how process enterprises really work - caliw.cali.org/conference/2000/program book/hbr2.pdf ·...

How Process Enterprises Really Work

by Michael Hammer and Steven Stanton

Reprint 99607

lt h o u g h r e e n g i n e e r i n g h a s i n s o m ecircles become a euphemism for mindless

downsizing, it has in fact done a world of good. It hasenabled companies to operate faster and more effi-ciently and to use information technology moreproductively. It has improved the jobs of employ-ees, giving them more authority and a clearer viewof how their work fits into the operations of the en-terprise as a whole. It has rewarded customers withhigher-quality products and more responsive ser-vice. And it has paid big dividends to shareholders,reducing companies’ costs, increasing their rev-enues, and boosting their stock values.

Most of all, though, reengineering has changedthe perspective of business leaders. No longer doexecutives see their organizations as sets of discreteunits with well-defined boundaries. Instead, theysee them as flexible groupings of intertwined work

108 Copyright © 1999 by the President and Fellows of Harvard College. All rights reserved.

What do IBM, Texas Instruments, Owens Corning, and

Duke Power have in common? They’re all redesigning

their organizations around their core processes – and

reaping enormous benefits as a result.

and information flows that cut horizontally acrossthe business, ending at points of contact with cus-tomers. Reengineering, in other words, has allowedexecutives to see through the surface structure oftheir organizations to the underlying purpose: thedelivery of value to customers in a way that createsprofits for shareholders.

But this new process view of organizations hasnot yet been fully realized. Many companies haveintegrated their core processes, combining relatedactivities and cutting out ones that don’t add value,but only a few have fundamentally changed theway they manage their organizations. The power inmost companies still resides in vertical units –sometimes focused on regions, sometimes on prod-ucts, sometimes on functions – and those fiefdomsstill jealously guard their turf, their people, andtheir resources. The combination of integrated pro-

by Michael Hammer and Steven Stanton

How ProcessEnterprisesReally Work

A

harvard business review November–December 1999 109

cesses and fragmented organizations has created aform of cognitive dissonance in many businesses:the horizontal processes pull people in one direc-tion; the traditional vertical management systemspull them in another. Confusion and conflict ensue,undermining performance.

That’s not the way it has to be. In recent years,we’ve seen a number of companies make the leapfrom process redesign to process management.They have appointed some of their best managersto be process owners, and they have given them realauthority over work and budgets. They have shiftedthe focus of their measurement systems from unitgoals to process goals, and they have based compen-sation and advancement directly on process perfor-mance. They have changed the way they assign andtrain employees, emphasizing whole processesrather than narrow tasks. And they have made sub-

tle but fundamental changes to their cultures, stress-ing teamwork and customers over turf and hier-archy. They have emerged from all those changes astrue process enterprises – companies whose man-agement structures are in harmony, rather than atwar, with their core processes –and they have reapedenormous benefits as a result.

Creating a Process EnterpriseTexas Instruments’ calculator business is one suchprocess enterprise. In the early 1990s, the once-thriving unit was in trouble. Plagued by long cycletimes in new product development, it was losingsales to more nimble competitors. Managementsaw the problem and took action, redesigning theproduct development process from scratch. Newcalculators would now be developed by teams of

AR

TW

OR

K B

Y C

HR

ISTO

PH

ER

ZA

CH

AR

OW

people drawn from engineering, marketing, andother departments who would work together in thesame location. Each team would have full respon-sibility for its product from conception through

launch, including suchhighly specialized activitiesas producing documenta-tion, creating advertising,and even developing train-ing materials for teacherssuggesting ways to inte-grate calculator use intotheir classes. Because eachteam would control everyaspect of its process, all de-velopment activities would

be performed in a coherent, streamlined fashion,free of all the old bottlenecks and delays.

That was the theory. But it didn’t work out thatway. The first pilot teams not only failed to achievethe desired reductions in development times, theybarely managed to operate at all. They were, in ef-fect, sabotaged by the existing organization, whichviewed them as interlopers. Functional departmentswere unwilling to cede people, space, or responsibil-ity to the teams. The technical writers and designerscharged with creating documentation got instruc-tions from the product team and then got conflict-ing orders from their supervisors in the marketingdepartment. The corporate training unit refused torelinquish control over the development of trainingmaterials, and the advertising department insistedon continuing to create product advertising. An ef-fort that had been intended to create harmony inproduct development instead created discord.

