There is a better way to reengineer business processes than that now in progress at most big organizations. It produces results in months, rather than years. It links process reforms tightly to competitive strategy, and thus delivers precisely the changes likely to provide the greatest benefit to the company as a whole. It fosters continuous rapid change among frontline employees. In short, it works.

Three simple steps underlie the new approach (Exhibit 1). First, diagnose the most critical problems and opportunities facing the company and sketch out a possible solution. Second, quickly, over a few months, translate the sketch of a solution into new work processes and systems; include new ways of working as well as new computer systems, and begin using both for real. If, for example, the problem at hand is inventory management, use the new process to manage a few products at several plants and warehouses. Given this real-world experience, determine where these new processes and systems succeed or fail, and quickly fix the failures. In other words: do it, then fix it. Third, scale up systems for rollout across the whole organization. Communicate the proven success of the trials in order to build momentum for change.

The essence of the "do it, fix it" approach is rapid learning from doing. Speed ensures that change is always relevant, it forces tradeoffs so that limited resources are devoted to pursuing goals of real value, it allows top people to participate in change, and it builds unstoppable momentum. Trying out new ideas in the real world allows their shortcomings to be rooted out by the harshest of tests—real-world experience—and their successes to be proven beyond challenge from the most cynical critics. By stressing speed of change and using the real world as a laboratory to learn from, the do it, fix it approach makes change and improvement a constant fact of corporate life.

Simple though it is, the do it, fix it approach can be uncomfortable to adopt because it requires a fundamentally new attitude to business problem solving and risk. But it is more than worth the effort. For the new approach overcomes the chief drawbacks of conventional reengineering (Exhibit 2).

Perhaps its greatest advantage, however, is the tight and natural linkages it creates between strategy and process change. At each step along the road, the do it, fix it approach asks executives to determine which changes will make the most difference to their company as a whole—not just to individual departments. Then it pushes responsibility for deciding how to grasp these opportunities on to the shoulders of the executives who will carry out the plans they create. The result is that strategy is constantly honed by the experiences of line managers, and vice versa. The corporation learns as it improves itself, and it improves itself by learning.

Who dares wins

Do it, fix it is intimidating. It demands real executive leadership and extremely rapid, effective systems development

So, if do it, fix it is so simple and so effective, why don’t more companies work this way? For three reasons: the do it, fix it approach is intimidating, it demands real executive leadership, and it usually requires extremely rapid and effective systems development.

While it is empowering to realize that your ideas will be acted upon, it is also deeply scary

This approach intimidates because it puts ideas into practice—with real profits at stake—breathtakingly quickly. It is one thing to walk into a conference room to brainstorm new approaches to, say, order fulfillment and customer response secure in the knowledge that several months of analysis and planning will follow to check and countercheck your ideas. It is entirely another to walk into that room knowing that in a few weeks you will be training salespeople to implement your ideas. While it is empowering to realize that your ideas will be acted upon, it is also deeply scary.

To make change less scary, most current big reengineering projects go to great lengths to lay out in painstaking detail the process as performed today. Then they analyze and plan out in even more minute detail how it should be done tomorrow, and all the steps in the implementation of the change from one to the other. Unfortunately, these sensible-sounding measures that are undertaken to reduce risk very often ensure failure—or at best a mediocre solution, slowly implemented. Particularly when computer systems are involved, it is not uncommon for new processes to take three, four, or more years from conception to first deployment. During that time, changes in the market or the company can render the new approach irrelevant. While managers analyze and plan, opportunities slip away.

Although it seems riskier, do it, fix it in fact increases the chances of success even as it reduces the cost of failure

Although it seems riskier at first glance, the do it, fix it approach actually increases the chances of success even as it reduces the cost of failure. True, it rushes new processes into pilots before they are completely debugged. But building and trying pilot operations for real requires only a few person-months of effort and eliminates the need for much analysis and planning, which can absorb person-years of effort before implementation even begins. In practice, the do it, fix it approach involves little risk, because the new systems are run in parallel with the old until they have fully proved themselves. Complete switch-over takes place only after executives are sure that the new work processes and systems will work better than the old.

