Printer Friendly

Microsoft: a search phoenix.

For a year or more, Microsoft has been developing a new information retrieval system. This findability solution was codenamed in the context of a super-secret development activity as "Oslo." I perked up with the reference to Oslo, a city in Norway. Microsoft bought Fast Search & Transfer, a Norwegian company based in Oslo, in 2008 for $1.2 billion. In September, Microsoft rolled out Oslo as Delve, the search service for Office 365.

According to the Free Dictionary, "delve" means to search deeply and laboriously. Microsoft has worked hard to minimize the laboriousness often associated with finding information to answer a business question. The "deeply" portion of the definition may foreshadow the richer functionality of Azure Search.

Deep Dive

Delve can index content and make it searchable for an Office 365 user if that content is stored on OneDrive (Microsoft's cloud storage service), Yammer (Microsoft's social networking system), email, or SharePoint (Microsoft's CMS). Planned enhancements include support for email attachments, content created in OneNote (a tree-form note-taking program), and Lync (a messaging and collaboration system). Julia White, general manager of Office 365 technical product management, says:

   Delve displays information
   that is most relevant for each
   person based on the work they
   are doing and the people with
   whom they are engaging. With
   Delve, information finds you
   versus you having to find information.
   All of this is presented
   in a beautiful, card-based
   design that is easy to
   understand and use. It also
   presents intuitive ways to view
   content, so you no longer have
   to remember where stuff is
   stored or who shared it with
   you. And, of course, Delve only
   enables you to view content
   you have access and permission
   to view....

Delve is a slight departure from the old-school keyword approach of Fast Search & Transfer, which became available 17 years ago. Microsoft moves along the "discovery" trajectory. The idea is that a user does not have to guess the secret combination of words that will unlock an answer. According to a Microsoft Office 365 blog post by Ashok Kuppusamy, a group program manager in Microsoft's FAST engineering team, "The goal for Delve was not just to re-imagine search, but to help people get their work done in a quicker, more informed, and even delightful manner" ( 11/introducing-codename-oslo-andthe-office-graph).

Smart Software

Delve incorporates Microsoft's Office Graph technology, which makes information access more intuitive by discerning connections and context. Combined with Delve's ability to personalize the retrieval service, smart software promises to reduce the frustration some search systems trigger in users. Graph supplies the intelligence behind Microsoft's search system. According to Kuppusamy, "The Office Graph uses sophisticated machine learning techniques to connect you to the relevant documents, conversations, and people around you.... Through the Office Graph, Delve is automatically populated with activities you already do every day, such as which documents you share, which people you meet with, and which documents you read."

Microsoft is leveraging innovations from its FAST division as well as from Yammer. According to a Microsoft Office blog post by Jared Spataro, Microsoft's general manager of enterprise social:

   First, we're extending Yammer's
   concept of the Enterprise
   Graph across Office 365
   to create something we're calling
   the Office Graph. When
   we acquired Yammer, one of
   the things we loved about the
   technology was how it mapped
   the relationships between people
   and information by simply
   recording likes, posts, replies,
   shares, and uploads. In
   what I think is a game-changing
   move, we're extending this
   idea to Office broadly, using
   signals from email, social conversations,
   documents, sites,
   instant messages, meetings,
   and more to map the relationships
   between the people
   and things that make your
   business go. Just as consumer
   social networks connect you to
   people and information you'd
   otherwise never be able to
   keep up with, the Office Graph
   will plug you into the very fabric
   that makes your organization

Delve can tap Yammer's message translation capabilities, improvements to file collaboration, platform updates such as embeddable feeds and the Like and Follow buttons, and continued development on various mobile platforms.

One notable innovation in Delve is the interface for trends that are taking place around an individual user and his context. The idea is that Delve takes note of who the user is, his information behaviors, and other signals. Armed with this knowledge, Delve proactively assembles information that's likely to be relevant to the user without the user having to formulate a keyword query or navigate from application to application to locate needed information.

Not surprisingly, the inner workings of the Delve system are not directly accessible to the user. Developers, however, can interact with its components via an API. Delve promises to be a wellspring for Delve-savvy Microsoft-certified engineers and firms. My initial impression of Delve is that it is a search application built on the Azure Search system.

Federated Search

Of particular interest to me is Delve's inclusion of federated search. The idea is that a user needs access to diverse types of content. In the course of answering a business question, she will want to know which colleagues may have information that's pertinent to her needs and will want to access information in different file types. Some information may reside in a Word document. Other information may be in an Excel spreadsheet or a PowerPoint presentation. In its initial form, Delve manipulates a handful of file types but cannot process email attachments. As the product matures, additional connectors and conversion filters will become available. In an organization or a small business, important information resides in various file formats. Thus, federation becomes more useful when it can process a wide range of file types, such as database content and legacy system outputs.