The problem was not in the design of the process.The problem was that power continued to lie in theold functional departments. The business’s leaderssoon realized that it was impossible to superimposean integrated process on a fragmented organization.

Rather than give up on the process, they changedthe organization. The development teams becamethe primary organizational units. The mission of thefunctional departments was redefined; no longer responsible for the work, they focused on trainingpeople in the skills required by the teams. A new

management role – the process owner – was createdto oversee product development in the calculatorunit. Budgeting was done by process instead of bydepartment. Office space was reconfigured to betteraccommodate and support the process teams. Theunit’s senior managers took every opportunity tounderscore the importance of a process perspectivethrough formal presentations, writings, and infor-mal conversations.

As a result of the changes, the calculator unit hasbecome much more successful in introducing newproducts. The time it takes to launch new productshas dropped by as much as 50%, break-even pointshave been reduced by 80%, and the unit has be-come the market leader in product categorieswhere it previously had no share whatsoever. Theoverall return on investment in product develop-ment has more than quadrupled.

IBM went through a similar transformation a fewyears later. Seeing that its large corporate customerswere increasingly operating on a global basis, IBMknew it would have to standardize its operationsworldwide. It would have to institute a set of com-mon processes for order fulfillment, product devel-opment, and so forth to take the place of the diverseprocesses that were then being used in differentparts of the world and in different product groups.But the change effort immediately ran into an orga-nizational roadblock. IBM’s existing managementsystems concentrated power in the hands of coun-try and product managers, and they were reluctantto sacrifice their own idiosyncratic ways of work-ing. They simply refused to allocate the human andtechnical resources required to design and roll outstandardized processes.

In response, IBM changed its management struc-ture. Each process was assigned to a member of itssenior-most executive body, the Corporate Execu-tive Committee, making that member accountablefor the process. All members were required to re-port back regularly to the Executive Committee onthe status of the design, deployment, and imple-mentation of the processes, including the benefitsrealized. Each process was then assigned an owner,called a “business process executive” (BPE), whowas given responsibility for designing and deployingthe process, as well as control over all expendituresfor supporting technology.

Each of IBM’s far-flung business units is now ex-pected to follow the processes designed by the BPEs.Should there be a disagreement between a unitmanager and a process executive about the work-ings of a process, the two are expected to resolve ittogether. By shifting organizational power awayfrom units and toward processes, IBM has achieved

110 harvard business review November–December 1999

Michael Hammer is the founder of Hammer and Com-

pany, a management education firm based in Cambridge,

Massachusetts. He is the author of “Reengineering Work:

Don’t Automate, Obliterate” (HBR July–August 1990).

Steven Stanton is a managing director at Hammer and

Company, where he specializes in process management.

In many businesses,horizontal processespull people in onedirection; traditionalvertical managementsystems pull them inanother. Confusion and conflict ensue.

How Process Enterpr ises Real ly Work

its goal of standardizing its processes around theworld. The benefits have been dramatic: a 75% re-duction in the average time to market for new prod-ucts, a sharp upswing in on-time deliveries and cus-tomer satisfaction, and cost savings in excess of $9billion.

In 1997, Owens Corning found that its efforts toinstall an enterprise resource planning system werefloundering. An ERP system is, in essence, an inte-grative mechanism, connecting diverse departmentsthrough a shared database and compatible softwaremodules. It is impossible to get the full benefits of an ERP system without having integrated processes.But at Owens Corning, as at IBM and Texas Instru-ments, there was no one in the organization to speakfor processes. Departmental and regional managers,as a result, were either rejecting the new software orseeking to tailor it to the narrow needs of their par-ticular units. In response, the company’s top execu-tives reorganized people into companywide, cross-functional process teams and appointed processowners to lead them. The new organization providedthe impetus for a successful ERP implementation,which has in turn led to a 50% increase in inventoryturns, a 20% reduction in administrative costs, andmillions of dollars in logistics savings.

