SES Toronto 2010 Live Blogging: Managing a Global SEO Campaign

Anne F. Kennedy, SES Advisory Board, International Search Strategist, Beyond Ink USA

Ian McAnerin, CEO, McAnerin Networks Inc.
Crispin Sheridan, SES Advisory Board & Sr Director of Search Marketing Strategy, SAP
Michael Bonfils, International Managing Director, SEM International

Author’s Note: As an old-school SEO who has dealt with international SEO efforts for years and watched them evolve, I’m pleased to be able to cover this session, and I’m really hoping people are going to delve into content distribution, search engine geographic relevancy indicators, local search, mobile location-aware search, and alt-country search engines – not just ‘people spell things differently in England’ junk. But I’m not going to hold my breath… c’mon SES, surprise and entertain me!

Up first is Ian

Ian starts by listing countries with the top e-conomies – basically the G8 plus the BRIC countries, and a handful of surprises, Turkey and Vietnam, Poland and Netherlands.

Aso top ten languages on the net are English and Chinese, followed distantly by Spanish and Japanese, with some French German and Arabic gaining ground. Google is the dominant search engine for latin-based languages, ones that use our familiar alphabet. For non-latin languages, such as Chinese, Korean, Russian, Japanese, it’s usually a local search engine, or a highly localized Yahoo version.

Organic search share breaks down like this, globally – percentage splits

  • 86 to gogole
  • 6 to yahoo
  • 3 to bing
  • 3 to baidu
  • 1 to ask
  • 3 to others

in the US

  • 77 google
  • 10 yahoo
  • 9 bing
  • 1 ask
  • 3 others

Other parts of the world are less spread out

in Mexico

  • 98 to Google
  • 4 to Yahoo
  • 3 to Bing
  • In China91 to Baidu since Google left
  • 5 to others
  • 3 to yahoo

in Canada english

  • 82 google
  • 9 to bing
  • 6 to yahoo

In Canada, French

  • 75 google
  • 23 to bing
  • 2 to yahoo

So… don’t do GLOBAL search campaigns, do a multi-national, multi-lingual search campaign.

  1. Plan centrally – vision
  2. synchronize globally -strategy
  3. execute locally – tactics

General recommendations – localize with a local office, let local people write and translate the content, don’t let the global office do that. The inverse is that a local office should not make any branding decisions, that’s a global strategy issue.

There are translation issues – the UK vs the US english, translations of things like ‘up the creek’ if you happen to be a forieng company that has its office literally up the creek.  or ‘chevy nova’, where nova means ‘doesn’t go’ in spanish.

Good design is as much a cultural issue as it is a technical. Some cultures like pastels, latin american’s like bold colours, square corners – if you’re serious about internationalization and want to target micro-regions, find out what works best design wise.

Avoid generic ethnic images when localizing

Avoid using text elements (like paragraphs) for sizing, as the translated version will not likely turn out to be the same size.

Ecommerce details are critical – shipping information and zip codes are important – be very clear about where you ship to etc

Site planning best practices:

  • subdomain or subdirectory for language
  • domain for country (canadian french) (canadian english)

personal preference:

subdomain for language

domain for country

directory for topic

language.domain.ccTLD/topic – eg

there are four levels to this

  1. unilingual one country – easy
  2. one language, multi-national –
  3. multiliingual – one country: languages are not countries, they’re different
  4. multilingual multi-national: hardest to do but once done right, gets the best response additional advice-


  • sell in a mature markets, brand in emerging markets
  • never geolocate a search engine – they represent all users (this is contrary to Google’s advice, which says treat the like a user and redirect as you would a user, but in questions when asked about this he basically said that Google may say that, but haven’t implemented it well)
  • flags are for countries, not languages
  • ‘english site’ goes to home page, ‘english’ goes to same page, but in english
  • never translate – localize

Next up – Crispin

This guy works for SAP, it’s the big German software company, third biggest software company in the world.

The SAP websites that represent each country grew up organically in different countries over the years.

Start with business goals /outcomes – SEO is not just a technical thing that doesn’t relate to value, it relates to branding, leads sales, and becomes monetizable.

