Enterprise Search: X1 Argues Search and Discovery Are the Cure to Findability Ills. Maybe Not?

January 26, 2015

I read a white paper from a search vendor called X1 or X1 Discovery. The company was incubated in the same hot house that produced GoTo.com. As a result of that pay to play model, Web search was changed from objectivity to advertising. X1 search, if I understand the white paper, Why Enterprise Search Fails in Most Cases and How to Fix It (registration from this link required to access the paper) and the companion article “X1 CEO Message: A New Approach to Enterprise Search Resonates” is the future of search.

The fix is an interface that looks like this:

image

Source: “Why Enterprise Search Fails in Most Cases and How to Fix It,” page 3.

In the “X1 CEO Message” I noted:

So in view of this customer and industry feedback, we coined the phrase “business productivity search” to differentiate what X1 focuses on verses most other enterprise search tools, which are typically re-fashioned big data analytics or web search appliances. And the feedback we’ve received on this from end-users and industry experts alike is that this assessment hits the nail on the head. Business productivity search is not big data analytics and it is not web retrieval. It is its own use case with a workflow and interface that is tailored to the end users. X1 provides the end-user with a powerful yet user-friendly and iterative means to quickly retrieve their business documents and emails using their own memory recall as opposed to generic algorithms that generate false positives and a workflow ill-suited to business productivity search.

I am not convinced that search and discovery as described is going to address the core issues that plague enterprise information access. Specifically, the last few decades have beaten keywords to death. The users have expressed their views by grousing about whatever keyword system is provided to them, finding alternatives to keyword search, and shifting attention from keywords to more actionable interfaces provided by a group of vendors largely unfamiliar to the keyword crowd.

There is a role for keyword search, but that utility function can be provided via open source solutions ranging from FLAX to Lucene to SphinxSearch and other options.

What is not provided is the automated collection, analysis, and report functions of the next generation information access systems. I have explained the characteristics of the next generation information access systems in CyberOSINT, described at www.xenky.com/cyberosint. In this study, I profile more than 18 next generation systems, provide a schematic of the functions included in these systems, and provide examples of the outputs these NGIA solutions provide to their users.

What’s interesting is that each of these vendors supports keyword search in some way. Just as a modern automobile provides a lever to display a turn signal, NGIA systems include utility functions. But—and this is a big “but”—the NGIA systems address the needs of the user. The idea is that the user, without trying to guess the keywords that unlock what’s in an index, provide actionable outputs. A dashboard is one option. More useful outputs include dynamic PDF maps with data displayed on a mobile device. The maps update ass the information arrives or the user moves around. There are outputs that show the key players in a deal and provide one click access to supporting data. No search is required. Many of the NGIA system operate in a predictive manner. When the user looks at the device, the information is “just there.”

I appreciate the efforts of vendors like X1, Coveo, Attivio, and IBM Watson in their attempts to breath new life into keyword search. Just as the old marketing essay about buggy whips made vivid to tens of thousands of MBA student, when the automobiles appear, the buggy whip outfits may want to make seat covers.

The fix for enterprise search problems is not more keyword and point and click suggestions. The solution is a shift to the NGIA approach. And that shift, whether traditional vendors of search grasp it, has already begun.

Stephen E Arnold, January 26, 2015

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta