Anchor Text Question Options

[Nystul]

Due to the long name of my website (keyword), a couple of my backlinks are anchor text as thisisme, instead of this is me.

Just wondering, does google see the backlink anchor text as 3 words too?

Thanks.

[Randy]

No.

They'll see it as thisisme

[robchuah]

Exaclty. There is no way for google bot to segregate out the words thisisme.

[Nystul]

Thank you. Much appreciate.

[piskie]

I think Google does word splitting and word stemming.
Note: I said "I Think"

[Randy]

Two different things piskie. Stemming, yes. Splitting no.

Stemming is when they recognize a word and will then return pages that use variations or similar/synonymous words for the word they recognize. Often this means the plural or non-plural, depending upon which was in the original search request. But it can also mean other variations. If memory serves Google's help files used an example of stemming with returning pages that contain the word "dietary" when the original search was "diet", though their example was a longer phrase. The same would apply for things like diets, dieted, dieter, dieting, etc.

Stemming of course means they first have to be able to recognize a word. If they don't recognize a string of text as a word it will not have any synonymous words with which to initiate stemming.

Word splitting is a whole other matter. I've not seen any evidence suggesting any of the engines attempt to split words. Especially if the root word isn't really a word.

[piskie]

When I said Stemming and Splitting, I didn't mean to imply any connection in function or meaning. Clearly they are different.

However, I believe Google introduced Both at about the same time using Technolohy they Acquired when they bought out "Applied Semantics". This was primarily (I think) for Adwords, Adsense and associated services, but I believe Applied Semantics were also in possession of Context Sensitive Wordsplitting technology as well as very Advanced Word Stemming capability Algos.

I suspect that Google is Currently using Word Splitting as part of it's Organic Algorithm, but obviously if they are, they woulsn't announce it to the SEO World. You may Dissagree, but just as much as I can't prove it, can anyone Disprove it ??

[chrishirst]

The only kind of "splitting" Google does is for the "Did you mean: " feature for common errors.

For example;

http://www.google.com/search?hl=en&rlz...amp;btnG=Search
is different to
http://www.google.com/search?hl=en&rlz...amp;btnG=Search

http://www.google.com/search?hl=en&rlz...amp;btnG=Search
vs
http://www.google.com/search?hl=en&rlz...amp;btnG=Search

http://www.google.com/search?hl=en&rlz...amp;btnG=Search
vs
http://www.google.com/search?hl=en&rlz...amp;btnG=Search

[piskie]

Chris, although your examples show word splitting in this context, they in no way prove that it is:

The only kind of "splitting" Google does

These examples only go to show that Google DOES use "Word Splitting" and not the boundaries and context to which Word Splitting is applied to.

[chrishirst]

These examples only go to show that Google DOES use "Word Splitting" and not the boundaries and context to which Word Splitting is applied to.

Huh ????????????????
So you think that 3 results for a combined word vs almost 1/4 million for the 2 seperate words shows they are splitting words in queries ???


It's very very simple. Why would ANY search engine do word splitting as returned results???

It really would be wasted processing power.

what if a user types in redfordcar? do they want results for "red ford car", "redford car" (as in a search for a car in Redford) or are they looking for redfordcar??

The software engineers cannot afford to second guess what the user wanted, that way would lie disaster for the SE in question. Users would very soon be very pissed off if they had to jump through hoops to force their intended query to be recognised.

You wouldn't be able to search for
greendale (green dale)
Liverpool (Liver pool)
or
Cornwall (corn wall)

Try those as two word searches BTW


Yes they are very capable of splitting seperate words and returning results based on that. BUT they are very aware of their user base and so they offer the most likely alternative.

[piskie]

Chris, I think you've shot off at a tangent which is most unlike you.

Keep it simple one step at a time.
In the examples you gave Google had used word splitting. FACT.
Statement "Google does use Word Splitting". Fact. QED.

The extent to which Google uses Word Splitting or in what context does not alter the fact that you yourself established "Google uses Word Splitting". Trying other search examples which may not split words does not make your examples go away. Just like it is not possible to be a little bit pregnant.

So as Google does use word Splitting even just "Some" of the time, who knows where it starts and finnishes.

[Nystul]

Nice discussion pple. Learnt alot from this thread. thanks.

[BBCoach]

In the examples you gave Google had used word splitting. FACT.
Statement "Google does use Word Splitting". Fact. QED.

The extent to which Google uses Word Splitting or in what context does not alter the fact that you yourself established "Google uses Word Splitting". Trying other search examples which may not split words does not make your examples go away. Just like it is not possible to be a little bit pregnant.

So as Google does use word Splitting even just "Some" of the time, who knows where it starts and finnishes.

I wish that were true, but SEs do NOT try to interpret concatenated words. How do I know this for a fact? For the last 10 years every time one of our copywriters missed a space in the title description of one our products those products could only be found in SEs with the concatenated words (usually in the top 10) and not the two words (unless additional copy with the space was provided in the description) and those products never ranked well in SEs until they were fixed. Dang copy writers!

BTW, it's not something a programmer would do because all logic has to start with a basic truth (business rule). In this context, that truth being that the word was spelled right or wrong. If right, it finds matches and if wrong, it finds similar phrases and checks the dictionary for correct spelling and offer up "Did you mean...", or return we can't find anything on the planet that matches. BTW, SEs have armies of linguist on staff reviewing search phrases misspelled and/or mis-typed (concatenated) and are creating lookups based upon the most popular phrases, but that is not true interpretation or "splitting" of the phrase. That's treating it like it was spelled right in the first place. Chris is right. Computer interpretation would take too many CPU cycles attempting to understand every query string so humans step in and create a lookup based upon that phrase. You're kinda right too due to the human intervention in creating the lookups that do return relevant results for popular phrases, but "splitting" is not going on.

Do a simple test with the words baseball, base ball, basesball, buseball and buse ball. Notice anything? Google is pretty dang accurate with those words ain't it. Not a computer -- a human made those associations.

[MaKa]

Due to the long name of my website (keyword), a couple of my backlinks are anchor text as thisisme, instead of this is me.

Just wondering, does google see the backlink anchor text as 3 words too?

Thanks.

Abstract of all above answers: No Google doesn't do word splitting for backlink anchor text, Google does sometimes do word splitting for queries by providing a "Did you mean" alternative.

Two completely different contexts, it makes a different to trying to word split every word and part of each word that they index potentially increasing their index size a lot, or only queries typed in by people - taking only a fraction of a second for each query.

[piskie]

When I proposed that Word Splitting took place, I did actually qualify it with 2 words, Context and Boundaries.

If some splitting takes and is visbly demonstrable, it is possible that some splitting takes place that is invisible.

Of course, a space missed out between 2 words even if "Splitting" took place, would not rank as high as the correctly typed 2 word phrase nor would we expect it to. However it could (just could) be a minor contributing factor could it not. Also in a very uncompetitive area it could possibly (again just could) increase in influence.