Creating a process enterprise is an enormouslycomplex undertaking, as Texas Instruments, IBM,and Owens Corning all found out. Traditional orga-nizational units are naturally hostile to integratedprocesses, seeing them as threats to their power. Soorganizational and management structures have to be changed in fundamental ways. That doesn’tmean, though, that existing vertical units such asfunctional, regional, or product groups are simplydisbanded – in even the most process-focused busi-ness, vertical units continue to play essential roles.Rather, it means that horizontal and vertical man-agement structures have to coexist, not just in peacebut in partnership. Not only does a company haveto redistribute management responsibility, it has tochange its basic management systems, and even itsculture, to support a new balance of power.

The Role of the Process OwnerThe most visible difference between a process enter-prise and a traditional organization is the existenceof process owners. Senior managers with end-to-endresponsibility for individual processes, process own-ers are the living embodiment of a company’s com-mitment to its processes. To succeed, a processowner must have real responsibility for and author-ity over designing the process, measuring its perfor-mance, and training the frontline workers who per-

form it. A process owner cannot serve just as an in-terim project manager, active only while a new pro-cess design is being developed and put in place. Pro-cess ownership has to be a permanent role, for tworeasons. First, process designs need to evolve asbusiness conditions change, and process ownersneed to guide that evolution. Second, in the absenceof strong process owners, the old organizationalstructures will soon reassert themselves.

The advent of process owners is a dramaticchange for most organizations because it separatesthe control over work from the management of thepeople who perform the work. Traditionally, a geo-graphical or functional manager oversees both thework and the people who do it. In a process enter-prise, the process owner has responsibility for thedesign of the process, but the various people whoperform the process still report to the unit heads.That kind of split in authority may be hard formany executives to imagine, but there are compa-nies that are making it work today.

One example is Duke Power, a true pioneer of theprocess enterprise. The electric utility arm of DukeEnergy, Duke Power serves nearly 2 million cus-tomers in North and South Carolina. In 1995, withderegulation looming, it realized that it had to do a much better job of customer service if it was to survive the onslaughtof competition. But theexisting organizationalstructure of CustomerOperations, the businessunit responsible for deliv-ering electricity to cus-tomers, was getting in theway of service enhance-ments. The unit was di-vided into four regionalprofit centers, and the re-gional vice presidents,overwhelmed by an endless stream of adminis-trative duties, had little time for wrestling withthe details of service provision. And even if theyhad, there was no way to coordinate their effortsacross the regions. No one, in short, was responsi-ble for how the company was delivering value tocustomers.

To solve the problem, Duke Power identified fivecore processes that together encompassed the es-sential work that Customer Operations performedfor customers: Develop Market Strategies, Acquireand Maintain Customers, Provide Reliability andIntegrity, Deliver Products and Services, and Calcu-late and Collect Revenues. Each process was as-signed an owner, and the five process owners, like


Texas Instruments’business leaders soon realized that it was impossible to superimpose an integrated process on a fragmented organization.


the four existing regional vice presi-dents, reported directly to the head ofCustomer Operations.

In the new structure, the regional vicepresidents continue to manage theirown workforces –the process units haveonly small staffs – but the process own-ers have been given vast authority overhow the company operates. First, theyare responsible for designing their re-spective processes. They define howwork will proceed at every step, and theregions are expected to follow those de-signs. Second, and just as important, theprocess owners are responsible for set-ting performance targets, establishingbudgets, and distributing those budgetsamong the regions. In other words,while the regions continue to have au-thority over people, they are evaluatedon the basis of how well they meet thetargets set by the process owners, andtheir budgets are in large part roll-ups of the monies disbursed by the processowners. The regional vice presidentshave no choice but to work in partner-ship with the process owners.

