| This article is also available on our | | | | in their specific areas, so that they |
| website: PROACTION - Generating Best | | | | can demonstrate the system and train |
| Practices. It is an excerpt of a paper | | | | others, as well as effectively determine |
| originally written by George Miller, | | | | implementation approach (with consulting |
| Founder of PROACTION. It has been | | | | help), priorities, etc. Next, IT people |
| modified and updated by Paul Deis, | | | | need to be ready to perform technical |
| PROACTION CEO. | | | | work and ongoing support. Project |
| INTRODUCTION | | | | leaders/middle management need to |
| This case study describes lessons | | | | understand the big picture and how their |
| learned from the ERP implementation at a | | | | own areas will work and interface with |
| leading metal container products | | | | other areas. Executives need to |
| company. It is a $200MM, 53-year-old | | | | understand the big picture, what systems |
| company, headquartered in the USA, with | | | | can and can't do and what are the |
| facilities in multiple states and in | | | | priorities and tradeoffs. |
| Europe. It is a leading producer for the | | | | Finally, end users need to be as |
| wholesale, retail, OEM and industrial | | | | comfortable as possible with their new |
| markets. The company was in the midst of | | | | tools BEFORE going live. It became |
| a successful turnaround during the | | | | apparent 2 ½ months before |
| project, which was initiated primarily | | | | implementation that additional training |
| to deal with the Y2K problem, as a new | | | | was needed for effective utilization of |
| management team provided new direction. | | | | the system. Management was quite |
| During the project, the company was also | | | | serious about rectifying the situation. |
| transitioning to focus factory | | | | The Human Resources Dept. managed this |
| management, flow manufacturing, product | | | | effort. Standards of performance for |
| rationalization and reorganizing | | | | each employee were developed. The |
| business units, all this in the midst of | | | | employees, trainers, supervisors, |
| a major facilities consolidation. In | | | | managers, and ultimately, executives, |
| spite of all this, the project was | | | | were held accountable for employee |
| probably in the top 10% of ERP efforts. | | | | competence. Formal knowledge |
| There was a constant personnel shortage | | | | assessments were conducted and remedial |
| and competition for mindshare. Due to | | | | action taken. The vast majority of |
| inevitable time and resource | | | | employees were ready on the "go live" |
| constraints, the phase I implementation | | | | day, and it showed in the results. |
| consisted only of replacing legacy | | | | One notable shortcoming in training |
| systems and the most urgent partial | | | | resources: nobody-company personnel, |
| reengineering of at-risk processes. The | | | | Oracle or third party person-seemed to |
| company was successful in transitioning | | | | know the entire system well enough to |
| to the improved system and initiating | | | | provide the big picture. We had to spend |
| follow-on process reengineering. | | | | an inordinate amount of time and money |
| Currently (mid-1999), system | | | | consulting with all kinds of people to |
| implementation in other North American | | | | get cross-functional, cross "module" |
| and European sites is underway. | | | | answers. We still have more work |
| The founders sold the firm to the | | | | required to fully develop such people. |
| present owners in 1996, who faced a | | | | However, the company does have a couple |
| number of challenges to bring the | | | | of excellent "generic" business systems |
| company up to their expectations: | | | | gurus available, which has proved |
| • Need for more emphasis on quality | | | | invaluable. We recommend spending the |
| • Debt load to service initially | | | | time and money to beg, borrow, steal or |
| inhibited capital appropriations | | | | develop cross-functional experts to know |
| • Difficult facilities consolidation | | | | your business and applications systems. |
| project | | | | Our Superusers and Project Leaders must |
| • Aging, obsolete and unresponsive | | | | have really been good, because a lot of |
| business systems, with a Y2K problem | | | | them have already been promoted/moved on |
| • Product development | | | | to better jobs. We have realized this |
| • Competition | | | | and are working to develop a better |
| In that context, the company was | | | | ongoing training/successorship approach. |
| attempting to modernize, rightsize and | | | | Any organization needs a comprehensive |
| streamline for the future, but nearly | | | | generic business systems education |
| lost the race with time. About six | | | | program to ensure that people are aware |
| years ago, the company engaged a major | | | | of the state of the art. An APICS |
| consulting firm to study major | | | | trainer is providing overview education |
| improvement opportunities. One of the | | | | for the company. Some seminars are |
| chief recommendations was to upgrade ERP | | | | being attended on things like |
| systems, but due to many other | | | | forecasting, JIT, etc. A wider-reaching |
| priorities, the company only started | | | | program is still needed. |