The idea of providing a single search box or a single discovery interface to a professional is an important one. A year ago, my team and I interviewed members of a 20,000strong trade association. The top request for a search and retrieval system was to have one search box that would deliver results from different sources. The users did not want web access. I checked our data to refresh my memory of their wish list:

* Office document content created in standard office applications such as Microsoft Word

* Adobe PDF files

* Information in the organization's SharePoint system and its OpenText RedDot software

* The membership directory, training programs, and conference information stored in databases

* Videos of the organization's conferences

* Email with attachments such as spreadsheets, links to webpages, and PowerPoint presentations

Federated search is available from specialist vendors. BrightPlanet, Dassault Systemes, Deep Web Technologies, and XML data management system vendors such as MarkLogic, among others, provide ways to run a query across different content sources. However, each of these systems has limitations. If Microsoft can deliver a rich federated search system, the company could leapfrog many enterprise search competitors.

Imagine a version of Delve that can access content in Microsoft SharePoint, a content management system on steroids. With a single point of access, a user can shave minutes or hours off of some research tasks. If the Delve personalization works in a manner similar to Amazon recommendations, a SharePoint search will learn from the user's actions automatically and in the background. Microsoft's smart software can assemble information that is likely to be of interest to a particular user and in a form that's more useful than an old-school, Google-style results list. The idea is that the user can "discover" or recognize information. As our research revealed, many users think that crafting a query that unlocks information hidden behind a search box is frustrating.

Universal Search

Federated search has been one of the underpinnings of Google since 2007. Google dubbed its federated search "universal search." In a 2007 blog post titled "Universal Search: The Best Answer Is Still the Best Answer," Marissa Mayer, who was then Google's VP of search products and user experience, wrote:

   Back in 2001, Erie [Schmidt]
   asked for a brainstorm of a few
   'splashy' ideas in search. A designer
   and product manager
   at the time, I made a few mockups--one
   of which was for 'universal
   search.' It was a sample
   search results page for
   Britney Spears that, in addition
   to web results, also had
   news, images, and groups results
   right on the same page.
   Even then, we could see that
   people could easily become
   overwhelmed with the number
   of different search tools available
   on Google--let alone those
   that would be created over the
   next few years. This proliferation
   of tools, while useful,
   has outgrown the old model
   of search. We want to help
   you find the very best answer,
   even if you don't know
   where to look.

In 2007, I thought that if any organization could make universal search a reality, it would be Google. Today, Google requires searchers to use different indexes and run queries in these silos of information. Most users are unaware that Google's universal search remains a goal. If a user wants to run a query on Google Books, he has to navigate to Google Books. The same requirement applies to Google's patent index. If a Google user wants to search the index that's created when Google indexes blogs, the task is not something the average person will find intuitive. I have to check my notes in order to run a query that's limited to blog content.

The process involves navigating to the Google News page (another content collection outside the reach of universal search). You can find this page at Next, run a query, any query. When the results appear, Google displays a secondary series of hotlinks beneath the search box. Select "Search tools." That click displays another series of hotlinks. You will need to click on "All news" and then click on "Blogs." For my queries, the blogs of major media properties such as ABC News and The Washington Post were consistently the most highly ranked results. To get nonmainstream blog content, I had to fall back to the time-consuming process of reading result pages, clicking on items, and then reading the source documents. I may not understand universal search, but the blog search process is an exercise in universal frustration. Has Google's approach to search become "dumb"?

Some federation is possible via the Google Search Appliance (GSA), but savvy GSA specialists have to fiddle with the software widgets to allow it to present a single list of results. Along with the license fees for the GSA, the engineering work adds to its cost. Although some search advisory services stress how economical the GSA approach is, its cost in a medium-sized organization can easily hit six figures. To get some sense of the basic license fees for the GSA, go to and run a query for Google Search Appliance. The first hit for me returned a GSA 5005 (an entry-level device) for backing up an online GSA 5005. The discounted license fee is $286,143.69. Tailoring the GSA to handle disparate content types adds to this base cost.

Reaching Beyond the Search Box

Why do Google and its Google Search Appliance as well as many search vendors fall short in their effort to deliver universal search or federated search? Microsoft has learned from its experience with Fast Search & Transfer. It has also acquired technology that blends social content with basic information retrieval. For example, if a user wants to know about people who may have information that's relevant to her query, Delve delivers a relationship display.

Delve does not abandon the search box, but it does push the keyword approach to finding information to the sidelines. The emphasis is on discovering and revealing information that's relevant to a user. When the user looks for information, the Delve system hits the ground running, providing information that seems to anticipate what the user wants or needs. Information is presented, not searched.

How important is this shift from keywords to a present-and-discover approach? I think that many users will welcome the new strategy. What happens when Delve migrates to SharePoint? Today, a number of search vendors compete to license software to SharePoint customers who want an improved search experience. If Microsoft's discovery approach is successful, it may force many of these vendors to refine or develop entirely new add-ins for their SharePoint customers.

In my opinion, Delve is a search cocoon. Finding information becomes more comfortable. The sharp edges of a Boolean query and the drudgery of clicking and browsing through results lists are mostly removed. The Delve interface may deliver the summit for users' findability needs. Microsoft is one of hundreds of vendors that have tried to reinvent search. Maybe this time it has out-Googled Google and other enterprise search vendors to boot.

Stephen E. Arnold ( is a consultant residing in Harrods Creek, Ky. Beyond Search, his widely read daily blog, is located at
COPYRIGHT 2014 Information Today, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Arnold, Stephen E.
Publication:Information Today
Date:Nov 1, 2014
Previous Article:Conferences around the world.
Next Article:Super STEM sources.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters