Roadblocks on the Information Highway

By Charles E. Gardner

SECTION FIVE:
How to fix the problem

Many IM departments just don't seem to understand that the point of giving someone a powerful PC is the expectation that they will create an application to do something productive with it. Twenty years ago telephone companies gave up trying to control what kind of phones people used--they gave everyone a modular jack, a dial tone, and then concentrated on providing new services and improving their network. IM departments everywhere would be wise if they followed that example and gave their computer users an RJ-45 modular connector and an Internet dial tone, and stop trying to police what is connected. IM should focus instead on helping the end-users define what database and other applications are needed, and then supply the resources needed to develop them--themselves.

The tools needed in the Information Age of the '90s

An effective information sharing infrastructure is one designed around the functional needs of the organization as a whole, while maintaining the flexibility to meet individual needs. It is impossible to define an effective infrastructure without consulting the users and listening to what they need.

How Effective is E-Mail?

Everyone is enamored with E-Mail, but most do not realize it is just series of linked text databases with an industry standard, or "generic," access scheme. How much does E-Mail really contribute to productivity? People send E-Mail requests for information and the recipient finds the needed data on a printout or computer then types a reply. Wouldn't be simpler and faster for person asking the question to find the information in the database themselves? E-mail is a valuable communication tool, but the ability to communicate should not take precedence over having something meaningful to say: you can say it best with databases.

Shared Databases--The Key To Effective Automation

The ability to electronically share information--data you can use productively-- dynamically with other users over a LAN opens several new horizons for PC users. Finally, after 15 years of "automation" it is no longer necessary to use printed reports to get information from person to person. User A can put information into a database, and users B through Z can access it on their screens, add additional information, or just view it. Users find and print only the information they need, when they need it, eliminating the need to search through printouts to find information. If people who now receive a report have access to the database that creates it, the distribution of the report can often be eliminated entirely.

Information has traditionally been collated and organized manually before being processed and filed. These time consuming steps are eliminated because information in a database can be retrieved and sorted in any way. If a process is dynamic, such as purchasing, status reports are out of date soon after they are printed. The second significant advantage of shared databases is that information can be entered at random and used by others immediately.

Keeping information in a database and retrieving it only when needed, instead of relying on reports, ensures that decisions are based on the most current information. Having information in databases completely eliminates the need for paper files.

Adding status functions and scripted command sequences activated by a single keystroke or a mouse click makes it possible to create automated listings (e.g., outstanding vouchers, project deadlines) so managers can quickly focus on the problems that need their attention most.

Most tasks can be structured to allow management "by exception." That's just an MBA's way of saying, "If it ain't broken, don't fix it." Every decision making process is defined by parameters. By designing those parameters into a database it can automatically find the situations that are out of the norm and need correction. You can't do that without a database status tracking system.

Most people have never been exposed to either databases, or that way of thinking. Adding new technology without giving people a new perspective how it can solve their problems solving is a waste of new technology. Remember why Ross Perot sold so many IBM mainframes--he taught people how to solve their problems with technology.

Effective use of shared databases always requires changes in procedures

Spreadsheets have automated many redundant tasks but there has been little real gain in overall efficiency because information still travels between departments on pieces of paper and data is reentered redundantly. Shared databases can eliminate all of the duplicated keystrokes and a great deal of the paper passing, but their effective use nearly always require changing the way people work. That, not adding a computer, is the most difficult part task in effective automation.

Consider a typical existing ordering procedure: A purchase order is prepared (on a typewriter or with electronic form generation) and seven copies are printed. After approval, the original is sent to the vendor, and copies are sent to the receiving and voucher units. Copies are filed in the originating and approving office (nobody knows what to do the seventh copy, but they've always made one, so they file it too).

When the goods arrive, the receiving unit forwards them to the ordering office and generates a report for the voucher unit on a typewriter. When the voucher unit receives the invoice from the vendor, it retrieves the purchase order and receiving report from its file cabinet. If there are any discrepancies between the P.O., receiving report, and invoice they send them all to the ordering office for clarification. By time the paperwork is returned the prompt-payment discount period has expired and the agency has lost a $300 credit.

