My last post was about unifying search across multiple domains to create a better search experience for your visitors. It’s no secret that real estate businesses typically have multiple domains, so I thought it fitting to write about the challenge of unified search here at Active Rain. However, any business with two (or more) domains is challenged to create a unified search experience, so this series of posts is important to every business segment.
At MyST we spend more time innovating than any other activity including selling; our CFO reminds us of this all the time. But we’re on a never-ending quest, and our recent research into custom search engines has uncovered some startling revelations that will change how we think about search.
Background
Until recently, our research and product development has focused primarily on helping business and marketing professionals create a sustainable and dominant presence in search engines and other Web 2.0 systems that help people find information. Our objectives have been constrained to produce content optimized for discoverability in Google (worldwide). I use the term “worldwide” to indicate the entire Google index; the one most people on the planet look to for recommendations.
As I mentioned in a previous post, few months ago we started to see a pattern; website (and blogsite) visitors that were increasingly dissatisfied with the second phase of search; this is the search process that begins after Google (worldwide) has recommended a handful of domains based on a given query.
The Two Phases of Search
The topology of the web has been slowly transforming into a sea of domain expertise nodes – clusters of smart and very focused sites created by individuals, teams, businesses, governments, clubs, social networks, etc. Blogs, websites, and social networks are the predominant implementations that enable domain expertise to surface on the web rapidly and become more findable. But the emergence of clusters representing focused expertise has created a new challenge – with all these loosely-coupled information silos, how do we manage the ability to search across multiple [disparate] applications without including the entire public web?
The search process is typically a two tiered experience; for most people it starts with a generalized search that includes more than 100 billion pages on the Internet (and growing). People have learned that to get close to what they're looking for, they must focus their keywords and use more of them. But the results generally point to pools of content that are likely to contain answers. The second phase of the search process begins once a pool of information is recommended; typically this requires a deeper search within a company's domain. But (as noted above) the definition of a “company’s domain” has shifted with the advent of multiple domains and loosely-coupled information sources. This is where the trouble begins and the visitor experience may suffer.
While Google may find reasons to highly recommend one or two pages in the same domain for a given query, it typically won't consider all related company domains or sub-domains within a given search result. In fairness, Google wants to provide a diverse selection of good recommendations – this is a sensible approach. Imagine a single company dominating the top ten results for a particular query - Google users wouldn't find this behavior very useful. Given the way Google works, searchers may be able to get close to the content they want, but the second phase of the search experience (which begins when they land on one of many possible domains of a company) is typically lacking in many ways.
This is not surprising; search solutions are usually implemented differently from domain to domain, and the variety of publishing, content management, and web applications in use by a company tend to exacerbate search continuity challenges. Few systems have the ability to easily integrate or scope searchable content across specific collections of domains. Visitors [however] have far greater expectations – they expect to find answers within the domain(s) that Google recommends. In many cases, the first recommendation may not be exactly what the visitor was looking for, but most people are patient enough to search one more time using a site search feature. This represents a sizeable opportunity to engage the visitor, but if you want to capture attention and keep visitors from using the back button to go back to Google (worldwide), a unified search experience for 100% of your company's resources is necessary. The more resources you can offer them for easier discovery, the more likely they'll continue to investigate your content.
Participating in the new topology of search will require comprehensive organization of your customer-facing content, and Google is providing precisely the framework to make this possible. Everything you’ve published (website, blogsite, social network, forum, reservations tool, partner content, online catalog, etc.) matter even more. Access to this focused sphere of content becomes more critical the larger the Google index grows (e.g., as the size of Google’s worldwide index increases, the ability to find information [directly] decreases). The trend concerning custom search engines is clear; it's a new extension of a proven idea that will radically shape how we find good content, and it’s already happening – people now search for good places to search for what they truly want to consume.
The Revelation
The second phase of search must be agile and dynamic. In the realm of custom search engines, there are domains, facets, and refinements and each must be configured in a context. Simply stated, if your business content spans five domains, you need a search engine that includes all five domains. But more important – the context may change depending on each visitor’s experience and wants.
To that end, our research forced us to coin the phrase “fluid search”. Before you hop over to Google and type that in, let me save you some time – there’s only 5,700 references; most concerning hydraulic and bodily fluids. Fluid search is a simple idea – create a user experience that possess enough agility to keep your visitors engaged. Here’s an example…
Imagine a visitor has landed on a specific blog post (from a Google query) about a new community being built in your town. While the post performed exceptionally well by ranking high in Google (worldwide), it’s just a start as far as the visitor is concerned; she wants more information and immediately clicks on a new unified search component feature that’s embedded right in the post. It says – “Click here for a deeper search on this subject.”.
With one click the search application provides a list of all posts in the blogsite and website related to the subject that attracted the visitor through Google. In this example, the list of resources is pretty thin because it’s a relatively new community. However, the search application has anticipated the possibility that broadening the scope of the search to include all pages linked to from the blog might produce additional interesting content, and it does.
The originally unified search application has expanded its scope to find three additional stories that were linked to from other blog posts that weren’t closely related to the original query. This is an important point worth introspection –
The search engine has magically transformed itself (on the fly) to consider off-domain pages that the collective businesses’ domains have linked to about the community in question.
This is the definition of a fluid search application. Building a search experience like this is not easy; it requires a bunch of stuff that we don’t have time to go into. Besides, the business requirements are far more interesting to consider.
We’ve learned that our customers want the experience described in the example, and we’ve built the infrastructure to achieve it. More important – we built it in a way that seamlessly integrates into MyST Blogsite services which are easily upgraded across our services. We also designed the technology to integrate with non-MyST blogging tools and web applications. Drop me a note if you’d like to be considered for beta testing this technology.
Comments(0)