| moving on the recommendations in | | | | Thorough Conference Room Pilot, testing, |
| mid-1997, with a looming Y2K deadline. | | | | simulation are needed, including stress |
| Therefore, schedule and implementation | | | | testing |
| priorities were somewhat mandated by the | | | | When all is said and done, the team must |
| situation. | | | | thoroughly exercise the system In order |
| The owners brought in new management in | | | | to know and assess its strengths, |
| the summer of 1998 and they began | | | | weaknesses and different ways they can |
| building their new team of high quality | | | | solve problems with it, in order to iron |
| professionals, including key valued | | | | out bugs, procedural weaknesses, needed |
| members of the existing team. | | | | organizational changes, documentation |
| Fortunately, the former management had | | | | and training requirements. Realistic |
| the presence of mind to initiate the ERP | | | | scenarios and data are needed. |
| project and to staff it with mostly | | | | The company got real serious about this |
| excellent people. The new management's | | | | a few months before implementation, |
| major initial challenges and priorities | | | | beefing up the conference room pilot |
| involved dealing with a difficult plant | | | | activities, making them a centerpiece of |
| consolidation, product shortages, | | | | the effort. This paid off in building a |
| customer issues, and that little Y2K ERP | | | | good implementation model, testing it |
| problem also. | | | | and helping in documentation and |
| The unresponsive business systems | | | | training efforts. |
| weren't solely attributable to ancient | | | | Two highly realistic stress tests |
| software. There were also data integrity | | | | exercised the system not long before the |
| problems in inventory, purchasing and | | | | go-live date. Minor defects, which |
| production. Practices and procedures | | | | might have caused major problems, were |
| were in need of updating. Much training | | | | quickly identified and corrected. |
| was needed. Existing systems tools were | | | | Conversion programs were also tested |
| not even fully implemented. There were | | | | again and again to avoid unpleasant "Day |
| philosophical differences between the | | | | 1" surprises. |
| way the business was run and how the | | | | If You Insist on a "Big Bang" |
| systems were designed to work, not an | | | | Implementation, then do it right |
| atypical situation in many companies. | | | | The author is not in favor of this |
| THE ERP PROJECT | | | | implementation approach, of putting the |
| The company had a "legacy" system, | | | | whole new system on-line at once. But, |
| consisting of Data 3 MRPII, HFA Order | | | | it had been decided to go that way long |
| Entry/Accounts Receivable, S2K | | | | before his arrival, so we all pitched in |
| (Infinium) Financials and the old ADP | | | | to minimize the risk. The most important |
| Payroll system. These were all pretty | | | | things to accomplish for that were: |
| good products in their time, but hadn't | | | | thorough testing, testing, testing, |
| been kept up to date, although vendors | | | | exhaustive conference room piloting, |
| did offer newer versions, but required | | | | sound technical execution, testing, |
| customization made some impractical. | | | | training and a good contingency plan. |
| For a variety of reasons, beyond the | | | | Did I mention testing, also? The |
| scope of this paper, the company chose a | | | | accountants were a little more |
| very expensive but good set of | | | | conservative, bringing up the general |
| platforms, including Oracle applications | | | | ledger first and running in parallel |
| and systems software, Hyperion Pillar | | | | three months earlier than everyone else. |
| budgeting, ADP PC-based payroll, ADP HR | | | | • Strong technical, systems |
| systems, HP hardware and HP-UX operating | | | | programming/DBA (Database |
| system and other mass storage, | | | | Administration) capabilities are needed, |
| networking and data warehousing | | | | at least for a system as complex and |
| products. | | | | flexible as Oracle. |
| The company set up an implementation | | | | • There are numerous tables, |
| team consisting of an executive steering | | | | parameters and security to set up and |
| committee, project manager (from | | | | maintain software versions and patches |
| Engineering!), and project leaders from | | | | to control, for various system and |
| middle management and 10 "Superusers," | | | | database "instances." A good DBA is |
| mostly high potential administrative | | | | also a consulting resource to IT, Users |
| people, to become their departments' | | | | and management. |
| experts in the new system. In addition, | | | | Conventional wisdom says that modern |
| there were a number of Oracle technical | | | | software packages take most of the |
| and applications consultants, as well as | | | | technical risk out of an implementation. |
| third party contract programmers, | | | | Conventional wisdom is incorrect. Unless |
| database administrators and business | | | | we tackled technical issues head-on, |
| consultants. | | | | such as system setup, multi-site |
| There was a fair amount of personnel | | | | definition, library/version control, |
| "churning," until the company got the | | | | conversion, etc., it would have meant |
