Search 2010: Five Game Changers
May 7, 2009
Editor’s Note: This is the outline of Stephen Arnold’s comments at the “debate”session of the Boye 09 Conference in Philadelphia, Pennsylvania, on May 6, 2009. The actual talk will be informal, and these notes are part of the preparation for that talk.
Introduction
Thank you for inviting me to share my ideas with you. I remember that WC Fields had a love hate relationship with Philadelphia. Approaching the Curtis Building, where we are meeting, I realized that much of the old way of doing business has changed. I don’t have time to dig too deeply into the many content challenges organizations face. If the publisher of the Saturday Evening Post were with us this afternoon, I think Mr. Curtis would have a difficult time explaining why his successful business was marginalized; that is, pushed aside, made into an artifact like the Liberty Bell down the street.
I have been asked to do a “Search 2010” talk twice this year. Predicting the future in today’s troubled economic environment is difficult. Nevertheless, I want to identify five trends in the next 20 minutes. I will try to take a position on each trend to challenge the panelists’ thinking and stimulate questions from you in the audience.
Let’s dive right in. Here are the five trends:
- Darwinism and search
- Real time search
- Google’s enterprise push
- Microsoft’s enterprise search
- Open source
I want to comment on each, offer a couple of examples, and try to come at these subjects in a way that highlights what my research for Google: The Digital Gutenberg revealed as substantive actions in search.
Search and Darwin
The search sector is in a terrible position. The term “search” has been devalued. Few people know what the word means, yet most people say, “I am pretty good at search.” That confidence is an illusion. The search sector is a tough nut to crack. Well known companies such as Mondosoft and Ontolica found themselves purchased by an entrepreneur. That company restructured, and now the “old” Mondosoft has been reincarnated but it is not clear that the new owners will make a success of the business. Delphes, a specialist vendor in Québec, failed. Attensity orchestrated a roll up with two German firms to become more of a force in marketing. A promising system in the Netherlands called Teezir was closed when I visited the office in November 2009. I hear rumors about search vendors who are chasing funding frequently, but I don’t want to mention the names of some of these well known firms in this forum. Not long ago, the high profile Endeca sought support in the form of investments from Intel and SAP’s venture arm. At Oracle, the Secure Enterpriser Search 10g product has largely disappeared. The strong survive, which means big players like Google and Microsoft are going to fighting for the available revenue.
Real Time Search
What is it? The first thing to say is that real time search is a terrible phrase. Riches await the person who crafts a more appropriate buzzword. The notion is that messages from a service like Twitter fly around in their 140 character glory. The Twitter search system at http://search.twitter.com or the developers who use the Twitter API make it easy to find or see information. A good example is the service at http://www.twitturly.com or http://www.tweetmeme.com. You look at Tweets (the name for Twitter messages) and you scan the listings on these services. Real time search blends geospatial and mobile operations. Push, not key word search, complements scanning a list of suggested hits. The mode of user interaction is not keyword search. This is an important distinction.
“Search” means look at or scan. “Search” does not mean type key words and hunt through results list. It is possible to send a Tweet to everyone on Twitter or to those who follow you and ask a question. You may get an answer, but the point is that the word “search” does not explain the value of this type of system for business intelligence or marketing, for example. If you run a search with the keyword of a company like Google or Yahoo, you can get information which may or may not be accurate or useful. You will see what’s happening “now”, which is the meaning of “real time”.
The Google
Google’s push into the enterprise is becoming more robust. Search vendors are quick to point out the shortcomings of the Google Search Appliance. The GSA is priced to help ensure that it is used for smaller content collections. But the Google OneBox API and the Google Apps Premier make it possible to do industrial strength integration, geospatial, and data management projects. Google can no longer be dismissed as an aberration. Google is mainstream with its Oracle tie up, its major accounts such as the White House, and it growing cadre of partners. Think hundreds, not a few dozen.
Microsoft
Microsoft Fast ESP has not yet delivered a fully integrated product. At the FastForward 09 conference, Microsoft announced a road map. In the future a solution to the 50 million document glass ceiling will be available. Today integration and quite a bit of work is needed to implement the system. In addition, there has been a revolving door for executives and a constant realignment of technical resources for the many “flavors” of search at Microsoft: Powerset, Fast, Vista and Windows 7, SQL Server, etc. The confusion in product lines and brands are visible symptoms of a fragmented, heterogeneous technical approach to search. Customers are pulled because of the low cost of initial acquisition of SharePoint, an amalgamation of legacy, third party, and new code. But the complexities of the system may overwhelm some licensees appetite for taming the unruly herd of components. Nevertheless, there are 100 million SharePoint licenses in organizations today, so SharePoint and its to approach search are going to be issues in most organizations. Here’s a diagram of SharePoint with its search component (in itself quite complex) from Impac Systems. My article in Beyond Search provides more information at http://arnoldit.com/wordpress/2008/06/28/sharepoint-placemat/
Open Source
Open source has not become the preferred way to reduce enterprise software headaches, but right now open source search systems such as Lucene and SOLR and FLAX are getting longer and harder looks by large organizations. The reasons are varied. Some organizations want to get out of the vendor’s walled garden and have control over the source code. Other organizations are resisting the increasing cost of getting support, maintenance, and customization support from the vendors or authorized resellers of a product. In the UK, Lemur Consulting has a number of high profile accounts and I learned recently that the company had a banner year despite the tough financial climate.
Wrap Up
There are other themes and trends. But I have to limit my remarks due to time constraints. Let me make three broad statements in closing and then turn the program over to the panelists:
First, search is difficult. Let no one convince you otherwise. It is not technology that is the problem. Too many people overlook the need for sound management. You can read the recommended management methods in Martin White’s and my new study “Successful Enterprise Search Management” at http://www.galatea.co.uk. Quick fixes cost money, some jobs, and in the case of one health care company lives. Search must be approached with appropriate resources.
Second, the vendors can use marketing lingo that makes every system seem to be the same. That’s not true. There are deep and fundamental differences. For example, Exalead and Perfect Search have architectural innovations that allow these firms to handle certain types of data operations more efficiently than some other firms. Endeca’s core system works well as an ecommerce system, but Endeca is not Google in terms of technical set up. The marketers use the word “search” which makes every system seem identical. These systems are not the same. The wrong system of a particular job yields an expensive problem.
Finally, vendors want to lock customers in. The push to vertical integration is designed to make it very tough to reverse out of a search system. Who practices this method of keeping customers? Google, IBM, Microsoft, Oracle, and SAP, among others.
The trends, therefore, are no significant change in the difficulty of making information findable.
Stephen Arnold, May 6, 2009
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
