Search and SEO: What's Happening

With so many great applications available and enhancements on the horizon, I am consciously taking a step back to survey the landscape.  Our teams have been hard at work building and improving things like search refinements, code search, autocomplete, and so much more—all to help you find exactly what you’re looking for on MSDN and TechNet.  There is one piece that I haven’t mentioned yet, however, which has the potential to greatly impact discoverability—and that is our newly initiated Search Engine Optimization (SEO) campaign.

For those relatively new to the term and the mythos surrounding it, search engine optimization is the process of optimizing your site’s content, architecture, quality and quantity of links, and more to build two key things: Relevance and Trust.  As you start to filter through the wealth of SEO guidelines, blogs, forums, and hearsay that I do on a daily basis, you’ll see those two terms pop up quite often.  Relevance and trust are what search engines (in a more contextualized manner) use to display search results (and rank them) once the information is indexed by the search engine.  (Are you still with me?  Good!)  The key here is that good SEO is by and large a process of simply making your site better for your customers…which is to say, my primary concern (and dare I say, shared by all of us here).

It’s no secret—we host a huge amount of content and resources on MSDN and TechNet.  It’s a major feat in and of itself for any search engine to parse all that data…but the major initiative we are working on in the coming months is to optimize our content to build relevance and trust in our own site search, as well as in other search engines.  The goal of SEO is to optimize our sites and related content so that (combined with our enhanced site search platform) it’s ultimately easier for you to find exactly what you’re looking for.

The SEO campaign will primarily be focused on the following:

·         Optimizing our page titles so you get a better match in search for the query you enter

·         Adding descriptions of our content so you know more about that result on the search results page

·         Building links to our content and sites so that they are indexed in search engines faster, bringing you fresher results

·         Optimizing our site architecture to enhance discoverability of related resources and surface only the most relevant content

We are working to build and implement on SEO best practices across the board.  We believe there is a great opportunity to build trust within our sites and relevance of our content to bring you better results in search.  As we move forward, I’ll be sure to note improvements and (as always) I’ll be looking for feedback…

Have you ever been stuck in the middle of a coding project and needed to find a class definition or perhaps a function call?  In an effort to provide a better experience for our customers looking for code on MSDN, we have just launched the MSDN Code Search Preview.  You can try it out here.

Why Code Search?

MSDN site search provides results from content across MSDN and related websites.  But what if you’re looking for code, specifically?  The MSDN Code Search Preview leverages a user interface that best displays code results and also allows you to narrow your search to code samples.  After entering in a code-related query (such as “createfile”) you can then filter your search by language and more to find the code sample you need and help you troubleshoot and complete your project.

How Does MSDN Code Search Preview Work?

To use MSDN Code Search Preview, just visit MSDN search.  By entering in a code-related query and clicking “Code Search Preview” you can search for code within the MSDN Library. 

There are some great features to take advantage of in the MSDN Code Search Preview.  Just a few of these features are described below:

• Advanced search.  Use the slider to choose an exact match or a “fuzzy” match to your query.  Add a language and project filter and you can quickly narrow your search.
• Search by language.  Looking for ASP.NET examples?  Coding in C++?  Filter your search using the “Language” drop-down list.
• Search using qualifiers.  Filter your search by function calls, definitions, and more.
• Search by technology area, called “project” in Code Search.
• Create a permanent link to your results.  Use the “Create Link” button to create a permanent link to your search results to share with others or come back to them later.
• ...and more to come! 

We’re currently planning to include more code results from sites like MSDN Forums, MSDN Code Gallery, and CodePlex.  You can also tell us what you think of the MSDN Code Search Preview and provide additional feedback by participating in the MSDN and TechNet Search Feedback Forum.  Stay tuned for more updates…

In a series of upcoming search "bake-offs," I'll be comparing Google results to our own site search results.  Up first is "volume activation" which provides a good example of TechNet site search demonstrating up-to-date results above and beyond Google's.  If you are looking for anything related to "volume activation" and specifically "volume activation 2.0" you'll find that Google's results still yield results from the old location at technet.microsoft.com/en-us/windowsvista/:

It's also assumed that a search for "volume activation" that provides me with a step-by-step guide for troubleshooting as the first link isn't ideal.  I'm more than likely looking for a "hub" or general information and resources on volume activation from Microsoft.  If you look at TechNet results, you'll see a much different results set that provides you not only with the correct link for volume activation in the TechNet Library (http://technet.microsoft.com/en-us/library/bb892849.aspx) but also other resources expected based on that type of query, such as the Overview Guide, Deployment Guide, and Planning Guide:

Stay tuned for more bake-offs on MSDN and TechNet!

