The #SearchLove Chronicles – Day 2 – Afternoon Sessions
Big sessions kept on coming at Searchlove 2012. After the first morning and afternoon sessions which were more about strategy and the morning sessions on the second day which were more about tactics the afternoon again saw more in-depth tactics.
Lisa Myers: International SEO. One Size Doesn’t Fit All
International SEO is a bit complicated, if you don’t understand it.
Country code domains
ccTlds is the best way to go ig you have an office (or will have) in the relevant country and it is a growing business. Especially if you have a marketing budget in those countries (or in house capabilities).
Having a marketing department in a country and using a subfolder is not really suggested by Lisa.
And remember the strong signal any cctld have.
They should be used when:
- budget restrictions and no local marketing
- Sites that will never have local offices
- Initial feo targeting and time restraints
- Informational sites
Not use it for eCommerce. Canonicalization can be a dramatic issue in International SEO.
- Technically in the middle between ccTlds and subfolders but not the ideal solution
- Easy to implement and can have separate hosting but don’t benefit in whole of the link authority on main domain.
Webmaster Tools for geotargeting
Going in Settings you can geotarget your domain/subfolder/subdomain.
But pay attention. Are you really targeting a country or maybe are you targeting a language? John Mueller told us that 25% of sites do wrong geotargeting.
Using multiple sitemaps
- Create separate sitemaps for different folders on the same domain
- Then submit them to webmaster tools and set the geotarget setting.
- follow this link: http://goo.gl/qWW1n
It is the new way of saying “It’s the translated page innit” or “it’s suppose to be indeed in this country“.
The Hreflang attribute is simply a way of telling the search engine this is a different version of a page.
When to use it?
- You translate only the template of your page
- Your pages have broadly similar content with a single language
- Your site is fully translated
Hreflang can be implemented as a HTML link element in the header, but now more easily it can be implemented using Sitemaps (recommended).
The Mediaflow has created a tool in order to create these sitemaps (check their site)
Classic… correct URL in SERPs, but the Title is from another “country” mirror of the site. This an issue when you target same language but different countries (i.e.: US and UK).
- If the content is largely similar, Google doesn’t change the title and meta information
- So you still need to have content that is unique!
Hreflang + Canonical
Quite confused communication.
- In most cases there is no need to use these in conjunction with each other for geo targeting issues
- Use it if you want to launch a new site, as Andy Atkins Kruger recently suggested on Search Engine Land.
The language issue
Language is important. So target differently when it comes to different languages.
Easy to understand, but so many times not considered by businesses.
If you want traffic Germany you must speak their language. Funnily enoigh the search engines also believes that a site that has content in German might also be targeting Germany. Pretty obvious, but obvious things are not paid attention to.
Please, don’t use big Brands as a model for International SEO.
For instance, Nike uses IP detection. NOOOOOOO
- Several big issues with IP Detection but mainly the problem that bots usually crawl from US IP, so therefore might never see your other content
Speak their Language! But don’t directly translate and for F#%* sake don’t use Google translate for it.
You should have a budget in each country, who deals with correct translations.
Having links from other sites in the language (and ccTld) that you are wanting to rank in is Hugely important. It’s like a neighbour saying: Yeah, they live close to me.
- Native speaking people, that get’s SEO and most importantly can write WELL
- A good project management tool
- Great SEO tool (as SEOmoz and Linkdex as Lisa uses)
- Streamlined reporting on links
Market are different. Methods working well in UK maybe don’t in Scandinavia or any other country. Also the competitiveness is different.
Raise the profile of content people. Get your content outreach people also to blog on clients websites.
- Don’t limit yourself to specific subjects
- You can reuse ideas or even translate articles (the magazine industry do it ALL the time). Obviously giving credit when necessary.
A strategy in International SEO must alway consider a Platform Level SEO, which both Global (for all languages) and Local, for the localization.