The new structure has proven to be a great success, focusing the entire or-ganization much more directly on thecustomer. Virtually every activity in-volved in serving customers has beenredesigned from the ground up. For ex-ample, the process owner for DeliverProducts and Services, Rob Manning,has worked with the regional units,with suppliers, and with his own ten-person staff to devise a new way to or-ganize warehouse facilities. Parts thatwill be required by installation crews,for example, are laid out the night be-fore for easy pickup in the morning, sothat the crews can load their trucksand be on the road in 10 minutes, a fraction of the70 minutes it used to require. The crews can domore installations in a day, so customers don’t haveto wait as long to get service.

Manning has also revamped the way the companyworks with its building-contractor customers. Asrecently as late 1996, Duke Power was meetingonly 30% to 50% of its commitments to those cus-tomers – laying cables by a certain date, for exam-ple. That created difficulties, as those customersbased their construction schedules around DukePower’s promised dates. The problem was that the

people making the commitments did not have anaccurate picture of the availability of individualfield-workers. They could not ensure, therefore,that the required skills would be in the right placeat the appointed date. Manning and his team de-ployed a new scheduling system that providesmuch more detailed information about the avail-ability of field personnel, enabling more specificand accurate assignments. They also designatedpeople to negotiate commitment dates with con-tractors and keep them apprised of changes. Finally,they underscored the importance of meeting com-



To succeed, a process owner must have real authority over designing the process, measuring its performance, and training the frontline workers whoperform it. A process owner cannot serve just as an interim project manager.

mitments to customers by measuring the percent-age of deadlines met and by publicizing each re-gion’s results on a daily basis. Duke Power nowmeets 98% of its construction commitments.

A New Style of ManagementDuke Power has learned that becoming a processenterprise is more than a matter of establishingnew management posts and rejiggering responsibil-ities. As lines of authority become less clear-cut,the way managers interact with one another andwith workers also has to change. Style is as impor-tant as structure. Process owners, for example,can’t simply order process workers to do their bid-ding. They have to work through the unit heads –the regional VPs, in Duke Power’s case. Manningsays that his role requires “three critical skills: in-fluence, influence, and influence.” Unit heads, fortheir part, have to negotiate with the process own-ers to ensure that the process designs are sound, theprocess goals reasonable, and the resource alloca-tions fair. The split in authority, in other words,makes cooperation unavoidable. If you don’t worktogether, you fail.

Duke Power’s managers, like those of most com-panies, were not accustomed to such a collabora-tive style. At first, the process owners and regionalVPs tended to act more as rivals than as partners.The problem wasn’t resolved until all the managerssat down together and developed a document theycalled the “decision rights matrix.” The matrixspecified the roles the different managers wouldplay for each of the major decisions made in the or-ganization, such as changing a process design, hir-ing people, setting a budget, and so on. It detailed,for example, which managers would actually makethe decision, which had to be consulted before-hand, and which had to be informed afterward. Ineffect, the matrix was the organization’s road mapfor managerial teamwork. Today, the managersrarely have to consult the matrix –they’ve internal-ized it. But the specificity and clarity of the matrixgave the managers a concrete sense of how the neworganization would work, and the very process ofcreating it gave them an appreciation for the new,more collaborative style of management.

The five process owners also had to learn to col-laborate closely with one another. Processes, afterall, aren’t islands onto themselves. They overlap,since the same workers are often involved in severalprocesses, sometimes simultaneously. At DukePower, for example, the same group of field person-nel installs lines (part of Deliver Products and Ser-vices) and maintains them (part of Provide Reliabil-

ity and Integrity). Initially, that overlap created aconflict. Installations almost always had hard dead-lines, reflecting customers’ need for precise com-mitment dates, but maintenance jobs often did not.As a result, maintenance kept getting pushed to theback burner. The two process owners got togetherto work out a new arrangement: certain field per-sonnel would be dedicated to each process, and therest would form a floating pool available to work oneither process. TheProvide Reliabilityand Integrity pro-cess owner alsoagreed to scheduleroutine mainte-nance in the springand fall wheneverpossible, creating greater installation capacity dur-ing the summer, when demand was highest. In addition to meeting informally to solve particularprocess conflicts, the five process owners meet reg-ularly in formal sessions with their boss, the headof Customer Operations, to review and coordinateoperational plans, budgets, performance measures,and the like.

