AODC day 2: Optimising your Content for Google Search

This week I’m at AODC 2010: The Australasian Online Documentation and Content conference. We’re in Darwin, in the “top end” of Australia. This post is my summary of one of the sessions at the conference. The post is derived from my notes taken during the presentation. All the credit goes to Joe Welinske, the presenter. Any mistakes or omissions are my own.

Joe Welinske presented the last session on Thursday, titled “Optimising the Googleability of Your Content”. The subject was search engine optimisation (SEO). Joe is really excited about this topic. He thinks it’s one of the most important skills for a user assistance expert to acquire. Potentially, it also offers a good opportunity for consultancy.

SEO is complex, but we need to learn about it. Joe calls it “embracing the beast”: If people are going to be using Google, then we want our material to come out at the top of the search results.

These are the topics he covered:

• Get all your content on a public-facing server.
• Learn about Google Search.
• Learn about SEO (search engine optimisation).
• Find out how to use other search engines.
• Create a custom Google Search for your UA (user assistance) material.

Getting your content onto a public-facing server

This is the single most important thing that has to happen: Your content has to be on a public-facing server.

There are ways to mirror your internal content onto public-facing servers. Joe talked about a few of them. Ultimately, the best way may be to migrate all content to web-based standards.

Security issues when publishing content on the web

Many organisations keep their content behind the firewall, or have privileged content requiring login. Security is a very real concern. Moving to a public-facing server may be a big change for traditional content developers, and a challenge for organisations that have up to now kept their documentation hidden from public view.

Organisations need to take a detailed look at all the help components they provide and see what can and what can’t be published on a public-facing server.

We should be aware that our users are using Google anyway, and we should ask ourselves what these users find when they do a Google search for information on our products.

We could put together examples of Google searches, to show management what people will be finding, given that they can’t find our own documentation. For sure, people will be posting information about your systems to answer other people’s questions.

Google search indexing

Note that some content will not index effectively. Joe gave some tips on the type of content that Google can index well and the type that may cause problems.

Good candidates for Google indexing:

• Web sites.
• Web-based help, such as output generated from Flare or RoboHelp. Note that frames may make the content less favourable for indexing. Google has a hard time making sense of content in frames.
• Eclipse Help is a good candidate, because it’s XHTML.
• PDF is good, because Google has put effort into indexing it. Some PDF formats may be less effective.

What does not work well:

• Microsoft Help (CHM and HLP) — This format has some HTML but is mostly locally-installed and uses a lot of proprietary code.
• Apple Help, Oracle Help for Java and JavaHelp.
• Flash — Google has done some work to be able to index text in Flash files, but images and videos not.
• Almost any proprietary markup — Google has problems interpreting it.

Bottom line: The more open standards you use, the better off you will be.

What about loss of context?

When we put together online help systems and context-sensitive help, we have control over the path the user follows. But when the user comes in via Google search, we don’t have that control.

Some tips from Joe:

• Provide navigation elements on all pages.
• Add branding, so that people know that they have found your site and thus the official documentation.
• Include the date last updated and the version of the application that the help covers.

Examples of SEO

Joe showed an example of a Google search he entered. He asked how to print a calendar in Excel. The first result returned was a link to Microsoft Office Online, telling you how to print a blank calendar. This was exactly what Joe wanted to do. What’s more, the content was an exact copy of what’s in the HTML Help provided with Excel, but laid out differently. So Microsoft has their SEO sorted out, and they are mirroring local help content in the web-based help.

Similarly, he entered a Google search for printing a Google calendar. Again, the Google web-based help was right at the top of the results. Google provides only web-based content, so they don’t need to support mirroring of content.

In contrast, when you search for “print an ical calendar”, there’s nothing near the top from Apple. So Apple needs to work on there SEO in this particular context.

How Google search works

It’s a combination of brute force and extremely clever technology:

• Brute force: Google runs a gigantic server farm in Washington, that buys electric power directly from the hydroelectric dam nearby.
• Extreme smarts for indexing information: Google has clever and secret algorithms for ranking content.

How does the ranking work?

Joe mentioned the following factors that influence the ranking of search results:

• Fresh and real content — The search index recognises a page as offering a solid body of information, not just fake content. This is an area we technical writers excel at. Link farms out there put together random text and try to fool Google. But our standard documentation will match up as high quality.
• Nomenclature — We use the right terms and the right labels in our documentation. If people use these terms in their searches, our content will quickly match.
• Links pointing to our content — In Google’s eyes, links indicate what other people think of the information on the page. When many people link to pages, this makes it more likely in Google’s eyes that this is well-respected content.
• Google Ads — There is some indication that your pages will be better indexed if they contain Google Ads.

Joe gave us these sources as guidelines on optimising your content for search:

• Google webmaster guidelines
• Google ranking factors, from Vaughn’s Summaries

Joe had a number of more in-depth hints on things like:

• Including the “robot.txt” and “sitemap.xml” files.
• Registering your site at “google.com”. (It’s free.)
• Including metadata such as title, description and keywords.
• And more. I’m sure Joe would be delighted to pass on this information if you contact him.

A tool to help analyse your pages for SEO

Of huge benefit is a tool called Go Daddy Search Engine Visibility. It goes through your pages and finds any search engine deficiencies. There are other tools that do the same thing.

Other search engines

Joe touched on other search engines, such as Yahoo and Bing. These are also important. Each has a separate registration procedure.

Getting other people to link to your site

Send people links to your content, such as sending links via email. Submit links to relevant sites. It’s a tough process and there’s no easy way to do it.

You can also think about community building:

• Start a LinkedIn group.
• Add YouTube videos with links in the metadata.
• Add a Facebook fan page.
• Tweet links via Twitter.
• Encourage bloggers.
• Supply RSS feeds.
• Consider translated content, such as having even just a single page about your documentation, in different languages. Google indexes different languages too. Also, people who speak that language will at least learn about your documentation and your product. They can then click through to read the information in English.

Creating a Google custom search

It’s easy and free to register to create a custom Google search. Register your site. Google supplies you with the code you can use to put the search box onto your site. Google will index your site for you. Your search box will bring up results for your site only.

My conclusion

This was intensely interesting and full of information. I could not hope to capture everything that Joe told us. Thank you Joe! This presentation gave me a lot to think about and experiment with.