- Recruiting native speakers
- Organisation of link development
- Reporting (especially value of links)
- Making sure everyone writes to the highest standard in languages I don’t speak
Paul Madden: Building an Outsourced & Automated Infrastructure from Scratch
Advice. It may be better not to do what we hear here but to learn from it. It is probably best not to do Blackhat SEO.
Paul Madden narrates his history as an Adsense farmer. It seems he had quite an easy life a few years ago. He decided to outsource and automate his link brokerage, for instance. On the positive side we can learn how we can scale operations from his experience with that and not necessarily in a black hat sense.
His example (pay attention to the process): Blogger outreach. Without process nothing works. All automated systems start with a robust manual process. Try to “deskill elements” and make it so simple even he (Paul) can do it. Find sites, contact, manage responses, place link or content. Make this more and more granular until you have every step covered well.
How to find sites to outreach to, are they ok? How do we approach them? and so on.
Split it into stages. Some tools to help you:
Find sites > Human approach
Contact Sites > Scripts
Manage and place a link or content > Human
Data collection via bookmarklets.
Why? No toolbars. No need for tech support. Easy to maintain. Hiring a team you’ll never meet. Human Resource sucks, says Paul. So go to oDesk. Pay attention on how you create a job posting in oDesk. Don’t get too specific. Simple postings work best.
Remember: hire fast, fire fast… give good but vague feedback if you have to let someone go early. Who to choose? India, Pakistan, Philippines and Bangladesh. All these four countries are good for basic research and data collection tasks.
Oh… don’t rely on the avatar.
Finally define a team.
A team does not need to know the final purpose of the job. They don’t even need to know the other teams. Then appoint a team leader and control what the team does. Create a hierarchy.
Safeguards. Secure the people via NDA who know the value of the data (For example, the team leader, the dev resource and any content resource) with a tool like RightSignature.
And now you have to train the trainer. Don’t ever mass train. Train the leader and he will train the his team. Show you’re in control until you have confidence in the managers. For controlling Paul uses the oDesk Work Diary, which screenshots the screen of each person working for you at random intervals every few minutes.
Feedback loops. Set KPIs and measure them. It’s the team leader’s responsibility to report the KPIs to Paul for comment. As everything settles down, start looking for new automation possibilities. Its always good to run a happy team, that’s why feedback from the teams is important. And try to match your expectations with theirs.
As your data builds, mine that dataset. New teams can extract more and more info and opportunities from it. Promote from within. Divide into specialist teams, for instance.
Finally get stuff coded to automate more. Hire a team not individuals on oDesk as that allows you to get a single contact who has developers who can handle multiple coding languages etc.
- AuthorityLabs API
- Bing Search API Data
Pre-qualify the data
- Twitter API
- Mozscape API
- We can try to reproduce the “spirit” and the methodology also in order to optimize our own job as white hat SEO’s.
- Especially important to create a granular and deskilled process.
- Even the team management described by Paul can be helpful especially when you have to deal with massive site issues, which require mostly automatic monitoring, reporting and adjustments.
Phil Nottingham: The Building Blocks of Great Video
Phil introduces us a real case: Kurtz and Blum. They did 53 videos for a very cheap cost per video. Videos helped organic traffic gaining 14% traffic and raising rankings.
Where does video fin in the of world of SEO?
Two main problems.
First one: Rand Fishkin… Content Marketing as the centre of the Web Marketing eco system. Video is great content! Video is not Content!!! Is a medium, a form of doing content. Images and videos are a language.
Common “we need to do video” attitude sucks.
Ask yourself is a content lose something if it were just text and image, instead of video. If video is not really nothing, than it’s better not going for it.
Second one: How most marketers see Video.
Video is Brand Awareness. Video is about much more than viral.
Rich Snippets. As we know they drive more traffic from search. So… embed the video on the page you to rank, then host on your own server or with providers like Wistia or Vzaar or VimeoPro or Brightcove.