If a company is going to make itself over into aprocess enterprise, it needs to change not only theway its managers interact with one another but alsothe way they relate to frontline workers. Processteams composed of individuals who have broad pro-cess knowledge and who are measured on processperformance have little need – or room – for tradi-tional supervisors. The teams themselves take overmost of the managerial responsibilities usually heldby supervisors. Supervisors, in turn, become morelike coaches, teaching the workers how to performthe process, assessing their skills, overseeing theirdevelopment, and providing assistance when re-quested. At Duke Power, in fact, the once ubiqui-tous foreman position has disappeared entirely, re-placed by a new role –the process coordinator.

Because the coordinator coaches rather than con-trols the people who perform the process, Duke’straditional ten-to-one span of supervisory controlhas widened dramatically; the typical process coor-dinator supports 30 to 40 people. (In some compa-nies, the number is as high as 70.) There are alsonow far fewer managerial levels at Duke; instead ofsix levels between the front line and the regionalvice president, there are only three. And as the pro-cess owners have taken over some of the former re-sponsibilities of the regional vice presidents, theVPs, too, have become more focused on trainingand developing their people. One Duke Power exec-utive calls them “super coaches.”


Process owners can’t simplyorder process workers to dotheir bidding. The role requiresthree critical skills: influence,influence, and influence.


The process owners also play an important, if in-direct, role in managing frontline workers. They actnot as coaches but as, to use Manning’s word,“evangelists,” promoting the process designs andrepresenting the interests of customers. As Man-ning puts it, “My job as a process owner is to con-vince the people who operate within my processthat there is no greater calling for them than to dowhat the customer needs them to do and that thebest tool they have is the process we have giventhem.” Manning performs this role by designingand delivering training programs to process work-ers; by setting performance targets; and by regularlytalking with them, keeping them informed ofchanging customer needs and listening to their con-cerns and ideas.

Traditional styles of management, tosum up, have no place in a process en-terprise. Managers can’t command andcontrol; they have to negotiate and col-laborate. They can’t wield authority;they have to exert influence. Any com-pany hoping to turn itself into a pro-cess enterprise needs to understand thechanges in managerial style that willbe required and their implications forstaffing and training. Few managerswill be able to make the transition eas-ily, and some may not be able to makeit at all.

The Question of ProcessStandardizationCompanies made up of many differentbusiness units will face an importantstrategic question as they make theshift to a process enterprise: Should allunits do things the same way, or shouldthey be allowed to tailor their pro-cesses to their own needs? In a processenterprise, the key structural issue isno longer centralization versus decen-tralization – it’s process standardiza-tion versus process diversity. There’sno one right answer. IBM, Duke Power,and Progressive Insurance, for example,have opted for standardization. Theydesignate a single owner for each pro-cess, and that person develops and in-stalls the same process design through-out the company. American Standard,in contrast, has different process own-ers and process designs in each of itsmajor business units.

Process standardization offers many benefits.First, it lowers overhead costs, since the process re-quires only one owner with one staff, only one setof documentation and training materials, and onlyone information system. Second, a company withstandardized processes presents one face to its sup-pliers and customers, reducing transaction costsboth for them and for itself. By standardizing itsprocurement process across all its business units,IBM has been able to create a single list of approvedvendors, enabling the company to aggregate its pur-chases and giving it much more leverage over sup-pliers. Owens Corning has standardized its orderfulfillment process across all its divisions, whichshare many of the same customers. That’s great for



Traditional styles of management have no place in a process enterprise.Managers can’t command and control; they have to negotiate and collaborate.

customers – they only have to submit one order, re-ceive one invoice, and pay one bill. It’s also great forOwens Corning, which has saved millions in logis-tics costs by consolidating shipments from differ-ent divisions.

Third, and perhaps counterintuitively, processstandardization can increase organizational flexi-bility. When all business units are performing a pro-cess the same way, a company can easily reassignpeople from one unit to another to respond to shiftsin demand. Its organizational structure becomesmuch more plastic.

