Finding Solutions

Business Management asks Andrew McKay, General Manager and Senior Vice President, Attivio, Inc. and Laurent Simoneau, President and CEO, Coveo about the value, potential and future of enterprise search.

“Not only is content scattered across multiple systems, but users now also need to access this content from multiple devices.”
-Laurent Simoneau, President and CEO of Coveo

BM. The value of search as a true enterprise platform has been touted for years, yet few organizations have seriously embraced the opportunity. Is this a technology issue, a problem with awareness of the benefits or a lack of understanding regarding how best to go about an implementation project?
AM. Three reasons. The first is, organizations not knowing where all the unstructured content lies in the enterprise and which of it has any value. Content management systems are easy to create and generally departmental. As such, there is little central control, and many stake holders are involved. In contrast, business intelligence systems are tightly and centrally managed by technical necessity, so enterprise-wide investments are a more natural fit.

Second, enterprise search tries to sell solutions but delivers technology. Business owners are tired of hearing 'we can do it', which leads to costly implementation projects. Search as a platform is important, but in today's economic climate, it must first demonstrate it can deliver solutions easily and efficiently.

Third, search is simply not enough. True insight resides in the integration of structured and unstructured content; bridge technologies do not work. Until this happens, search will always be the 'poor cousin' of business intelligence.

LS. First of all, you have to recognize that this is a technology that's still in its relative infancy. To date, the perceived market leaders have been seen as complex and costly, and the platform players lack some important functionalities because they have concentrated on the infrastructure play at the expense of addressing specific business pains.

So, the typical current responses to enterprise-wide RFPs have either been too general, because no one is really able to cover everything perfectly and in a timely fashion, or too limited in terms of performance, scalability and/or functionality to be considered for complex projects. Added to that, the requirements gathering, design and deployment factors frequently make an enterprise acquisition so complex that they slow down or kill the whole project.

BM. The ability to analyze unstructured data is a key - both in terms of meeting compliance targets and gaining a competitive advantage. What are the benefits of a converged enterprise search/business intelligence solution in this regard?
AM. The cost of non-compliance does not care about structure, and less than full coverage means increased exposure. But full coverage is expensive because it requires two incompatible technologies - one each for search and business intelligence - even though the rules and processing are the same. Convergence brings technical simplicity and consistency: there is one integrated implementation.

Convergence also brings deeper insight into customers and markets through integrated analysis. Consider the request, 'give me all the articles that are favorable about my top 10 selling products.' Search can solve one half of this request; business intelligence the other. But neither discipline easily solves both. Now imagine adding your call center, CRM and web analytics data. Integrating call center transcriptions, customer history and order data provides more personalized, responsive service. Integrating web analytics, sales data and market intelligence helps create intelligent product bundles that increase average order value.

LS. Well, the first thing to recognize is that, while the content ecosystem has all sorts of data, end users don't know or care how data is stored - they just need to get as complete a picture as possible, as easily as possible. So the first and foremost benefit of a converged solution is giving them exactly this in a just-in-time fashion.

These 360-degree views require structured and unstructured content scattered across multiple, heterogeneous systems: CRM, ERP, email exchange server, financial systems, intranets etc. With unbiased search technology, that is designed to handle all of the various kinds of data, users can get complete information when they need it; for example, through the kind of hybrid query that Coveo can execute: search emails for customers that are related to product X and have more than Y$ in the pipeline.

BM. In your opinion, what are the most exciting developments/capabilities currently being integrated into the enterprise search function?
AM. There are several incremental advancements in linguistics and text mining, and although they are interesting academically, they do little to address the problems that plague enterprise search today. Most studies show that half of enterprise search users are dissatisfied with their implementations. More dramatic improvements are called for.

Two innovations worth noting are SQL and workflow. To truly integrate structured and unstructured content, the search engine must treat structured data with the same flexibility as business intelligence. This means allowing on-the-fly ad hoc querying, which really means supporting SQL, notably JOINs.

To be more relevant to the enterprise, search must get in front of business decisions before they happen. Providing alerts and low-latency index updates are important, but interacting with business processes requires managing sequences of events. This is accomplished through a workflow engine.

LS. In a word, mobility. Not only is content scattered across multiple systems, but users now also need to access this content from multiple devices. Consider traveling executives who require an instant, 360-degree view of the customer that they're about to meet. Their devices are Blackberry or Windows and they must access the CRM, the sales intranet, the support database and their mailboxes to find the right information.

Enterprise search engine technology already provides an unbiased layer on top of multiple enterprise repositories. Beyond that, the search interface is the single point of access to mobile-enable dozens of different applications in a seamless, secure and simple fashion. Coveo enterprise search products have built-in mobile interfaces that provide exactly these kinds of capabilities.

BM. With the rise of Web 2.0 tools and technologies, content creation, management and analysis is easier and more accessible than ever before. What impact is this having on the enterprise search market, and how do you see the market evolving over the next 12-18 months?
AM. The rise of Web 2.0 tools will accelerate the proliferation of content, demanding more from search technologies. We will see more collaborative data, owned by specialized groups rather than individuals, adding greater understanding of context and leading to further specialization of search requirements.

The line between intranet and Internet will disappear (other than for privacy) as enterprises depend more on the web and the growing migration to SaaS for critical information. The search and business intelligence markets will begin their merge as they both become more operational in real time. Deep analysis will become even more important, and continue its push towards better predictive capabilities.

LS. The classic intranet and CMS ecosystem is now complemented by enterprise blogs and wikis. Because these new repositories have a different structure and a different set of meta data, enterprise search technologies need to add more granularity to the social relevancy algorithms, provide new ways of navigating through the results, and treat content from behind the firewall and in the cloud as a unified resource. Because the content is organized more and more around persons and users, it makes sense to navigate through dynamic expertise networks, built from the context of a search query.

Extra info...

Value of enterprise search
The key value for enterprise search is productivity gains for employees:

• A recent Accenture survey of 1000 middle managers found they spend as long as two hours a day searching for information and that more than half the information they find during searching is useless.
• The Butler Group claims as much as 10% of a company's salary costs are wasted through ineffective search.

According to IDC, the amount of information created and copied in 2010 will surge more than sixfold to 988 exabytes (988 billion gigabytes), representing a compound annual growth rate of 57%.