New Info Mix: Search and Enterprise Publishing

October 31, 2008

The 20-somethings who want to rescue enterprise search from the swamp in which it is lost have to look at another player in the enterprise information game. The technology in question focuses on printing “dead tree” documents. These documents are mailed to millions, often tens of millions of people. Here’s an example. You lease an automobile. Each month the financial institution or entity brokering the deal sends you a statement. The real estate on these paper documents can be put to better use that reminding you to pay within 20 days. The financial institutions want to remind you to refinance your house (there’s a great idea) or remind you to take the vehicle to so and so’s dealership for the 6,000 check up. Some financial institutions want to greet you personally, with a cheerful “Hello, Tabitha.”

Time Magazine, the Wall Street Journal and other “dead tree” experts have figured out how to personalize certain print runs of their publications. Print personalization in the regional editions of Time Magazine have been around longer than some of the snappy Web personalization services. Advertisers like the idea of buying a specific demographic. Financial institutions, electric utilities, health care insurers, and other firms that mass mail invoices to known customers want to personalize as well.

A Web content management system is a toy compared to the industrial strength systems required to crank out the personalized invoices for a company like Honda, Prudential, or one of the few remaining banks that can pay for postage. The systems to handle this procedure have to connect to various back office systems, provide graphic tools, offer work flow controls to move information object A to financial analyst B and then to invoice C. Mistakes when it comes to billing customers are not acceptable. Web content management systems are able to produce Web pages, but it takes a broader, more comprehensive system to perform industrial strength enterprise publishing.

Now here’s the surprise that most of the enterprise search companies are ignoring. These systems include search functions, but it is a utility within a broader, mission critical enterprise system. This is not the world of the marketing department or the Web specialists. These systems make the day for the organization’s chief financial officer, regulators, and attorneys.

typical output

Source: http://www.exstream.com/Applications/Telecommunications/bundledservicebills.asp

One of the early players in this enterprise publishing game was Optio Software, which was acquired by Bottomline Technologies this year. Streamserve, a company based in Sweden, is a leader in this sector. IBM and Ricoh hooked up and created InfoPrint last year to pursue this market sector. Xerox has long been a player with its now-long-in-the-tooth “Docutech” systems. Exstream Software, headquartered in Lexington, Kentucky, sold to Hewlett Packard earlier this year for about $1.2 billion.

HP promptly hooked Exstream into the company’s printer division, and announced on October 27, 2008, that HP has inked a deal with Graph Expo, a large UK printing outfit to produce digital information systems that can spit out a book or perform certain types of custom publishing at high speed, with great programmability, and at lower cost than existing methods. HP and its shareholders learned a while ago that there’s real money in printer ink. In fact, I describe HP as an ink vendor, not a computer company.

You can read the HP news release here. I haven’t seen much coverage of this announcement, but there may be some pick up in the printing-centric Web logs and trade publications later today and whenever dead tree publishers get around to covering the story. On the surface, it’s not too exciting.

I think that dismissing this story is a bad idea, particularly for companies in the search, content processing, and text analytics business. Here’s why:

  1. Most vendors of enterprise search have not entered the enterprise publishing sector. Some of the firms with which I have had contact are generally unaware of these systems, their inclusion of search as a utility, and the systems’ ability to output Web pages, reports, and invoices. This cloud of unknowing is one that should be dispelled but the ostrich approach to business is often a favorite of search vendors, their advisors, and the conference organizers who seem indifferent to this major shift in enterprise information systems.
  2. Enterprise publishing systems carry hefty price tags. Because the systems are mission critical and make it possible to cross sell or run ads in most output from the system, seven or eight figure deals are not uncommon. Enterprise search and content processing systems that purport to index “all information” for the organization may gain credibility in some parts of an organization, but at the CFO level, enterprise publishing gets the attention of the woman who writes the checks.
  3. The end-to-end model seems to becoming popular. I may be reacting to news stories that flow through my intelligence system. But Autonomy has emphasized its big deals like the one with an unnamed pharmaceutical company. That’s a big deal and it appears to be an end-to-end play. Brainware just nailed a $2.0 million deal for an end-to-end solution. The Exstream Software deals are, as I understood the briefing I got earlier this year, end-to-end.  The question becomes, “Where do specialist search, content processing, and text mining companies fit in?”

I have watched Mark Logic grow in the last couple of years. The firm is careful not to stray far from its positioning as an XML database. But, when I look at the firm’s demonstrations, I see a company that is poised to challenge some of the bigger players in the enterprise publishing market. Will the company knock off the HP and Exstream Software duo? I don’t know. But when I hear about findability solutions, I think about meta-plays from companies like Mark Logic, HP, IBM, and Streamserve, not the old school chatter about latent semantic indexing, taxonomies, and clustering. The real money is in what I call “metaware”. Watch Beyond Search for more on this “space”.

Stephen Arnold, October 31, 2008

Comments

2 Responses to “New Info Mix: Search and Enterprise Publishing”

  1. John Rueter on October 31st, 2008 11:36 am

    Thanks for pointing out StreamServe in your blog post, Stephen! I’m so glad to see that people are starting to think about enterprise document publishing. In many respects this is the last frontier of business processes that are finally becoming automated and integrated within the organization. Not only are there substantial cost savings from saving trees by going from print to electronic, but companies are realizing tremendous benefit from the customer opportunities represented by placing target messaging on bills, statements, invoices etc. that are appropriate for the customer.

    While leading the marketing efforts at one of the big enterprise search firms, I always used to say that finding information is a good start, but the real question is: What do you want to do with it? That is where enterprise document publishing comes in….whether e-invoicing, personalized customer correspondence, automated distribution across all communication channels…

    Document composition, processing and production has always been a backwater in the organization, people haven’t really paid much attention to the costs, because it is a “cost” of doing business. The promise of enterprise publishing is that you can take any input and distribute to any output, reducing operational costs and fostering better customer experiences. The benefits are real, the results are tangible.

    This is actually an exciting space to be in and for me it feels like the early days of enterprise search, when we were trying to establish and demonstrate its value to the business world.

  2. Stephen E. Arnold on November 1st, 2008 9:22 am

    John Rueter,

    No problem. Feel free to send me info about Streamserve and what is new at the company.

    Stephen Arnold from a country with no vowels in its name, November 1, 2008

  • Archives

  • Recent Posts

  • Meta