Daily newspapers shouldn't be
modeled on Model Ts
Current newspaper production is an exemplary model of Industrial Age manufacturing.
A single copy of each newspaper is laboriously produced by hand every single day. Copies then are mass produced on a press, and the reproductions are shipped to customers.
Then the original is discarded.
The entire process begins again the next day -- from scratch.
Oh, sure, some newspapers do what is called zoning, where multiple editions are produced with some stories changed or moved around. The Waterbury Republican-American (my paper in Connecticut), for example, is considered an aggressive zoner for a paper of its size.
That means we produce four versions of the newspaper each day -- and we can manage that many only four days a week. (In effect, we are telling our readers what Henry Ford once told his customers: "You can have any color, so long as it's black.")
Not only is the production process wasteful, it is customer-hostile. For example, even with zoning we produce a one-size-fits-all newspaper for the 175,130 residents of Litchfield County.
We do the best we can, but we lose the business of anyone not suited by our definition of the best one-size-fits-all paper.
If you think about it, the way we put out a newspaper is not that different from the way old Henry cranked out Model Ts. We start by having craftspeople -- reporters, photographers, advertising reps -- build each part pretty much by hand.
Those parts then flow through an assembly line, where line workers -- copy editors, paste-up workers, supervising editors -- do the same operation over and over (and over again) as each piece comes to them, then is sent on to the next person on the line.
The only wonder is that more of them don't get an automatic weapon and go postal.
Enter the Information Age.
The Industrial Revolution centered on manufacturing, and therefore assembly lines and other production techniques.
The Information Age is centered on data, and therefore information systems.
By definition, then, an Information Age newspaper system is centered on data acquisition and manipulation -- not production.
Data acquisition we understand; we know we need a reporter on the scene, a camera on the spot. But data manipulation?
Think of it this way: Every day we gather scores of stories. We print them.
And then we throw them away.
Oh, sure, there are forward-looking newspapers that ship their stuff to Lexis-Nexis and Dialog. But exactly how many real people do you know who can access any of those things?
Even worse, think of how much of our business is dictated by production. Think of the endless debates over news hole. Think of the difficulty any news organization has launching new products.
Even worse, think about how often we must tell our readers "No."
"I want more sports." "No."
"I want more business news." "No."
"I want more local news." "No."
"I want more school news." "No."
We all know why things are this way: It's difficult to impossible to redirect the assembly line.
So let's get rid of it.
Instead of an assembly line, we'll have a matrix of interlocked databases. Some will contain works in progress. Some will contain already published data: Text, graphics, photos, sound, video and data models. Some will contain research data: The census, crime statistics, voting results, and similar source information.
Other databases will contain finished data waiting to be published.
A master database controlling these container databases will be able to output to print, voice, on-line products, wireless products, telephony, broadcast -- you name it.
Everyone from reader to reporter will be a client for those databases. Writing a story? The production database will get it to your editor, then make sure the right version gets on the correct page.
Need a new product to meet a changing competitive situation? The database will build it on the fly.
Want more sports and less ... whatever? The database will do that for you.
Want e-mail or pager notification of events that are important to you? The database will alert you as soon as they happen.
Not just digital delivery
A database-driven production system may sound like an expensive waste until you are downloading your product to subscribers' wrist-computers and other Dick Tracy stuff, but such a system also makes possible the production of new and better print products.
You can also use it to build tailored newspapers -- editions constructed so each reader gets the sections she wants -- and targeted newspapers -- editions constructed so each advertiser gets the readers he wants.
The benefits of such a system are obvious. All news organizations throw away more material than they print -- the sheer volume of the wires makes that true -- only to see specialty publications cherry-pick subscribers using much the same data.
Think, for example, of the volume of sports data that goes straight into the bit bucket; then think of the sports specialty publications that are sold in your area.
So a sports edition seems like a no-brainer. But there are large, expensive pitfalls in this rosy scenario. In a nutshell, the problem is using an industrial age technology -- the printing press -- to produce an Information Age product.
Theoretically you could use the press to print a custom paper for each and every subscriber, but you'd sure have fun explaining the resulting sea of red ink in your budget.
You could spend a ton of money on new equipment, such as high-speed ink-jet printers that allow you to insert custom printing on the fly. With it, you could merge your subscriber names into personalized ads, making your paper look just like ... every sweepstakes letter your readers have ever received.
There's value in targeted and tailored products. But in the end, newspapers are what they are: An Industrial Age broadcast medium where one voice speaks to many listeners.