Consider a modified system using a shared database: The ordering office accesses the agency-wide purchasing database using a Windows PC, and the user's password allows access to the order screen. The order information is entered. The vendor's name, address, and contract information is retrieved automatically by entering a four-digit vendor code. Fund cite and other information is also imported automatically. Meanwhile, the approving officer is also accessing the database (she keeps the database open in the background on her Mac and checks the P.O. layout every hour or so). All the orders requiring her approval appear at the click of a button. She enters a code that inserts her name into the records. One copy of the P.O. is printed and mailed to the vendor. The P.O. form was changed last week when someone suggested using "window" envelopes, so no envelope needs to be typed--the change to the P.O. layout was made in a couple of minutes by the purchasing manager, who had created the system.

When the goods are received, the clerk on the loading dock enters the order number and the P.O. appears on his screen. He checks the contents against the P.O. If the goods match what was ordered he enters his ID code to acknowledge the shipment, and forwards it. If the goods do not match the order, he checks with the ordering office and rejects the shipment. He clicks a box on the P.O. record to alert all users of the problem. The voucher unit won't pay the invoice because both the ordering and voucher offices will see the receiving "flag" when they access the database. The voucher department generates payment it when the invoice is received and checked against the database. Since incorrect shipments are always rejected by receiving and also flagged, there are never any follow-ups. This ensures the payment discounts are always received and there are no mispaid invoices. Under this system a payment could actually be generated when the goods are received, eliminating the voucher process entirely. Invoices could then be thrown away or vendors could simply be told not to send them. Tom Peters, in his book "Thriving on Chaos" related how the Ford Motor Company successfully radically streamlined its purchasing by using such a procedure.

All parties involved in the purchasing process, including overview and control elements, can tap into the database at any time to monitor status and cost. The information is as current as the last entry. The ordering office can check order status and retrieve information on total amount purchased for any period. The budget officer can access the database and obtain accurate information on obligations and liquidations for the entire agency or a single element. The head of IM can even search for orders of No. 2 pencils to see who needs computer training.

Changing how people work, not adding technology, is the road to efficiency

In this example the increased responsibly undertaken by the receiving clerk--the new procedure of refusing to accept goods not matching the order--is responsible for eliminating the biggest bottleneck of the existing system; voucher processing. Change, not automation, had the greatest effect increased the efficiency. The database merely facilitated the process.

As the purchase order example illustrates, database solutions such as timekeeping, purchasing, and budgeting are most effective when implemented on an organization-wide basis. This implies the need for a central database. Economies of scale and other factors might lead to the conclusion that an expensive centralized mainframe-based relational database system is needed. But, it is important to justify, rather than jump to that conclusion. Although an organization has 5,000 computer users, only 50 may need access to the purchasing a database application at any given time, and most may use it infrequently.

The author has created such a database purchasing system for his office. It took less than a week to design it and get it running. Anyone on the LAN can use it to order and follow-up on the status. Previous orders of the same item over the past two years can be found in seconds and a summary of purchases, by vendor, can be produced for any period.

Redesigning an effective Information Technology infrastructure

Getting automation back on track

The model for automation should be an old fashioned barn raising, not the delivery of prefabricated barns by helicopter. The most effective organization structure for managing technology and using it effectively is a rectangular matrix in which computer specialists work on a task-oriented basis in the bureaus--creating new databases for end-user needs--while at the same time remaining linked on functional basis to a central IM planning process. The first step towards improving the automation effort is to involve knowledgeable end-users in the strategic planning of the information infrastructure. If this had been done before the implementation of LANs, the power-users most likely would have suggested organizing them along functional, rather than bureaucratic, lines and would have pointed out that the most important component was not the hardware, but the shared database applications running on the servers which would allow information to be used effectively. They would have told IM what hardware and software was most effective for solving the problems in their office. If they had received it, most of the problems would have been gone within a year.