Because they take months rather than years, do it, fix it projects can—and must—be led by experienced, senior frontline staff, people used to committing the company’s resources day by day. Their leadership and experience creates change that moves faster than markets. Their knowledge allows computer systems and business processes to be developed in parallel, enabling unexpected insights gained from experience with one to be used to improve the other. Once they have devised a successful pilot, there is nobody better qualified or more able than these line executives to persuade their colleagues to adopt the new ways of working—or, better still, to improve upon them.

The remaining crucial element in the do it, fix it approach is systems development. Only a handful of business processes do not require some sort of information technology support, and the handful is growing smaller with each passing year. Until recently, though, there was simply no way to build a quick, cheap working version of a new computer system in order to try out a new idea.

Today, however, personal computers and client-server technology can create sturdy systems sophisticated enough to do real work, and achieve this in months, not years. Better, these systems can often be improved and extended to form the basis of actual production systems—not just thrown away after they have proved their point. Not only does such quick construction enable rapid change, it also creates a more sustainable rhythm of change than do conventional approaches. Instead of a dramatic, exhausting upheaval every few years as whole systems are completely transformed, the combination of do it, fix it and this new way of developing IT creates continuous, rapid improvement in which corporate processes, executive knowledge, leadership, and information technology all improve together, each fuelled by improvements in the others.

One senior executive who led a successful reengineering effort using the do it, fix it approach described its advantages as follows:

The beauty of the approach was that we could never have predicted how messy the real world would be, and how many times we would have to change what we were doing to fix problems or chase new opportunities. But we were moving so fast that by the time we spotted new challenges, we were already reaping measurable rewards from our previous efforts. And that’s what gave us the courage to go on. We had put the reward back into the risk/reward equation.

Here’s how

The first and quickest step toward change is diagnosis. In one or two months, gather the facts top management need to set their priorities for change. Then, take their list of opportunities and set specific, measurable, and ambitious objectives for the solution. Aim for the best you could possibly do, even if that seems too good to be true (after all, the whole point of piloting is to discover whether something can be done or not). Sit down with a small group and sketch out a solution to the problem so that it can then be tested more broadly.

Typically, the most important factor in deciding how to approach change is computer systems. Changes in computer systems fall into three broad categories:¹

Supplemental systems plug in new functionality as needed, leaving untouched everything already done adequately

Supplemental systems. If the process at hand is already largely automated, there are obvious advantages to reusing the existing code and hardware as far as possible. Client-server-style approaches to systems improvement can be used to do just that. The idea is to plug in to the old system just as much new functionality as is needed, while leaving untouched everything that is already done adequately. Two sorts of enhancement have proved particularly useful: those that create new user interfaces to old systems, and those that make old systems smarter. The first approach, front-ending, can, for example, pull together a collection of customer-oriented databases to enable sales staff quickly and easily to gain an overview of the whole of their firm’s past relationship with a customer—providing obvious opportunities for improving sales and customer service. The second approach, wraparound, might transform, say, the logic that calculates product pricing, while leaving untouched the surrounding systems handling ordering and distribution. Both approaches typically leave 80 to 90 percent of the old system untouched—which means they can build new capabilities much more quickly and cheaply than starting again from scratch.

Standard systems. If a supplemental system is not cost-effective, if the process at hand confers little strategic advantage, and if other companies do more or less the same thing in more or less the same way, then the best approach is to try to take advantage of the process’s very ordinariness by using packaged software. Excellent software packages already exist for a variety of processes, including order fulfillment, raw materials procure-ment, and production planning. By ruthlessly changing business processes to fit the software—rather than reengineering’s traditional vice versa—companies can get state-of-the-art processes at minimum cost. And they can use the effort they save to be more distinctive in areas where distinctiveness really matters.

New custom systems. In a small minority of cases, new systems still have to be custom-built from scratch. Either there is no existing automation—which is very rare in today’s companies—or the old system is simply so decrepit that it can no longer be improved cost-effectively. But even when you are starting from scratch, the lessons of do it, fix it still apply. Build a prototype; put it to work; fix it; build on it; scale it up; and roll it out.

The heart of the reengineering process lies in building unstoppable momentum for change

Having diagnosed the problem and chosen the relevant technological approach, a company’s next steps are obvious: do it and fix it. If you don’t have a pilot solution that is changing the way people work within six months, you’re probably not being ambitious enough. Moreover, the momentum of change should build on itself. Within months, the first to be trained in using the new process should themselves be working on improvements, and training others. Change should cascade through the company, accelerating as it goes. The heart of the reengineering process lies in building unstoppable momentum for this change—and that relies at least as much on communication and vision as on process design. Examples best show the way.