Which brings us back to new media. A database-centric system makes it as easy to output to the audiotext system and electronic products as to the press.
Much of this is possible now through the use of Standardized Generalized Markup Language. SGML is a series of tags that can be interpreted by a wild variety of output devices. There's a little code next to a headline, for example, that says "Main Headline."
On the press, that will come out as 72-point Caslon demibold, while the voice synthesizer will read that in a big bold voice for audiotext customers.
If you think about it, SGML is fairly close to the old familiar Atex or TMS coding where you'd define your main headline font as "F1" or whatever, and you'd stick <F1> into your copy and get 72-point Caslon demibold at the typesetter.
SGML is the basis of the new Universal Text Format wire coding designed to replace ANPA 1312, the wire service text transmission standard.
Yet output and production are not the same. Think of it this way: Do you think The Adventures of Huckleberry Finn is a great book?
Would you like to have it read it to you by your audiotext system -- start to finish?
That's obviously ridiculous. Print and voice are different. In other words, SGML is no magic bullet. But it eases the problem of multiple outputs.
The magic bullet is the database. The database will allow us to build products on the fly. The database allows us to do narrowcasting, the opposite of broadcasting.
Time to think about information
Information has a time component as well as a data component. This is something we already know. Think about it:
The fact that the stock market crashed in 1929 is history.
The fact that the stock market crashed yesterday is news.
The fact that the stock market will crash in an
hour ... ? Now, that's worth money.
While the production angles are interesting, it is the database linkages that are most interesting for proponents of computer-assisted reporting. A true client-server system will allow any reporter, editor or producer in any office to access any piece of data when they need it -- without wrestling with Structured Query Language or other roadblocks.
How will this work? It will work because the system is structured around our data, rather than around production. It will work through the use of client-server technology perfected by large industries such as banking and insurance.
It will work because, frankly, the technology is not that cutting-edge. (Let's remember: As an industry, journalism likes to keep its feet planted firmly in the previous decade.)
There are some who would argue that such a system already exists, and end users already have access to it: The Internet.
Not quite.
We've already talked about two dimensions of information: the information itself, and time. There is also a third important dimension: context.
Context is a rare and wondrous thing on the Internet. More than one on-line wag has said that the 'Net has become so vast and meaningless that for many people it is a substitute for life.
Context is so obvious in current media that we tend not to notice it. We know without thinking that the most prominent story in a newspaper or the first one on a news broadcast is the most important (at least in the view of the people in charge of the product) and the stuff buried in the back is ... well, the stuff that deserves to be buried in the back.
We also judge context by the messenger who brings us information. We have different expectations for the information we glean from a network television news broadcast than from an infomercial. We have different expectations for the information we receive from the New York Times or our local newspaper as we do for one of those supermarket tabloids bannered with "Wife kept dwarf lover under bed for seven years!"
But the 'Net's great virtue is also its great weakness. It is possible to search the world in moments and gather tons of information, thousands of documents, on almost any subject imaginable.
It can also be difficult to impossible to ascertain where that information came from and whether any of it is reliable. The vast searching tools such as Wais -- Wide Area Information Services -- Gopher and WebCrawler index and retrieve data indiscriminately, without judgement as to the source.
And even if you know enough about the 'Net to be able to determine that "LOC" in the "Http://LOC" address for the document you just downloaded stands for Library of Congress, you still don't know whether your particular find originated in official government records or the science fiction section.
It is this lack of context that has led Nora Paul -- director of the Poynter Institute library, news research guru and 'Net surfer extraordinaire -- to say frequently that using the Internet for research "gives me the creeps."
One of the most important roles in new media will be putting things in context on-line, just as we put things in context now by putting the items we deem most important at the front of the paper or newscast.
Which brings us back to our database. A solid client-server system will allow us to provide information to our viewers and subscribers at the speed of light, rather than the speed of gossip.
There's an old story (told in a slightly different form by Harvard Business School's Theodore Levitt) about the people who ran the Pennsylvania Railroad. Around the turn of the century they were approached by some of the inventors working with the newfangled automobile technology with a brand new idea: the truck.
PRR officials sent the inventors away. "We're in the railroad business," said the railroad men.
So they were, and so they went out of business when the transportation business -- which they should have been in -- shifted to trucks.
Now the turn of a new century is upon us; indeed, the turn of the millennium. And it's time to decide what business we're in.
It's time to decide what age we're in. If we're going to run an information business in an Information Age, it's time to rebuild our businesses around an information system.
That means building our businesses around databases.
-- Christopher J. Feola
| 