Computing's MVP--The Power User

Many other writers have described the way IM departments develop applications as the "Field of Dreams--Build it and they will come" approach. They have IM's attitude towards the user correct, but analogy all wrong. If you have seen the movie you know that it's the guy out in the tall corn who is building the ball field and the guys selling the tractors who think he is crazy. The guy in the field is a power-user whose creativity is being handcuffed by IM purchasing / support restrictions. He's out in the field building applications on faith and little else, with the hope that someday IM will give him the equipment and software necessary to implement them effectively. If the goal of automation is the effective use of technology then the power-users are the most valuable players in the game. For years they have been using computers to solve problems and eliminate redundancy, and most have "reinvented" their slice of the bureaucracy several times over as more powerful hardware and software was made available to them by IM. In most cases they have fought hard for the equipment they have, and although it is not always what they asked for, they have used it to its maximum potential.

Because these power-users have learned how to eliminate redundancy in their own sections they can easily see how it can be eliminated where bureaucratic lines intersect and information is passed. But they can do little about it because they have no voice in the overall planning process, and the IM bureaucracy tolerates no interference from "amateurs." IM doesn't even know who the power-users are. They have never tried to find them, or bring them all together to formulate an effective automation strategy. By tapping the talents of the power-users IM would be harnessing the most knowledgeable segment of the user population and would find that many of them have already developed applications for their departments which could be applied on LANs organization-wide with little or no modification.

Application specialists are needed in the Bureaus

What is needed in the bureaus is not legions of hardware support specialists but people who know how to create spreadsheets and databases and solve problems with them. IM has historically failed to meet the application (i.e., problem solving) needs of the organization because it has never been is equipped, and never will be, to support the application needs of the end-user. The needs of the user have exploded exponentially because each new application that is added to an office brings to mind five more that are needed. The only possible solution to this dilemma is to teach the users to solve their own problems.

A new IM philosophy and structure is needed to meet user needs

The goal of the redesigned central information management structure should be identifying and supporting those in the organization who can develop and support their own computer applications. Getting the technology specialists off of their equipment-oriented treadmill and into the field working on applications will expose them to the problems that need to be solved and create a productive synergy between the people who know the problems and the ones who have tools to solve them. Keeping these specialist linked on a functional basis to a small central planning function in IM--with meetings, high-level training seminars and a functional performance evaluation process--would keep the skill levels of the information specialists current and the strategic planners in touch with user needs. Most importantly it will prevent the current duplication of effort between the bureaus.

The biggest shortcoming of the current IM structure is its inability to provide custom designed applications to the users. If the matrix organization model is adopted and power-users are called-up to the major leagues to the become application facilitators in the corporate IM culture, this problem will, to a great extent, be solved. The dynamic synergism between knowledgeable bureau application facilitators and a redefined and streamlined central strategic planning staff should result, in short time, in a comprehensive overview of how information moves in the organization and how its various elements interact on a functional and information sharing basis. This will reveal where organizational changes, hardware, software, and shared database applications are needed. It is vital that the technical planning group also include those outside the technological and management realms to ensure that the input and insights from all areas and levels of organization are incorporated. Increasing in efficiency will only come through change. Change is not possible without involving people on all levels.

Equipment and software standards--Handcuffs on creativity

People's technology needs, their level of technical sophistication, and the technology which is available to them all change too rapidly for a bureaucracy to keep up with. By the time an agency's standards are adjusted to accommodate newest technology it is already out of date. Twenty years ago telephone companies gave up trying to control what kind of phones people used--they gave everyone a modular jack, a dial tone, and then concentrated on providing new services and improving their network. IM departments everywhere would be wise if they followed that example and gave their computer users an RJ-45 modular connector and an Internet dial tone and stop trying to police what is connected. IM should focus instead on helping the end-users define what database and other applications are needed and then supply the resources needed for them to develop their own solutions.