High quality video without draining your bandwidth. HTML5 player with Flash fallback. And don’t use iframe option of Wistia o vimeopro.
Then do video XML Sitemaps. Provide Google with video metadata, define the ideo thumbnail. Use the Distilled video sitemap generator.
Use also Shema to have videos indexed.
What about conversion? A user who sees a video, especially about appliances, tends to convert more and spend more.
You must bridge the gap between Initial Interest and Conversion. You don’t must create commercials.
Empathy and Hard Facts work well in video marketing online.
But what about the costs of a professional video? Actually you can do good videos for the web without spending too much. The camera is not a problem, more are the lights and the audio equipment.
You can also use graphic animations without paying that much with shuttershock.com.
Important: remember to have transcripts of your videos. Google needs them. But also provide unique and relevant text for product pages.
Pro Tip: don’t put “conversion” cideo on YouTube:
- Cannibalizes rankings for your own site
- Hard to convert from YouTube
- Can devalue your own site
So, does YouTube suck for SEO? No… if you don’t a presence there, do you really have an SEO strategy.
Optimize for YouTube. First, consider it as Inbound TV. So it is importante to have a creative story attached to your brand..
Tutorials also work well.
So.., let’s be practical
- Use keyword YouTube tool
- Include keywords in the file name
- Include a closed caption (transcript) file “keyword stuffed”
- Optimise title, description and tags as per traditional SEO
- Include naked URL link in meta description
- Get links, views and shares
- Measure user engagement, not views! (YouTube Analytics)
Lower bounce rate for video:
- Keep strings(indents short
- Focus on quality of audio
- Get straight to the meat and potatoes
And if you put in YouTube, put everywhere else.
Links and Social Shares
Links and Social correlates quite well with videos.
How and when do people link to a video? You must think to video as part of a page type. For instance suggested is the new integrations of HTML5 and JQuery to make video part of a page and making liver and more interactive.
Make sure links only go to your site, not hosting platform. That is why is important to have pro video hosting. In Vimeo you can restrict visibility on Vimeo itself, so people must click and go to your site to view it.
Add a custom embed code on page.
Exception to the rule: YouTube playlist for Content Curation.
Remember, you need a strong page to see your videos ranking. And if you have hosted on YouTube and people embed the video but link to YouTube not your site… reclaim the link.
If you want traffic and conversion to your site host your own videos.
Patrick McKenzie: Eating CRO, Real World Case Studies for 20-100% Increase in Revenue
CRO is not a question of buttons’ copy.
Instead it is test hyphotises about customers, not about pages. And fix that which is broken.
Be aware of not burying the trial signup, not having clear call to action and to have made too difficult to purchase our products/services. Be aware of navelgazing and having a terrible UX which doesn’t make users understand what to do with your pages.
Try to have clear pricing, clear calls to action, continues selling and use social proofs.
- Obvious UX issues should be your #1 priority
- Need to have someone with responsibility and authority to monitor performance of the website
- As consultat: it is possible to meaningfully impact the culture of a client, if you can demonstrate the value of your work.
Focus on the first five minutes experience in a site to understand what’s wrong.
- Guide people’s purchasing decision.
- Ask for referrals, explicitly
- This is a major engineering challenge (months of works in some cases) but when you generate more incremental revenue than a product launch, then they can afford it can’t they.
Target Measured Funnel Weaknesses
Tip: try to fix the low hanging fruit situations in the funnel, because they can increase your conversion rate a big percentage.
Continue digging when analyzing a site for CRO.
Remember that your clients do not live on your website. They are not ready on first visit to make a big-ask decisions. They might be seaching in advance of need, for instance.
Don’t make it hard to buy.
High ROI things to test:
- 37signals style photos on landing pages
- Value proposition headlines for plans and pricing page
- testimonials aligned directly against customer objections, as close as possible to where those objections are.
And this is all, folks. It was a pleasure to write the chronicles of Searchlove this years. I hope they helped you enjoy somehow the Seachlove experience.