Grasp the moment

For one large retailer with geographically dispersed operations, the challenge of change was redoubled by the fact that the company was doing very well. But managers knew that to stay abreast of the best, they would have to provide better customer service, with lower levels of inventory. They drew together a team led by a senior line executive with buyers, systems support people, a financial analyst, and a transportation manager. They set the team a ten-week deadline to convince top management that change was worth pursuing. In fact, the team managed that in eight weeks, and spent the remaining two in brainstorming the design of a totally new logistics process.

With only a sketch of a solution to guide them, the team selected a region and product category for a pilot of the process. Half of the team flew out to the pilot site, and devoted themselves to communicating and testing their vision with the line managers with whom they would have to work on the pilot. The other half began creating a pilot of the new system, first using Excel spreadsheets and Access databases. Just 68 days after the end of the diagnostic period, the pilot—programmed in Visual Basic and running on a PC networked into the company’s major systems—went live and real goods started shipping from vendors to warehouses and on to stores under the management of the new system.

The next two months were spent in frantic improvement and communication. Several improvements were made to the computer systems and to the business process itself. When the team encountered an unusual product or store that did not fit neatly into the logic of the existing system, changes to accommodate important exceptions were usually made within 24 hours. Even as they improved the new process, the change team also proselytized its virtues to the executives who were next in line to adopt it.

By the end of three months, the team had proved that the new process did indeed achieve higher service levels (up over 10 percent on average) without increasing inventory. They had also gained a consensus among both top management and colleagues that the new process was a great improvement over the old practices. Today, the firm has taken advantage of that consensus to drive the rollout of the new process. Members of the original team fanned out across the company to spread the gospel of both the new process and the new way of managing change. Then members of the second-generation teams also spread across the company—and so on, with ever-accelerating momentum. Even as it implements the process, the firm improves it continuously.

Making standards work

The case of a major electronics manufacturer provides an excellent illustration of the advantages of the standard system approach. The diagnosis in this case was short and straightforward: the company was suffering from its own complexity. Growth had left it serving a wide variety of different customers, ranging from those buying products off the shelf to those demanding elaborate and expensive custom-manufactured solutions. After-sales service arrangements ranged from sending parts in the mail to dispatching customer service engineers anywhere in the country within a guaranteed four-hour period. To add further complexity, the firm merged with another company and thus had to cope with multiple systems for order entry and for designing products to meet customers’ unique needs.

Enough, executives soon decided, was enough. Most of the firm’s competitive advantage came from its people and its product technology, not from service and delivery. So executives decided to devote their efforts to unifying and simplifying their business processes and systems. They chose the best off-the-shelf software package they could find. Although they initially left open the possibility of customizing the software, they demanded such high financial returns to counter the high costs of customization that, in the end, no customization was undertaken.

Executives scrapped eighty overlapping (and often conflicting) systems and replaced them with two packages

Executives scrapped 80 overlapping (and often conflicting) systems and replaced the lot with two software packages. All order-entry and fulfillment work was ruthlessly standardized to fit the capabilities of one of these packages. Product design tools were similarly replaced with an object-oriented package, also purchased off the shelf.

Determined though they were to fit business processes to the capabilities of the packaged software, executives still had plenty of room left for process innovation. Product design, for example, was moved from the beginning of the order-delivery process to the very last moment so that customer changes could be accommodated at the lowest possible cost.

By completion, inventories had been cut; logistical costs were reduced, and the time required to order and install a new system had fallen to a fraction

Use of the new processes began only five months after the onset of reengineering, and rollout was complete far earlier than originally expected. Pilots were organized to fit specific segments of the overall business process: installation, order management, inventory management, and after-sales service. Only after all proved successful did the new processes replace the old ones. By completion, inventories had been cut by over 45 percent, logistical costs by over 35 percent, and the time taken to order and install a new system from 60 to 14 days. Not surprisingly, customer satisfaction had reached recordbreaking heights, and so had the job satisfaction of the people who had undergone the gut-wrenching experience of changing just about everything about work as they knew it.

