A tough truth about complex and integrated software applications is the lack of expertise and professional depth available to implement and maintain them. This explains a lot about why small business units and project teams often find and deploy their own software tools to get work done.
I am particularly concerned at the lack of will by organizations to fund implementation of applications like enterprise search to aggregate at the retrieval-end the content stored within disparate applications. No rational business planning can justify having workers sift through multiple repositories, each with a separate sign-on, search interface, and search engine protocols just to find a single document. True, organizations need highly competent professionals to meaningfully implement, tune, and administer enterprise search engines. They require the expertise of search analysts, taxonomists, librarians, IT specialists with security, platform, and software development training. However, developing a team of six to twelve “search engineers” to give workers in a thousand person company quick access to relevant content is an ROI no brainer when we know workers waste significant (5 – 15%) amounts of working hours hunting for stuff.
This week’s Information Week article by Nicholas Hoover on Web 2.0 contained a comment about Wells Fargo “…on another Enterprise 2.0 front, integrated search, the company has limited employees’ ability to search across data repositories because of the complex authorization schemes needed to keep people from accessing information they shouldn’t.”
Today’s (Feb. 27th) New York Times headlined with a story about Microsoft buying “a specialized search engine tailored to deliver useful medical information to consumers,” Medstory, Inc. The story goes on to cite comments by Esther Dyson who refers to the technology as “an ontology engine.” This underscores another truth about quality semantic (natural language) search; it depends on the existence of meaningful, topic specific vocabulary and concept maps to work well, a complexity in narrower markets.
Finally, we have seen the recent shift of companies like FAST moving from a strategy of selling solutions directly to enterprises for the purpose of aggregating content through a unified search portal to focusing on niche markets and highly tailored search architectures.
These are just three cases of a shift among search companies to leverage their search technology IP in more lucrative offerings. The losers will be organizations that really do need to deliver content more holistically to workers through a single search engine. Yes, security is a concern, and skilled search technologists must be hired and dedicated to delivering search options that tie directly to business operations.These efforts are not one-off projects but need to be sustained as permanent infrastructure. If you are in a position to influence search procurement solutions make your case for the most suitable software that will really help deliver the best retrieval option company-wide. Be realistic about funding and staffing; then go for it. If Enterprise Search is what you need, make sure that is what you get and deploy.
I agree with the observations in your post which we here at Recommind are also seeing as trends in the marketplace. You touched on the topic of security, which is certainly a concern that must be addressed when discussing enterprise search and categorization products. There are solutions on the market today that not only provide their own layers of user and information security, but also respect and enforce the native security controls at the application/database layer. Put more succinctly, today’s leading solutions can and should provide more security, not less, than the “siloed” approach they replace. And make no mistake about it – a “security by obscurity” approach is no longer a viable defense to shoddy information/records management and access practices, as the fallout from the recent revisions to the Federal Rules of Civil Procedure is making all too clear. The essential decision enterprises must make is this: should they continue trying to “hide” information with minimal ability to police and audit what is actually happening with their information, OR should they make their information available to those who need it, in context, while implementing, enforcing and auditing the access and use rights their business requires? Invariably, enterprises are seeing that the latter approach is the only route they can take to remain competitive while effectively managing the risk inherent in today’s business environment – and is an approach which should also, somewhat ironically, drastically improve the security of their information.
But it is important for companies large and small to realize that working through these issues and finding an acceptable “information access and management policy” which can be implemented, enforced and audited is not only a healthy exercise, but an absolute necessity in today’s competitive, highly litigious and compliance-driven market. An enterprise search implementation is the perfect time to undertake such an exercise; after all, if employees are a company’s biggest asset (as most CEOs are wont to say), why would any company refuse to make the most useful information available to them in the most actionable way possible? Isn’t that a little bit like buying the fastest car you can afford…and then refusing to put any gas in it?