Google Explains Ranking (Basically)
Today, Google Fellow Amit Singhal posts on the Official Google blog explaining Google’s ranking system. But before you all start salivating and clicking, note that this is a very “high level” overview—no secrets given away. Still, it’s a good reminder of the basics of Google’s ranking system.
Singhal writes that Google’s ranking system is based on three guiding principles:
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- Best locally relevant results served globally.
- Keep it simple. [Thanks for not calling us stupid.]
- No manual intervention.
Now for the breakdown: [musical break, awesome choreography]
Ahem. Now for the interpretation:
Best locally relevant results served globally.
Despite the fact that Singhal says this one is “obvious,” to me this is the most oblique of all of the statements here. But yes, apparently he does mean that the best “local” result should be the best overall result:
We often call this the “no query left behind” principle. Whenever we return less than ideal results for any query in any language in any country – and we do (search is by no means a solved problem) – we use that as an inspiration for future improvements.
‘Kay, maybe I’m dense, maybe I’m just too bounded by the denotations of the words that he’s using, but . . . what? What he’s saying here sounds like “When queries don’t return good results, we want to improve.” That’s awesome. But what does that have to do with “local” and global results? Or does he mean that the best “local” result (results served in other languages/countries, not we would actually consider “local searches,” I guess) should be as good as the best overall result? That raises a slew of other questions.
Keep it simple.
The first principle “is obvious” and this one “seems obvious.” (I think I’m in for a headache, especially when he asks “Isn’t it the desire of all system architects to keep their systems simple?” Oh, if only.)
No, really, this is simple:
We work very hard to keep our system simple without compromising on the quality of results. . . . We make about ten ranking changes every week and simplicity is a big consideration in launching every change. Our engineers understand exactly why a page was ranked the way it was for a given query.
But no, they won’t tell you if you corner them at a search conference.
No manual intervention.
Singhal gives two reasons for this: first, that any one individual is too subjective to render good, objective results and second:
often a broken query is just a symptom of a potential improvement to be made to our ranking algorithm. Improving the underlying algorithm not only improves that one query, it improves an entire class of queries, and often for all languages.
This does come with a caveat:
I should add, however, that there are clear written policies for websites recommended by Google, and we do take action on sites that are in violation of our policies or for a small number of other reasons (e.g. legal requirements, child porn, viruses/malware, etc).
Singhal promises more fun in a future post, promising to “discuss in detail the technologies behind our ranking and show examples of several state-of-the-art ranking techniques in action.” Excellent.