Viewed with hindsight, the barriers to change were high. The old way of doing things felt comfortable, which was hardly surprising, as it had been devised to support long-established work practices and procedures. But the speed and decisiveness of change galvanized the workforce nonetheless. At first, in early brainstorming meetings, managers were reluctant to pursue so radical a course as to scrap many of the systems that they themselves had built. But a few crucial calculations of the business value of the proposed changes overcame their doubts, and, once they were committed, the sheer speed with which the solution was tried, fixed, and steadily improved was both liberating and invigorating. Having solved pressing problems, the company has gained the nerve to tackle other opportunities with the same decisiveness.

Managing the whirlwind

As they look back on their first experiences with the do it, fix it approach, the overwhelming impression of even the most successful firms is a feeling of amazement at just how different the change felt from the usual pace of corporate life. It was at once exhilarating, empowering, and somewhat terrifying—as if someone had suddenly tipped the contents of the committee room into a Silicon Valley entrepreneur’s garage. Seldom before had people worked so hard, or achieved so much. With hindsight, several broad lessons stand out:

Commit to goals, not plans. To force managers to take responsibility for solving their own problems, analysis needs to stop, and the real work of change begin, as soon as the conclusions are agreed upon. This tends to be long before detailed plans have been worked out—which is precisely the point. Starting to change people’s work lives before plans for the change are complete is the only way to drive home the point that corporate change is never finished—and that the only person who can fill the gaps in your job description is you yourself. For three to six months, this creates more anxiety and uncertainty than many executives have ever experienced before. It also generates more real achievement and self-confidence than most have ever experienced.

Pursue ambitious targets. Set over-ambitious goals. Really stretch yourself and your organization. Most of the time, you will be amazed as you reach seemingly unattainable goals and go on to achieve more ambitious goals yet. (In fact, the regret most commonly expressed by leaders of reengineering efforts is that "We didn’t reach far enough.") You will also reinforce the point that success in day-to-day business is largely a matter of reacting intelligently to things that don’t turn out exactly as planned. But whatever the case, it is important to manage the risks of failure. In particular, this means setting the highest goals for the pilot, but at the same time limiting the breadth of the project and the resources devoted to it initially so that even total failure is not cripplingly expensive.

Learn from doing, especially when it’s painful. Pushing line managers to begin using a new process before all the details have been worked out is painful. It is also the best, indeed only, way to involve line managers in creating processes that truly meet their needs—instead of what analysts perceive their needs to be. One retailer’s deadlines pushed a team to set up a new product line with a process barely sketched out. The experience proved a real endurance test: it started with a 100-hour week of confusion and frustration. But from this experience were born a system and a user interface that enabled those who followed in the team’s footsteps to set up a new product line in about 30 minutes—a result even better than the original ambitious goal for the project.

Make line managers do the work; use staff people only to help. Line managers already know most of what they need to do. They can move more quickly in implementing solutions than any staff analyst, who will always take longer to understand an opportunity. Line managers have the confidence to adopt radical solutions, and the experience and judgment to recognize what works and what doesn’t. More important, when a solution does work, they have the credibility among their peers to drive a rapid rollout campaign.

Start communicating well before you know all the answers. Since many people will ultimately be caught up in the change created by new processes, it is crucial that all factions be involved in discussions about problems and opportunities as early as possible. Executives might feel rather feeble standing in front of a group and answering even simple questions with "We don’t know yet." But the willingness to acquaint people with the issues—and to demonstrate that even the people at the top don’t know all the answers—is crucial to building credibility and kick-starting participation in change. In practice, once the project gets rolling, communication can easily absorb a third of the time of the leadership team.

So just do it, fix it, and do it again

Do it, fix it makes change a continuous challenge for everybody rather than a periodic obsession for a few planners

The do it, fix it approach pushes every company to behave as only the highest-performing companies do today: to overturn traditions as a matter of course, to set and achieve aggressive objectives, to move quickly, to take managed risks, and to learn from successes and failures alike. It does so by transforming challenging goals and timetables into an ally of change rather than an obstacle, by ruthlessly pushing aside time-consuming details of little worth to focus on big opportunities, and by making change a constant challenge for everybody, rather than a periodic obsession for a few planners. Reengineering as commonly practised today just doesn’t do these things. That is why reengineering must itself be reengineered.