In the very real world of business IT, the concepts, topics and products marketed and sold in isolation are not generally employed as such in practice. Our research repeatedly confirms what any CIO knows; the true skill in exploiting technology lies in making it work in context with all the other systems, processes and tools already in place.
And so it is with search technology, which despite having been around for a while, still feels like a tool or capability seeking the right level of understanding to be applied most effectively. From the outsider's perspective, many people may assume they understand search ‘because of Google'. However, Google is about finding ‘anything' and arranging it into some loose order based on whether Google thinks it is important, a phenomenon familiar to most.
In many business scenarios, it all needs to be a bit more deterministic. A person looking for something needs a degree of certainty that they will find as much as possible of what they are looking for, even if they didn't know exactly what it was when they started. This is a very different starting point, requiring far more precision than the shotgun-based approaches of the 'Net.
So what is going to work in business when it comes to search? We know from our research that exploiting information in a business context is about finding everything relevant, dispensing with the irrelevant and getting the right information to the right person. This is an important starting point when we think about how to consider implementing search.
Looking at the challenges of exploiting information assets more broadly, our research has identified ‘classification' and ‘trust in results' as two of the major hurdles to overcome. The two are linked. If you have no way of ensuring that all relevant information has been included in a search, then by default you have no idea if what the results turn up is what you actually wanted. In the business context, users need more than a formula-driven lottery based on the sheer quantity of information available. If search doesn't turn up the goods, it can quickly become no more than processor-hungry shelf ware.
So can you get into a position to exploit search? The answer is yes, but it requires starting with knowing what your information needs are, rather than hoping that a tool can mine your fragmented pools of information and make sense of it all, or indeed attempting to classify every single item of information in the business just in case it is of value.
For example, by engaging with sales, marketing, field service, product development and customer services teams, you can build a picture of information requirements: from the specific items they need and the way they are used in the context of business processes, to the systems they use to input, process and retrieve information. This may span a range of activities from communicating with customers, looking for advice from remote locations or creating sales or marketing collateral.
Here is where thinking about ‘information classification' in the right way comes to the fore: understanding your requirements can help you identify key information assets and provide an objective means with which to categorise them from the ‘top down'. Thinking about ‘classes of information' from this requirements perspective will help you identify the information assets you need to treat as a priority. We are not saying this is a new approach - at its heart, it is called good librarianship. What we are saying is that it is important to put the horse before the cart.
Working through things in this way will also allow a more informed choice to be made when it comes to choosing the right tool to deliver against the requirements you have identified. It is worth avoiding the temptation to dive straight in. The accessibility of underlying information repositories and the links between them must reflect the use cases, roles and information types you have identified. Then comes the (long) list of must-do's: for example ensuring that security, identity and access management, compliance and data protection policies are compatible with the strategy, rather than being elements to work around.
Little, if anything in business-oriented IT today matters more than the ability to exploit information assets. Trying to gain, or improve this in isolation of all the other processes and communications that keep a business running may doom an initiative from the start, or severely restrict the additional value it can deliver. Adding a new capability on top of an un-optimised business and IT infrastructure doesn't solve anything.
Indeed, getting one's house in order first is the only way to ensure that investment in new processes and tools can deliver what was intended. Taking a top-down approach might feel like going the long way round, but the answer to ‘What are we trying to achieve?' lies here, not in a toss-up between different search tools. Unless we take that longer route, search could remain either a well kept secret for those in the know and a poorly understood tool not quite able to deliver on the promise for everyone else.