”Search” vs. “Searching”: Bringing Enterprise Search Back to Reality
Silvija Seres is frank about the issue of “search as pure-play.” “We have had conversations at FAST about distancing ourselves from the word ‘search,’ even though it’s an important part of our name. People think of search as being simple...type in a box and get results. Yet we’re thinking of search as being a whole new organizing and collecting principle that can be a whole new platform strategy. The meaning of the word ‘search’ will rise up from the search-box connotation,” she says.
Dennis Moore, general manager of emerging solutions at SAP, also thinks search occupies a different category
than other software tools. “Search is not a standalone thing; it’s the way information workers work. Search is a means to an end; it’s not an end. In one year we will still be talking about ‘enterprise search’ in its own category. But in five years, there may be one or two vendors who OEM engines for enterprise search to other companies. But it will remain an essential feature of any
serious enterprise platform. The platform vendors need content management, they need BI, they need application server, they need lifecycle management...and they need enterprise search. That will make it a very competitive environment,” Dennis predicts.
“The market is dividing into two directions,” says Matt Glotzbach. “There’s the large platform vendors who, not surprisingly, serve cross-functional search needs that can be used for any type of application. But there will continue to be a collection of niche vendors serving specific needs. By no means will the need for search applications in ecommerce, or BI, go away.
“Nearly every application that has data or information involved already has some search function in it,” Matt continues. “The greater need is to tie together and mix the various types of data. As much as we’d like to see a homogeneous environment within an enterprise, the reality is that it’s going the exact opposite way.”
Jerome Pesenti takes it a step further: “A vendor can’t specialize in a ‘vertical market’; there’s no such thing. No two companies are alike. Each has its own repositories, etc. You can’t offer them an out-of-the-box solution. It has to be suited to that specific company, not the vertical market it’s in. That’s what the specialty vendors bring to the table.”
Jerome continues, “It’s a matter of the specificity of granularity that a specialty vendor can provide. A customer that has a large but very similar product line with thousands of pages of documentation for each has a unique challenge. If you just threw a search engine at it, you’d get garbage from the output. You can’t just return a thousand documents as a result...that doesn’t work. You need the technology and expertise to adapt to that environment.”
SAP’s Dennis Moore also identifies a shift in the understanding of the function of search. “In my job, I look at ‘what’s next.’ And what’s next is empowering information workers. Perhaps the synergy of innovation in the consumer space and interest in the corporate worlds to empower the information worker has come along at the same time and created an interest in enterprise search,” he offers. “Originally, search was a way to find a document. But now the questions are more like Who’s in charge of this? Who’s the expert in that? Or How many of these do we have in stock, and how much would they cost if they were delivered on this date and those terms...? Enterprise search is the coming together of search and BI. You do some searches, then you also conduct some kind of action or transaction based on them. That’s very different than what consumers do on the Internet.”
FAST’s Silvija Seres interprets the difference between commercial search and enterprise search in lofty terms. “Search is the common language between customers, partners and employees. The flexibility you get from a powerful search system allows you to tailor the way you present information in a way that’s very unique to your business. It consists of a lot of preprocessing and postprocessing of the information, but it results in something that your customer or employees find best suited for their needs.”
Silvija further bisects the search markets: “Enterprise search should not work like consumer search. You would give up incredibly valuable context information. Where consumer search works best is in the hierarchy of documents on the Internet, using the implicit information within the documents’ links. You don’t have that same structure in documents within the enterprise. But there is a lot of other business information that should be exploited. There’s a completely different workflow that should be tailored for the workers.
“CIOs assume it’s easy, whereas it’s actually a very careful analysis exercise, where you have to be careful about data, about how it’s going to be used, how it needs to be preprocessed with special tags and classifications. We have a customer who uses search as the organizing principle for their infrastructure—search is the bridge among information sources, and search is also the expression language on top of that,” says Silvija.
Johannes Scholtes of ZyLAB adds another spin onto the conversation: “Embedded platform search and free-search in particular are much like the meteor that killed the dinosaurs: there’s a broad availability of free personal and open-source enterprise search tools that are able to perform a very serviceable array of tasks for many users. Buyers and vendors know that enterprise search platform providers are going to struggle in a market where the ‘good enough’ part of many solutions is free.”