In case you haven't seen this latest post on the MSDN & TechNet Search Blog, I thought I'd point out the latest improvements with MSDN & TechNet site search (also rolled out on Expression).  Our engineering team has been working hard on improving our existing search features--namely our autocomplete feature.  Autocomplete enable you to view terms that relate to the keyword query you have entered into the search box as you type.  Previously, autocomplete was limited to around 30k terms, so the terms that appeared in the search box drop-down were relatively limited.  We've now expanded that list to over 2.4 million terms. From the Search Blog:

"...this means you're at least twice as likely to get a completion when you're typing a search term, and the number of completions you see will be much longer...Note that we haven't improved the autocomplete in the masthead of msdn.microsoft.com and technet.microsoft.com pages yet--for now the 2.4M-term enhanced autocomplete is only on search results pages."

Stay tuned for more updates to MSDN and TechNet site search.  Check out the rest of the post for more info:

http://blogs.msdn.com/msdnsearchblog/archive/2008/05/28/released-improved-search-autocomplete-and-more.aspx

By publishing your content to the internet, you are by default releasing your content to the world.  The fact that content on the Web can be read worldwide in a matter of nanoseconds is one of the most amazing phenomena to come out of our lifetimes (I think).  But the challenge well beyond your writer's block is that you are writing a letter that begins "Dear world..." and sending it to over 6 billion people. 

Okay, not all of those people are fortunate enough to go online--and maybe that's not such a bad thing (Note to self: find out where all my time went) but of those online readers, how many are you going to reach based on the context and quality of your content?  Perhaps there's some value in region-biased search engines (my Live.com search is set to French and thus, will prefer results in French if I choose to do so).  And your content may be best-suited for a specific region.  But with search engines offering more integration with machine translated pages, how does your content hold up on the other side of the world?  If you've got good content, chances are that not-so-little unknown audience is out there looking for your content--and if you haven't made your content "O.K." you could be losing out on a large piece of the pie...

So what is O.K.?  By making your content "O.K." I don't mean to speak to mediocrity.  The phrase "O.K." or "okay" is virtually universal.  It is commonly understood in most languages (sometimes with a slightly different pronunciation).  The idea is simple: Make your content universal without making it boring.  Fortunately, most "good content" is interesting enough based on the audience's perception of that topic.  If I'm looking for information about how fruit ferments and the article I found is clear and concise and addresses my initial inquiry, I will naturally consider that piece of content to be good content.  That same piece of content transcends this "good" state to become O.K. when it yields a universal contextual relation between each component of the sentence.  (That was a mouthful.)  Perhaps some printable Binaca will help explain...

Here is not so good:
The server is a jack of all trades, running multiple virtual processes, as follows:

Here is pretty good:
The server is a jack of all trades.  It can run multiple processes in the following virtual environments:

Here is good and O.K.:
The server's capabilities are robust.  It can run multiple processes in the virtual environments listed below:

What makes the content "good" in the second example is that the sentence structure clarifies the context of "running" better for machine translation by clarifying that "the server" should be associated with "running."  The first sentence fails to do this and a machine translation could potentially associate "jack" with "running" (which may or may not be followed by "to fetch a pail of water").  What makes the content good and O.K. in the third example is that the sentence further clarifies the meaning of "jack of all trades" for the reader.  The last example is universally contextually relevant and, despite the edits, makes me want to read further.

Try doing some spring cleaning on a few top level pages and monitor the search results in relation to your site traffic...especially international site traffic (if you can get that data).  I think you'll be pleasantly surprised. 

As a writer, editor, search enthusiast, and techie, it has become apparent to me that the notion of "engineering good content" is quite foreign to many of us and I'd like to take a few minutes to explain what this is, why it's important, and how to go about it.  In the end, I'd like to set the stage for many posts to come...

First, there's a notion out there that a good search engine will yield good search results.  I'm going to step forward and say that this is not necessarily true--and in fact, it's becoming more and more false with the clutter of site content that exists in the Webbed world we live in today, where search results are in direct competition with search engines. 

To the point, this is the essence of "engineering good content" for the Web.  We no longer use the Web in the same way that we would to look up information in an encyclopedia.  (Remember that old, dusty Britannica set?  Remember your first online experience?  If you were like me, lost like a child in the woods, you may have used them similarly.)  In fact, we are forced to use the Web much like a database manager would--if that database manager were able to make queries without any rules (perhaps a terrible nightmare for the "database people"). 

It's important that we address this change: There is so much information that exists online, that we are inundated with the good, the bad, and the ugly when it comes to search results.  The search engines are burning coal like a steam engine stuck in the muck to force "good" results from often mediocre content.  The key here is that search engines are starting to become more like people in a way--they will look for a topic, based on a query, in a logical, sensical way (the infamous algorithm).  The problem is that the topics it has to crawl through are simply created and casually organized...and voila--that's the end of the road for that topic (and consequently, often the last time the writer ever sees that topic again).  This approach keeps us hanging in the chaotic, monkey-with-typewriter world we're in today.  To evolve is to focus on engineering this content beyond its inception--for search engines and the people that use them.