-numbers cannot be resolved by GNS. 

 

One solution to this threat (and more money as a result into Google's wallet) is to advertise with Google Adwords.   For instance, if your company name’s word-number isn’t resolvable with GNS, you can still advertise with AdWords to insure visibility.  For instance, the word-number for TextPayMe (839872963) doesn’t work with GNS, but smaller word-numbers do work. That is, TextPayMe can tailor its marketing strategy to win over with common keywords, such as Pay Sms. If you try the word-numbers for PAY SMS, or allnum:729 767, you’ll find a listing in the Sponsored Links a company that proves this method works: m-enable.com (wireless payment solutions).  It may be too difficult for TextPayMe to compete with the Google search results for the ALLNUM of 729 767; examples: SMP for Statutory Maternity Pay (SMP=767) or Pop (POP=767) or Ray (RAY=729) or Raw (RAW=729). However, these are words that aren't hot keywords, and so it is unlikely any organization will bid on keywords (for AdWords) like Ray or Pop or SMP. TextPayMe, therefore, can get the visibility they need by purchasing AdWords for relevant keywords, such as PAY and SMS.    

 

The AdWords solution is a quick fix until Google learns how to include names of newer tech companies in Google Number Search, if it ever does.  There is the possibility that Google never will not budge.  The search engine giant could stand to make additional revenues by ‘refusing’ to make any alterations to GNS, citing any excuse they want as part of a strategy to increase revenues.  That’s not likely though.  But IF the status quo remains, we’re talking billions of dollars of advertising in AdWords - companies that never needed to advertise because their SEO strategies have always worked for them will need to start advertising on Google. The problem with this solution is that common web-speak keywords will be overbid on and over-saturated.

So, the AdWord solution is not the best solution for the long run.  Perhaps the solution is this: If GNS cannot resolve the word-number of a new company, and instead returns a hodgepodge of numbers as a result, then perhaps the solution lies in numbered domains.  Let’s take WORDDIAL for example.  The word number for WORDDIAL, 96733425, doesn’t resolve in GNS.

 

 http://www.google.com/search?hl=en&lr=&q=num%3A96733425

 

However, WordDial owns the domain name, 96733425.com, which is the first result for the ALLNUM of 96733425.  This way WordDial can still direct business to their company despite GNS’s imperfections.  Alternatively, a company can put their word-number in their meta tags and such and see if that works. There is a problem, however, with changing metatags and buying numbered domains. These strategies would work only if GNS cannot resolve a number - when that happens, Google treats the query as normal google search and returns results that contain the exact number you queried as seen in the example above with WordDial. The reason why this happens is that GNS cannot find a quantity of relevant results that meets a certain threshold that is set by Google. This will likely happen if your company's wordnumber is more than six letters long, like WordDial. If it six letters long or less, buying a numbered domain will do little good. The reason is that GNS would not default to numbered results and instead provide valued results.

Read the second part of this post at Google isn't that dumb

Technorati Tags : textpayme,466453,96733425,adwords,worddial,gns