As compelling as the arguments for standardiza-tion are, process diversity offers one big advantage:it allows different kinds of customers to be servedin different ways. The industrial customers whobuy Texas Instruments’ digital signal processingchips to put in their cameras and cellular tele-phones require rapid responses to design changes,whereas the retailers who sell calculators demandfast replenishment of standard products. Trying toserve both groups with the same order fulfillmentprocess would backfire, leaving each dissatisfied.Recognizing that fact, Texas Instruments allows itsbusiness units to design and manage their own or-der fulfillment processes.

Some companies have decided to standardize cer-tain processes but not others. Hewlett-Packard, forexample, standardizes procurement to gain lever-age with vendors, but it allows a variety of productdevelopment processes, reflecting the wide varia-tion in its products and in the customers who buythem. Johnson & Johnson has largely standardizedits R&D processes throughout its pharmaceuticalbusiness units to encourage them to share peopleand ideas and to enable all R&D projects to be man-aged as a single coherent portfolio. At the sametime, different units go their own ways in designingsales and manufacturing processes tailored to theunique characteristics of their products.

Our rule of thumb is that companies should stan-dardize their processes as much as possible withoutinterfering with their ability to meet diverse cus-tomers’ needs. However, we have learned that it’susually harder to impose standardized processesthan to allow diversity. A corporate executive pro-posing standardization will almost certainly be metwith a chorus of “but we’re different” from divi-sional general managers. Some of the resistancemay reflect legitimate concerns about whether astandard process can meet the needs of differentunits and different customers – and in those casesstandardization may indeed be a mistake. But theresistance may simply be the death rattle of divi-sional autonomy. General managers are accus-

tomed to seeing themselves as entrepreneurs run-ning their own businesses; the corporate center issupposed to give them resources and demand re-sults but otherwise keep out of their way. Whilecorporate executives should be prepared for this re-action, they should not give in to it. The rewards ofstandardized processes are great, and they’re worthfighting for.

Making the TransitionMaking the shift to a process enterprise involvesmuch more than just redrawing an organizationalchart. The changes we’ve discussed are fundamen-tal ones, representing new ways of managing andworking, and they are not easy to make. They re-quire the full attention and commitment of the or-ganization. Unfortunately, most companies todayare swimming – or sinking – in a sea of change pro-grams. (One large retailer we’ve studied stoppedcounting after 250.) The proliferation of changeefforts causes harm in many ways: it consumes re-sources, creates confusion, and encourages cyni-cism. Before launching a process enterprise initia-tive, management needs to take a hard look at all itschange programs, pruning those that aren’t relevantto process management and merging those that are.Distractions must be kept to a minimum.

The move to a process enterprise should be con-nected with an overarching strategic initiative. AtAmerican Standard, for instance, the building of a process enterprisewas positioned as a way to achieve thecompany’s long-term goal of reduc-ing working capitalby slashing cycletimes and inventorylevels. At OwensCorning, the effortwas linked with theERP implementation. At Duke Power, it was tied toderegulation, and at IBM, it was connected to creat-ing a truly global business. Other companies havelinked their programs to a move into electroniccommerce, the implementation of a merger, or theintegration of a supply chain.

One particularly effective way to underscore theimportance of the effort – and to help ensure its suc-cess –is to appoint high-profile, respected executivesas process owners. By putting its best people in thesepositions, management emphasizes the high priorityit places on process management and ensures thatthe process owners will be taken seriously.


In a process enterprise, the key structural issue is no longer centralizationversus decentralization; it’s process standardizationversus process diversity.


In addition to being focused on the transition, or-ganizations need to have a realistic sense of the sac-rifices and disruptions it will entail. A shift to a pro-cess enterprise isn’t a quick fix; it doesn’t happenovernight. American Standard announced its trans-formation into a process enterprise on January 1,1995, but it hasn’t yet completed its journey. IBM,Duke Power, and the other companies we have dis-cussed are also still working on aligning some as-pects of their businesses with their processes. Exec-utives need to prepare themselves for years of effortand set the organization’s expectations accordingly.

