For want of a better name…
I have been trying to come up with a suitable name for this since Tuesday night when I first spotted it in some queries I was using to find some old posts I had written.
The first name I came up with was “HyperPersonal” because I was seeing it in more personal search results and I mentioned it in Dojo chat to get some feedback.
Then the chat switched to the evils of hosting WordPress on windows servers and the death of Search Monkey and other Yahoo APIs.
24 hours later I came up with another name “Hyper Exact Match” though that is terrible English but by this time I had decided that the matching was very specific to domain queries as some purely name or brand associations just didn’t work the way I expected them to.
I came up with this explanation for this:-
If a query can be determined to be specific to a particular domain, and the domain has multiple results for the query, Google instead of showing a link “More from this domain” will show up to 6 results.
Actually it can be more than 6, but that is as many as I had seen at the time for the limited queries I was testing on.
Dave later suggeted the following
@Andy, brand task search? Not really a ‘site search’
[2010-08-18 23:28:20] David Harry: Domain query search?
[2010-08-18 23:28:43] David Harry: the lack of a ‘site’ command kinda takes it out of a ‘site search’ imo
But I am going to stick to calling this Pseudo Site Search.
As far as I am concerned this is a site search for mortals who haven’t fully mastered Google query syntax but are showing intent to retreive information from a particular resource.
Meanwhile Malcolm got the jump on me and posted something before me.
Bill Slawski chimed in in Malcolm’s comments with this
This looks like the process described in Google’s patent “Query rewriting with entity detection” (US Patent 7,536,382). which was granted in May of last year.
For example, the process might identify Apple as a specific entity that is associated with a specific web site, and rewrite the original query to provide results from the Apple site. From the patent:
Some entity names are unambiguous and uniquely identify particular entities. A large number of names, however, are somewhat ambiguous or generic, making it more difficult to identify the entities to which they are intended to correspond when included in users’ search queries.
Systems and methods consistent with the principles of the invention provide mechanisms for determining the entities to which entity names correspond and selectively rewriting users’ search queries based on the entity names. Accordingly, a user’s search query may be restricted to a search of document(s) associated with the entity that the user intended in the search.
Bill has now followed up with a blog post of his own based around a number of Google (and Yahoo) patents on Entities.
So I spent 36 hours working out what to call it and whilst I have been pipped at the post (literally) I thought I would still write something and include some of the examples I dug up.
More From this Domain
This is how things used to be
Some domains you would get more than one additional link – I have seen 4… maybe 5 links in the past for huge sites, plus a suggestion for more content from the same domain.
This is what made me call this hyperpersonal – the search I performed as I was just deciding what to respond to a review request (that I turned down).
Sidenote: I am close to deleting my profile on all the paid review sites – the requests are most often a waste of time, and those that are interesting are on sites that now by default state you have to give followed links – if they think the links are more valuable than the feedback they know where to shove it.
This also shows how useful this is… [site:andybeard.eu pagerank] is a much more complicated query that you could get wrong
There are lots of popular sites that are .net (Slideshare, Problogger) or .org (SEOmoz)
How many times have you used a site query for the wrong domain due to this?
Michael as one of the biggest brands in blogging would be an obvious choice, especially as he tends to write only for his own blog Techcrunch.
It didn’t work for Michael…
It does work with Techcrunch… hmm
So this is what got me thinking that it isn’t so much “personal”… the “entity” needs a strong association with the domain name possibly, and not just with the website and contents.
If you have read Bill’s post then just try [bill slawski patent] and [seo by the sea patent]
Whilst Bill has written in other places frequently, I only associate him with one domain for patents.
News Query Space
[Chicago Tribune Obama] didn’t bring up the new results, so I stuck a .com on it to bring up something interesting
Nine results for Amazon in the UK for this query, but look at the subdomain switching between www and astore.
I thought I would grab a pretty one from the US
Entities, People & Exact Match Domains
There is nothing clear cut with this…
For instance you would expect Aaron Wall of SEObook to be a strong enough brand associated to a single domain for this to work.
[aaron wall pagerank]
[aaron wall google] (Google knol does exist still!)
If you slap a .com on, it works
Interestingly searches for a domain like that in the past would bring up references to the domain, not lots of results from the domain itself.
Google’s job is to provide relevant search results for a query (and make money doing it) – this does not mean that you will necessarily gain search traffic by having a strong brand associated to an exact match search query.
I have lots of weird and wacky things regarding search that I have discussed in various places (mainly private communities) that I still haven’t discussed here on the blog that I refer to as “Pinocchio SEO” that freaks highly competent SEOs out a bit, and I have no idea how this will relate to that.
I prefer “Pseudo Site Search” over “Query rewriting with entity detection” as I think that is the kind of query that is being returned… maybe a combination “Entity Search” would be better, but that doesn’t quite make the domain relevance/reliance strong enough.
Matt has an update on this as well with a quote from Google
Postscript: A Google spokesperson has confirmed that the search results discussed above are part of a ranking/user interface change related to domain-based intent:
“We periodically reassess our ranking and UI choices, and today we made a change to allow a larger number of pages from the same site to appear for a given query. This happens for searches that indicate a strong user interest in a particular domain.”
Hmm “related to domain based intent” just means that is the catchment for this… but I am sticking with my pseudo site search.