Forum Moderators: open
The sites concerned have non-UK specific domain names and so I guess Google determines their location by the IP address. If I use online services to locate the IP (I've used as many different ones as I can) then it always shows up as being in the UK (even down to the district where it is located in some cases).
What happened? Is anyone else affected? And does anyone know anything I can do (short of switching all the sites on my server to .co.uk domains) to fix this?
I know of some servers located in the same building that DO show up in .co.uk results, and also some that DON'T. This is really confusing and my traffic has taken a painful hit as a result. Any comments/advice would be greatly appreciated.
This issue has been discussed a number of times:
Why not appearing in Google's UK pages? [webmasterworld.com]
UK hosted site not appearing on Google.co.uk [webmasterworld.com]
Listed in [Google] .co.uk, but not in .com? [webmasterworld.com]
Non-US hosting: Does it hurt your rankings in Google.com? [webmasterworld.com]
Hopefully there'll be something in there to help you. The flaws in Google's implementation of regional identification are indeed frustrating. If switching to UK domains would help restore your traffic, maybe you'll have to consider biting this bullet.
If enough people are affected does this not warrant some investigation from Google themselves, after all, it's in everyone's interests for Google.co.uk to return UK sites regardless of the choice of domain name.
I'd like to think that a company so focused on delivering relevant results to searchers would always be looking to improve things. But if traffic is hurting, can you afford to wait for someone else to do something?
Just thinking about your first post though - doesn't google.co.uk default to searching 'the Web', rather than 'pages from the UK'? AFAIK, it would only be the 'pages from the UK' results that are affected by regional filters. You must have had a lot of traffic that chose to 'search UK sites only'?
I was aware of earlier postings on the subject, but it seems that no-one has satisfactorily resolved this issue.
It is a very difficult issue to resolve - I know because it is something I have been working on for some time. The 'dark web' problem is worse for smaller countries that don't have a good internet infrastructure but at first estimates, it is running at 15% for the UK. For countries like Ireland, the error margin on com/net/org domains can be as high as 50%. Google's methodology seems to centre on filtering websites with .uk on the end and websites hosted on UK IP ranges to generate its 'pages from $country' index. It is far from perfect and ignores websites hosted outside the UK. Furthermore it can induce problems by including websites that are relevant to other countries into the UK results simply because they are hosted on UK IP ranges.
If enough people are affected does this not warrant some investigation from Google themselves, after all, it's in everyone's interests for Google.co.uk to return UK sites regardless of the choice of domain name.
Determining what is and what is not a related website for one country is hard enough but on a global scale, you are dealing with hundreds of countries. I don't think that Google's current methodology is capable of doing this for each country. (I even e-mailed Google about a methodology that would work a lot better than its quick and nasty method this but got nothing other than the canned e-mail reply). The issue is closer than to cryptography than to search engine work.
Regards...jmcc
Two years ago, I picked up lots of nice traffic from keywordhere.com domains. Now it has fallen, with UK searchers looking for keywordhere.co.uk domains growing substantially. I really do think that if you are targeting the UK, you should use a country domain - not just because it will help on the SEs - but because your target market is looking for it - rather than 'just another .com'.
>>But if traffic is hurting, can you afford to wait for someone else to do something?<<
As far as I can see I have 3 choices - I can either hope that Google will look at their method of IP location (and as I say, looking up the location using online services ALWAYS returns the correct country (including [ripe.net...] )) which in all probability will never happen, I can settle for lost traffic from uk specific searchers, or I can rebuild my sites from scratch using .co.uk domain names. None of these is a particularly attractive prospect.
I also noticed something really strange - some of my sites DO appear in the .co.uk results, but always near the end of results pages, and never the home page - does this mean Google thinks some of the pages ARE in the UK? Or my PR in the UK is different than in the US?
This is very confusing.
I can understand that it is not an issue of paramount importance to Google - but it does affect search relevancy. And if there are so many good ways to locate countries by IP, why are Google using a bad one?
I can understand that it is not an issue of paramount importance to Google - but it does affect search relevancy. And if there are so many good ways to locate countries by IP, why are Google using a bad one?
Locating a website on the basis of IP would only work perfectly in an ideal situation where each website is hosted on the IP ranges of its relative country. However in the real world, the IP method that Google uses will provide a good starting point but it cannot account for the 'dark web' aspect. The availability of cheap hosting (eg in the US or the UK) outside of a particular country can force many websites out of that country and on to the IP ranges of the US and the UK. By Google's method, these websites would automatically become US or UK sites. That is the point where Google's methodology collapses. I really think that Google is just trying to solve it in conventional search engine terms (if indeed it actually is doing anything about it).
It seems to me that Google uses a very simplistic website - IP - country model. <cynical mode>It probably reckons that the 'pages from $country' aspect is not used enough to devote the resources to creating a better model.</cynical mode>
Regards...jmcc
I've seen quite a few that leads me to believe they don't. I *think* that they outsource it and the contractor attempts to trace the country.
igloo, have you tried [visualroute.visualware.com...]
Yep, resolves to London, UK, as does [ip2location.com...]
[geobytes.com...]
[antionline.com...]
etc.
I know a lot of these services as a result of anti-fraud research.
Determining what is and what is not a related website for one country is hard enough but on a global scale, you are dealing with hundreds of countries.
This would let the webmasters of sites which Google incorrectly identifies, to easily get round the problem.
How does that idea sound?
Perhaps a way for Google to let webmasters help them out in this respect is for them to recognise a meta tag that could specify which country the site was
based in.This would let the webmasters of sites which Google incorrectly identifies, to easily get round the problem.
How does that idea sound?
Idealistic. :)
There still will be a margin of error due to the sites that do not include the meta tag.
Regards...jmcc
I agree that this is idealistic, but as there is little mileage to be gained from saying you are in a country you're not, this could be a good workaround for people in my situation.