Have you ever wondered what makes certain search results appear above others
in my searches? (rankings) You may have put it down to “that’s the rocket
scientists algorithm”……I tend to think of it as a magic sauce/wine/sangria or
Grandma’s special pancakes – a formula refined overtime.
Here’s a little insight:
Microsoft’s #1 Priority for this release of Search (it’s been rebuilt from the ground
up from previous versions) is Relevance and Priority of results.
Getting more accurate results appearing at the top of result lists.
Some of the ingrediences going into the sensational pancake mix are (a partial list):
-
Click Distance – how far away you are from ‘authoritative sites’.
e.g. take your intranet, it’s got documents on it, links to internal and external
locations. As you drill down these links, essentially your getting further away. -
Anchor Text – html links act as annotations in what their pointing
to.
e.g. ‘Link to Footy Results’ = http://moss/sites/footy -
URL Depth – Generally URLs higher in the heirarchy tend to be more
relevant than those below. e.g. simply looking at the number of slashes in a URL this
may indicate that you’re going deeper and deeper into more specific areas, and loosing
relevancy (possibly).
We can also go and make certain sites/URLs as being ‘authoritative sites‘
that allow us to influence the search ranking and relevance where needed. - URL Metadata – direct matching on text within URLs
-
Metadata Extraction – automatically extract metadata like Titles,
Author, LastModifiedDate etc. from documents etc. -
Automatic Language Detection – results in your language bubble to
the top. -
File Type Biasing – generally speaking, PPTs will be more relevant
than XLSs for e.g. -
Text Analysis – Text ranking based on matching terms, frequencies
and word variations.+ more………
So just in case you’re ever wondering is there alot that goes into MS Search and getting
your results to the top of the list……absolutely!!! I dont quite have the formula
to Coca Cola yet…but working on it…….