Agency hardware and software standards, if they exist at all, should be defined by the needs of the most sophisticated users, not by the ability of the current IM equipment and support infrastructure to meet them. Standards should be minimums which define the support and training infrastructure, not caps that will prevent application development by the most advanced uses.

Users should be encouraged to push the limits of the technology, and "power-users" who develop shared-databases and other applications should be identified and given the resources they need to be effective.

The information managers of tomorrow should the good judgment to realize that if a power-user wants to deviate from the established IM standards:

  1. the standard hardware or software does not meet their needs;
  2. they know what they're doing, otherwise they wouldn't be asking for something different;
  3. what they are asking for is most likely something better that IM doesn't know about yet and should investigate; and
  4. they don't need IM's support for training (see #2).

The most effective system is the one that meets everyone's needs

It is amusing listen to Macintosh and Windows PC users argue over which system is best. This debate has been going on for ten years and neither side has budged an inch. The author has been saying for past ten years, "Within five years everyone would be using UNIX." It appears he finally may be right. There is a powerful new generation of desktop hardware emerging which will open new horizons for desktop computing and shared databases that can only be imagined. PC and Macintosh GUI database software will be able to run on Pentium /Power PCs, and multiple CPU enterprise servers will be become inexpensive and powerful ways to meet the information sharing needs in the 1990's. It is foolhardy for IM departments to approach this technology with a 1960s mainframe mind set.

Pentium and Power PC based workstations will be capable of running Windows, Macintosh, and UNIX but their real power will only be unleased when a new generation of software is written for them. Microsoft is backing the Intel Pentium. Apple and IBM are backing the Motorola Power PC. Which should IM support? The one that makes people most productive. Which one is that? Give them a choice and then ask them. For the past 20 years no one has asked and there has been no choice.

People know what tools they need but Management never asks them

The still prevalent practice hiring third-party contractors to conduct "need assessments" should be stopped immediately. It's more money down a rat-hole spent by bureaucrats that are too ignorant to make their own evaluations and too stupid to listen to people inside their organizations that can. The people inside the organization already understand both the problems and the available technical solutions. A large portion of IM's financial and personnel resources are dedicated to specifying and administering contracts, but large-scale contract solutions are seldom as effective as those developed internally by the users. The need for contract solutions could largely be eliminated by formally integrating the power-users already in the bureaus into the IM management matrix as application specialists. Most are already performing these duties on an ad hoc basis in their work groups, so all that is necessary is giving them a role in the strategic planning process to ensure they receive the resources they need.

Shared databases are the ultimate destination of the information highway. This makes the choice of a database development tools and who will develop the applications the main issue to be resolved on an organization-wide basis. Small, modular, solutions should be examined before making commitments to large, expensive ones. The functional inter-relationships between departments should define the overall architecture of the information system and should be the deciding factor when configuring LANs and developing databases.

Involving users in the design of new systems

The roadblocks on the information highway are the guys in front holding up traffic, but people resistant to change are the pot-holes. Even in cases where an application has been custom designed by someone in the same office with user input, the process of implementation often meets considerable resistance because users are nearly always reluctant to disrupt their familiar routines, no matter how inefficient they are. The most frequent answer to the question,"Why do you do it that way?" is, "Because we've always done it that way."

Seventeen years designing and implementing database-based systems at the user level has taught the author that the most effective way to implement change is to make people's job easier, not just more efficient. If that cannot be done, there must some other reward built-in the the system if it is to work. At each step of the process the, "what's in it for me?" hurdle must be jumped. For the purchasing order example to really work effectively the receiving clerk must be given a new title and a promotion to reflect his new responsibility and value to the organization; otherwise he will be unwilling to accept it. Most importantly, the people affected by the change must have a voice in the design process. They always have useful insights regarding new ways of doing things, and they will more readily support what they help create.

A wise automation developer is the one who makes the users think the new system was all their idea, and he just supplied the computerized tools to help them implement it.

[- table of contents -] [ mail to author]