He continues, “Think back to the enormous impact the Windows NT free search engine had when it was introduced in 1993. Many vendors have gone out of business since then, except for the ones that adapted to the changing marketplace. Companies that survived in 1993 and that will survive in today’s market are search vendors with specific technological niches, or vendors that focus on search-driven applications in which specialized search capabilities were integrated within an applied business activity—contract management, e-discovery and e-disclosure, litigation support, email management for compliance, historical archives, project and correspondence management in litigious environments and so on.”
What Makes Search So Special?
Naturally, this panel of mostly specialized search vendors has large stakes in the game. But they make a compelling argument that search has unique characteristics in the business environment that might not meet the eye of Joe SixPack.
“Producers and consumers of information are now one and the same,” says Michael Schmitt. “That puts stress on the technology underpinnings, to bring content in from a variety of sources, and make sense of all of it to empower the end-user experience. But there’s also stress on the editorial component. There’s brand image at stake, so there needs to be consistency and editorial credibility in the site. It requires an entirely new set of skills to address that, to relieve all the tension spots around that circle.”
Johannes adds, “For most common users, search used to be a separate key component of infrastructure, but over several years almost all ECM, CRM, ERP, HRM and other vendors have signed OEM agreements with the major search vendors. Since platforms from Microsoft, Oracle, IBM, etc., contain search out-of-the-box, the only search appliances that survived are ‘search-driven applications’ for which the standard OEM or platform search does not work. This can be because the data is of a special nature, or because one needs cross-platform search or because more is required than just search: for instance integration with security, data input and data output,” he explains.
“The market has essentially split into three segments,” says Jared Spataro of Microsoft. “At the high end there are specialty solutions. At the low end, commodity solutions. And in the middle, an emerging segment we’re labeling true ‘enterprise search.’ We expect basic search capabilities to commoditize, further fueling explosive growth driven by ‘bottom-up’ demand. The specialty segment will continue to focus on deep niche markets. The enterprise segment will be hotly contested by platform vendors, and this competition will eventually drive consolidation of the enterprise search decision into the broader decision for a more comprehensive platform for business productivity. In many ways, the same dynamics that have played out in the portal, collaboration and document management markets will play out in search,” Jared predicts.
Dealing With Complexity
Paul Sonderegger is VP marketing for Endeca, and as visionary about search as anyone I know. “To say that ‘search’ is at the root of all things is not quite right. Searching is at the root of all things. The difference is that searching is a basic human activity; it’s the way we solve problems.”
OK, but in the business context, doesn’t that lead to the application of tools? And the standardization on the ones that seem to bring the best results?
“That is a dead end,” insists Paul. “That’s what the original search vendors did wrong. Search is more than just the ‘search box.’ And by concentrating on the search box—which is a small corner of what people use to find information—the search box got relegated to a piece of functionality instead of what it really is, which is infrastructure. If IT departments concentrate too much on one piece of functionality to support search, they will be in the same trouble spot.”
Paul continues, “What they should look at are the many ways in which people gather information, and figure out how to most efficiently make tradeoffs based on what you know about the information. The piece of functionality that supports that is different for each of the different business processes that that action is embedded in.”
Google’s Matt has a cautionary message for the market, too. “The trap that enterprise search as a marketplace needs to avoid is assuming there is a difference between the way consumers behave and the way information workers behave. In the consumer space, it’s well-known that there’s a Darwinian aspect...if something isn’t working, users will try the next thing and there’s zero switching cost. Enterprises tend to think that’s not the case with them. But it’s a fallacy that users will accept the ‘IT-sanctioned’ tools and that’s that. That’s wrong; enterprise users will find their own ways to get things done.”
Matt Schmitt says the emergence of a “community mind-set” (my words, not his) adds further complexity to the matter. With external sources, such as social networking sites, influencing corporate behavior, the challenge is good old knowledge management, he says. “How can I take information from internal and external knowledge plus feedback about it from multiple sources, automatically draw relationships within those sources and share that knowledge? In effect what you create is a core competency within a community of users who have mastered the discovery of up-to-the-minute news that is pertinent to them. That is knowledge management at its best.”
“The search box will persist,” says Endeca’s Paul Sonderegger. “But the difference is that it will have to be fully integrated with the engine that cranks out other individual views. The search box won’t disappear, but it will be relegated into its appropriate function and integrated into other information-gathering applications.”