Not everything needs to be done at once, of course.Process owners should be appointed immediately,as they will guide the entire effort. A process-basedmeasurement system should be established at theoutset to track the effort’s progress. But expendi-tures on employee-training programs, compensa-tion systems, and other costly or complex infra-structural elements can often be deferred. (See thesidebar “The Infrastructure of the Process Enter-prise.”) Instead of trying to build a companywideinfrastructure at the start, it’s best to focus first onachieving some tangible benefits quickly. Without


MeasurementMost businesses lack rigorousmeasures for their processes. Theymay know their manufacturingcosts and their product sales downto the penny, but they don’t knowexactly how often they fill ordersflawlessly or precisely how long ittakes a new product to go fromconception to profitability. Indeed,they’re usually not even sure whataspects of their processes theyought to be measuring. Their mea-surement systems conform to thevery organizational boundariesthat their processes transcend.

In moving to a process enter-prise, therefore, managers need toconduct a thorough analysis to de-termine what aspects of processperformance are most directlylinked to achieving the organi-zation’s overall objectives. DukePower has conducted such ananalysis. It identified its overarch-ing strategic goals – such as pro-viding reliable and competitivelypriced electric power and hassle-free customer service – and has de-termined how each of its processeswould affect those goals. It then es-

tablished relevant process perfor-mance measures. For the DeliverProducts and Services process, themeasures include the percentage of projects completed by the datepromised to the customer, the per-centage of installations done cor-rectly the first time, and the time it takes the call center to respondto a customer’s inquiry. Measuresfor the Provide Reliability and In-tegrity process include the numberof outages, the number of outageslasting more than two hours, andthe accuracy of restoration timesgiven to customers who have lostpower.

Process owners not only use themetrics to track the status of a pro-cess and guide improvement ef-forts, they also disseminate themthroughout the organization to re-inforce people’s awareness of theprocess and to focus them on itsperformance. Since the same pro-cess measures are used to gauge theperformance of everyone involvedin the process, the metrics alsohelp to reinforce teamwork.

Compensation If frontline personnel and man-agers are to focus on processes,their compensation should bebased at least in part on how wellthe processes perform. All processteams at Allmerica Financial haveconcrete performance goals set bythe process owners, such as targetsfor the time required to process ap-plications and the percentage ofcontracts issued without errors.The team members receive bonusesbased on achieving those goals, andthe process owners can award addi-tional bonuses to members whomake outstanding contributions.At American Standard, the com-pensation of process owners isbased on three factors: process per-formance, business sector perfor-mance, and corporate performance.The heads of regional businessunits at Duke Power are assessednot only on the bottom line of theirregions but also on how well theymeet their process goals.

Traditional ways to measure perfor-mance, determine compensation,provide training, and even organizefacilities are tailored to verticalunits, not processes, and to individ-

uals, not teams. Companies mak-ing the shift to a process enterprisewill need to take a fresh look atmany of the basic elements of theirorganizational infrastructure.

The Infrastructureof the ProcessEnterprise


clear early signs that the desired gains will materi-alize, people will grow anxious and begin to resistthe changes, and the entire effort will lose momen-tum. At Texas Instruments, for example, the suc-cess of the product development process helped con-vince the organization of the virtue of processmanagement, and the company is now extending theapproach into supply-chain, retailer-engagement,and other processes.

Companies with many business units havesometimes found it useful to designate one unit totake the lead. That unit becomes a kind of organiza-tional prototype. Through its experience, the com-pany as a whole can identify and rectify problems,promote benefits, and set a course for others to fol-low. At John Deere, for instance, two divisions havetaken the lead in becoming process organizations:John Deere Healthcare and one of the equipment-manufacturing units, the Worldwide ConstructionEquipment division. Other divisions within the

company now have the opportunity to learn fromtheir experiences and build on their best practices.