We had over eighty languages to deal with, lots of duplicate content in English and Spanish – this is a classic international SEO issue. They have tried to reduce the amount of dupplication

SAP also has a complex content supply chain.


  • webmaster tools country flag, then country domains  – they own the cc’s but havent migrated to them yet
  • canonical tagging – tell the search engines which duplicate page is the ‘master’ page to rank
  • content
  • keyword cluster research
  • keyword mapping = existing content and indentifies content gaps
  • local link building
  • social media
  • local partners

We prioritize keywords – (cluster approach) compared the volume of keywords along with the value of those keywords (Basically the PPC value) – map this into xy and you find the lowest hanging fruit.

Some basic keyword research lets you decide to otpimize for ’small business software’ which has 100000 searches, instead of small business solutions

link building – deep and local + via “addthis” on your site allows users to share, bookmark and link to pages35k inbound links

Tools they use:

SEO insights



  • determine your business outcomes – value your micro and macro conversions
  • do your homework on the greatest areas of opportunity
  • measure test and promote your successes – talk to your bosses–

Michael is up next

This is a case study – a fictitious company ”Tom Whortons

background – a successful online site selling coffee beans to businesses online in Canada and the use, was given 1million dollars to expand internationally

they tried to accomplish this by

  • identifying markets
  • duplicated content and design
  • translated website and keywords

results were disastrous

no sales in brazil

lawsuit in Germany

Anger from bad translation in Korea

The issues that arise

  • launched in the wrong target markets – web analytics was the only research used to determine who to market to – just because you’re already getting a bit of traffic from a country isn’t an indicator that it’s a target market
  • Automatic translation failed
  • actual translations: cafe du cancer, non-homo milk, straight tea
  • ignored global seo issues – online payment disaster, only accepted credit cards
  • launched out of sync with other geographic areas budgetary chaos

how to address the challenges

  • Redefine target markets
  • Evaluate content / created a translation strategy
  • solve search and technical issues related to SEO

Redefine target markets:

new target market calculation method –

local competition, actually look at how it is done locally now +

technical factors – is it hard to put a server in china, etc +

web analytics +

keyword research data is most important – finding out how much traffic is available

all of that = new target market

challenge 2 – evaluating content

if local successful competitors are built around quality points, not price points, perhaps you should consider imitating – in Germany price points and looking cheap is a negative, not a positive.

Ecommerce options of the good competition –

the style and length of descriptive content

the way location trust indicators are displayed

find a ‘localization specialist’ – which is different from a translator. they need to be native to the country, not just someone who speaks the language. They also have industry knowledge, and because they’re culturally involved, some linguistic knowledge. Anybody always has to use third party proofreading. They may use semantic based machine translation, which is different that simply autotranslation.

challenge 3 – keyword research specialist

so put a search marketers and a localization expert in the same room together – you wont easily find someone who is good at  both, but b uild a team.

global keyword development

  1. plural vs singular
  2. dropped accents
  3. common mistypes caused by keyboard layouts
  4. broken compounds – mother-tongue, alley-way
  5. alternate characters
  6. inflections
  7. linguistic dialects and diversity

‘coffee bean’ in French by Google

‘grain de cafe’ – not bad

keyword specialist says – cafe en grain also, often without the accent on cafe – POW three new phrases to optimize for that no machine will tell you

solving some SEO technical issues

  • servers and data-centers – having a server in the US doesn’t always help, the best recommendation is have a box in every country, though that’s difficult sometimes
  • Domain names should be country-specific, it will help rankings, but also build local trust
  • Duplicate content – often a huge issue in international SEO – best thing is to change-up the entire content of the site
  • Global payment solutions – different countries have different ways to pay – china uses money transfers, for instance

Centralized project management

have a single point of contact for centralized communication  of

  • global search marketing teams
  • advertising agencies
  • collaborative tools
  • global best practices

take aways

  • use multiple avenues of research to identify your target markets
  • don’t trust a translator, ever
  • flollow global seo best practices – hosting, tlds, avoid duplicate content
  • align your processes

Leave a Reply

Your email address will not be published. Required fields are marked *