It has been more than a month after Google told the world what changes it has made to its algorithm for February. There have been 40 changes related to search quality for the month. Though the March updates are out already, I’m still talking about the February ones because, among those 40 for the Feb month, there were three improvements on the quality of local organic search results – after which I’m still seeing a few results in my country which keep me puzzled.
Here are the ones I’m talking about (of course, you might know already).
Improvements to ranking for local search results. [launch codename “Venice”] This improvement improves the triggering of Local Universal results by relying more on the ranking of our main search results as a signal.
Improved local results. We launched a new system to find results from a user’s city more reliably. Now we’re better able to detect when both queries and documents are local to the user.
Refreshed per-URL country information. [Launch codename “longdew”, project codename “country-id data refresh”] We updated the country associations for URLs to use more recent data.
To my best guess, these altogether mean that Google is better able to identify the user location, local queries, local documents and local websites more reliably by now – while ensuring that the results are strictly relevant to the query relying on the traditional ranking signal. We will get back to that traditional ranking signal stuff in a while.
Yeah, and it worked. A lot of webmasters did notice more localized results in Google search. That applies for generic queries as well, since localized results for those queries are returned based on the user location. I believe that the whole local results improvement stuff largely relates to returning more localized results for generic queries.
Localized? Not Completely.
So let’s get down to my case here. My location is Chennai – a South Indian metropolitan. So, my local domain for Google search is [co.in] and google (almost) always detects my location accurately.
I, too, have started seeing more localized results for many of the queries I type in Google. However, a recent search I conducted for “Plumber” left me puzzled.
Yeah, there are plenty of relevant results in the first page – a bunch of local listings and about 3 pages from India. I don’t mind the two Wikipedia pages appearing at the top, even. Because I know that the search for “Plumber” in India is far too less.
But I didn’t understand why the plumber of South Floria would show up in the first page for Chennai area. In fact, the service location of that website is appearing in every possible part of the result snippet. May be I should choose to view only the pages from India, which I did.
South Florida link now climbed up the results, outranking three localized Indian results. And there were two more pages from abroad – One serves Atlanda North Metro and another, Phoenix.
Clearly, I have been seeing pages from abroad for this query for over a month. Only the website was different which is serving Wilmington, DE.
Server Location Can Affect SEO (Hey, Matt says that)
Alright, the content of all these pages whatsoever don’t have anything to do with my location, but what if these websites are hosted in an Indian location, possibly? If Google attempts to find the locality of the document using the website’s IP… Yeah, it is possible. We’ll come to that in a while.
As it turns out, none of these pages are hosted in India. Their servers are in different locations across the United States – Scottsdale, Illinois and there was even one located in Quebec, Canada.
And the most puzzling thing is that these pages are not ranking for any custom location where either the above services are provided or their servers are located. If they are not ranking for their own localities, why are they ranking here? And what happened to traditional ranking signals now?
So, to put all these simply:
- None of these pages have content related to my locality
- None of these pages serve my locality (Google crawls and uses telephone numbers, address, etc.)
- None of the websites are hosted in India
- Still some/all of these pages appear in Google results for Chennai location and vary over time
As far as I know, there are a few factors through which Google could identify the nature of the domain.
Google has the ability to figure out the location using the physical address, phone numbers and zip codes (In the ‘Specific Factors’ look at the 2nd, 7th and 11th elements).
Matt Cutts himself confirmed that Geographic location of a web server will affect SEO. Here is an excerpt:
“Yes it does. Because, we look at the IP address of your web server. So if your web server is based in Germany we are more likely to think that it is useful for German users. That’s not the only country we are going to return you for and we also look at the tld’s, we also look at the .te .fr all those sorts of things. You can also specify in Google’s webmaster’s console and say yes my site my .com or .whatever is about this specific country. You can even do that for specific parts of your site like de.something.com or fr.example.com. So I would absolutely recommend that you use those tools. If you find a great deal in a particular country and if you really want to stay in that country with your web server, I think that’s fine. But if you are worried about it or you want to experiment you can certainly try switching the geographical location of your web server which is essentially changing your IP address and that might end up helping for various countries. So it’s the sort of thing that I would encourage you to experiment.”
Regional to the Data Centers?
After all these, I don’t buy the bizarre localized results (which are not ranking in their own localities) since Google has almost all the capabilities to identify local documents, queries and the user’s location. So, for I’m still an amateur, I asked about this to one of the best men I admire in one of his blog posts and sent a tweet. His reply included a snippet from a blog post of him about how Google’s data centers split regional and global data. Here’s the reply, in case you didn’t follow the comment link.
That particular query, and those particular pages don’t seem like ones that seem like they would be considered “regional” in India, but under that patent, if it seems like a lot traffic is going to those pages from India for one reason or another, they might be considered regional, especially if they aren’t getting a lot of query traffic from other places. If there aren’t a lot of people in India searching for plumbers, and these two pages are being searched for and selected very frequently by people at the company or companies working upon them, it’s possible that the algorithm that could be considering them as regional might just be.
They don’t look like they would be considered “global” sites that would be replicated across many data centers. They may not be very popular in the US either.
Here’s a snippet from my post:
Content that isn’t world-wide could be included within a particular index as regional content, and may be located within a regional index at a data center based upon being similar to characteristics of the queries received at that particular datacenter. For example, if 75% of web queries from Lithuania are in the Lithuanian language, then many of the pages within the data center for those searches may be in Lithuanian. Pages that are popular in Lithuanian that aren’t in the Lithuania language may also be included in the regional index for that data center if those pages aren’t popular enough elsewhere to be included in the global index.
Yeah, that makes sense. But I wonder how many people would be searching for plumbers and clicking on websites from the United States and Canada. It’s not utterly impossible, but I don’t think many people would be visiting those websites without a purpose. Oh yeah, there is one possibility, but that could generate only a small amount of usage data; relatively negligible, I would say.
Still I doubt if such a small amount of queries (given the demand trend) in Indian data center would significantly affect the results for a query as generic as “Plumbers”?
I would much rather prefer throwing a bunch of crappy content from my region than giving something geographically irrelevant to my search. Well, if Google identifies the nature of the content based on traffic sources, then I feel it’s nowhere related to their update.
What Bill said seems to be a perfectly valid point, but still “The codename: Venice” is making me to stick towards complete relevancy and nothing else.
Typical for India : Typical for USA
But let’s not get deceived. I did notice this kind of bizarre results for a few other queries as well – Filling Station, for example. I observed that queries that return this kind of results are typical ones for United States and Canada (say, “Petrol Bunk” is typical in India than “Filling Station”). I haven’t seen any strange results for queries that are typically used in India, at least not yet.
To Sum Up
so what do all these mean? Aren’t we completely localized? Or is Google just on the way to implementing its updates to country specific domains? What kind of traditional ranking signals is Google using for returning localized results?
While we are still wondering, Google has improved its algorithm with local intent in March as well.
Better handling of queries with both navigational and local intent. [launch codename “ShieldsUp”] Some queries have both local intent and are very navigational (directed towards a particular website). This change improves the balance of results we show, and helps ensure you’ll find highly relevant navigational results or local results towards the top of the page as appropriate for your query.
Better local results and sources in Google News. [launch codename “barefoot”, project codename “news search”] We’re deprecating a signal we had to help people find content from their local country, and we’re building similar logic into other signals we use. The result is more locally relevant Google News results and higher quality sources.