Because the changes involved in becoming a pro-cess enterprise are so great, companies can expectto encounter considerable organizational resis-tance. We have found, though, that it’s rarely thefrontline workers who impede the transformation.Once they see that their jobs will become broaderand more interesting, they are generally eager to geton board. Rather, the biggest source of resistance isusually senior functional executives, divisionheads, and other members of the top managementteam. These senior executives will often either re-sent what they see as a loss of autonomy and poweror be uncomfortable with the new, collaborativemanagerial style. If allowed to become visible, theirreluctance will soon be amplified throughout therest of the organization. CEOs, therefore, need totake particular care in communicating to unitheads, involving them in the change effort, and


FacilitiesIn most companies, people arehoused in vertical departments, according to their function, theirregion, or their business unit. Butbecause processes cut across thosevertical divisions, process workersneed to be drawn from them into anew location where they can workas a team. At Owens Corning, forinstance, many different employ-ees are involved in filling an order,from customer service represen-tatives to transportation coordi-nators to accounting personnel. In the past, each of those peopleworked in a separate location, sur-rounded by others in the samefunctional specialty. Now all thoseinvolved in order fulfillment are located together. By sharing thesame facility, they get a better viewof the entire process, and they are able to exchange ideas easily.American Standard has undertakena radical program of co-location,creating shared spaces for all of its process teams. When all work is process work, all space becomesprocess space.

Training and Development In traditional organizations, manypeople have relatively narrow jobsand need to know little outside thescope of their own department. Fora process team to succeed, how-ever, all the members must under-stand the whole process and howtheir individual efforts contributeto it. Usually, workers will need tobe trained to take on their broad-ened roles. Duke Power, for in-stance, puts all its linemen througha class called “Thriving in a Pro-cess Organization,” which givesthem a basic grounding in the elec-tric power industry, covering suchtopics as deregulation, utility coststructures, and customer require-ments. It also gives them an appre-ciation of the concept of a businessprocess, a detailed understandingof their own process, and trainingin the personal skills needed towork collaboratively.

Career Paths There is less need for middle man-agers in a process organizationthan in a traditional one. Processowners design and measure theprocess, and process teams carry itout, overseeing their own work andmaking all the day-to-day oper-ating decisions required to keepthings moving smoothly. As a re-sult, most of the rungs on the tra-ditional managerial career ladderdisappear. A process enterprisetherefore needs to develop new ca-reer models that are not based ontraditional hierarchical advance-ment. Allmerica Financial, for in-stance, offers employees two newcareer models. One is based onmastering a specific insurance dis-cipline, such as claims handling.Claims personnel who developgreater knowledge and skills are assigned more complex claims andget a higher base pay – without aformal change in level. The othermodel offers a career path throughmany parts of the company – fromclaims to IT to underwriting, forinstance.


gaining their full commitment. They should be pre-pared to dismiss anyone who steadfastly refuses tosupport the initiative. In our experience, it is notuncommon for anywhere from a quarter to a half ofthe senior team to leave –voluntarily or otherwise –during the changeover.

Looking to the FutureGiven the challenge of shifting from a traditionalbusiness to a process enterprise, some may wonderif it’s worth it. We believe that, for most companies,there is really no alternative. Process managementis not merely a way to address specific problems –poor quality, say, or high costs. It is a platform forcapitalizing on new opportunities.

Take e-commerce. The cutthroat world of the In-ternet places a premium on the swift and flawlessexecution of processes. As Amazon.com and othere-commerce leaders have discovered, if you deliverorders on time and with no problems, customers re-turn to your site. If you botch orders, customerswon’t give you a second chance. Putting a Web site

in front of a flawed process merely advertises itsflaws. The same goes for business-to-business e-commerce. If your processes are not totally reli-able, you can forget about being a supplier to Dell orany other of today’s turn-on-a-dime manufacturers.

But just as important as having smooth, efficientprocesses is being able to redesign those processeson the fly. From order fulfillment to customer ser-vice to procurement, operating processes are rarelyfixed any more. They must change their shape asmarkets change, as new technologies become avail-able, and as new competitors arrive. Look at IBM.Having successfully redesigned most of its pro-cesses over the last few years, it is now redesigningthem all over again to make them compatible withthe Web. Without the flexibility inherent in a pro-cess enterprise, it would be next to impossible forIBM, or any company, to shift processes quicklywithout disrupting its entire business. A processenterprise is the organizational form for a world inconstant change.

Reprint 99607 To place an order, call 1-800-988-0886.



how process enterprises really work - caliw.cali.org/conference/2000/program book/hbr2.pdf ·...

Documents