| right combination of Oracle and other | | | | trouble. Technical factors became a |
| outside consultants. Several internal | | | | non-issue precisely because we did apply |
| team members fell by the wayside, | | | | the resources to deal with them. |
| principally due to operational workloads | | | | We relied heavily upon outside resources |
| or people moving on. The team leaders | | | | in the beginning and gradually tapered |
| were especially vulnerable to their | | | | these off as we built the skills |
| heavy departmental workloads and some | | | | in-house. We need to continue to broaden |
| were not able to devote sufficient | | | | the skills base to ensure that multiple |
| energy and time to the project success. | | | | in-house employees can handle critical |
| To a significant extent, the Superusers, | | | | technical tasks, a common aspiration in |
| outside consultants and IT personnel | | | | mid-sized Information Technology |
| picked up the slack. | | | | departments. |
| Progress was relatively slow in the | | | | The company was successful in building |
| initial months, even requiring a | | | | an effective Oracle RDBMS/UNIX/Oracle |
| rededication at one point. But the | | | | Apps, technical capability, scaled |
| company and the team never gave up and | | | | appropriately for maintaining a |
| ultimately progressed to a successful | | | | "vanilla" version of the system. It |
| and fairly smooth implementation, but | | | | took more people than we would have |
| not without a struggle during the | | | | liked. |
| preparation. | | | | Ensure data accuracy |
| SCOPE AND OBJECTIVES | | | | Formal ERP systems are very sensitive to |
| A. Due to the aforementioned issues, | | | | data accuracy. We had to make big |
| project scope and objectives were | | | | improvements prior to implementation. |
| limited to the following: | | | | Even though our outside auditor |
| • Y2K compliance (actually the main de | | | | "blessed" our overall numbers prior to |
| facto schedule and objectives driver) | | | | implementation, these still did not meet |
| • Timely, accurate financials | | | | our far more rigorous internal |
| • Closed loop MRPII | | | | standards. We are only now approaching |
| • Strong IT infrastructure for future | | | | world class standards in most areas, but |
| growth, flexibility, integration, | | | | actually slipped in one area after a |
| stability and profitability | | | | reorganization, spurring a rededication |
| B. Not included, but desired: | | | | to data accuracy. In the future, we |
| • Business Process Reengineering | | | | will conduct rigorous quarterly internal |
| • Automated Data Collection/bar coding | | | | inventory accuracy audits and look at |
| • Freight/Warehouse system | | | | other areas as well. We are |
| • Flow Manufacturing | | | | simultaneously trying to reduce |
| • Full ERP | | | | dependency on formal data accuracy by |
| • Supply Chain Planning | | | | moving to more visual control approaches |
| C. Applications addressed were: | | | | and quick reaction systems. |
| • Financial: General Ledger, Accounts | | | | Minimize software modifications |
| Payable, Accounts Receivable, and | | | | Everyone we knew in other companies |
| Financial Planning/Budgeting | | | | making major changes to a software |
| • Manufacturing: Master Production | | | | package ended up spending huge amounts |
| Scheduling/Material Requirements | | | | of money and time and often had |
| Planning, Inventory, Work-in-Process, | | | | maintenance and operational problems. We |
| Capacity Requirements Planning, | | | | learned from their experiences, made |
| Engineering/Bill-of-Materials, Costing | | | | minimal changes and had minimal |
| • Order Management; Order Entry, | | | | problems. |
| Invoicing | | | | Control the scope |
| Addressed separately: Forecasting (Smart | | | | Although there were people who wanted us |
| Forecast), Payroll (ADP PC version), | | | | to do a lot more, most of us agree that |
| Human Resources (ADP), Aggregate | | | | scope control helped to get the job done |
| Inventory Management (IQR System), | | | | cheaper and faster than it would have |
| Returns and Warrantee (homegrown), | | | | otherwise. It would have really |
| Quality (Homegrown and PC packages), | | | | stretched things out and jeopardized the |
| OADW (Oracle Applications Data | | | | results if our plate had been even more |
| Warehouse). | | | | full than it was. |
| LESSONS LEARNED | | | | Have Contingency Plans |
| The company learned some important | | | | After seeing and hearing about others' |
| lessons during this project, some | | | | past disasters, we developed a |
| relatively painless, some didn't come so | | | | contingency plan for: what to do if the |
| easily. To save time, only the ones | | | | implementation failed; what to do when |
| deemed most important are listed below. | | | | the system crashes- for an hour, a day, |
| Major executive involvement is crucial | | | | a week, longer? This was addressed. |
| to success. | | | | There is also an alternate backup site |
| Executives control the staff, budget, | | | | in place. We can always improve with an |
| punishments and rewards. In the cases | | | | ongoing readiness audit. |
| where we failed to engage executives, | | | | Have a Go/No Go decision point |
| results usually suffered. When engaged | | | | We did it and are glad we did. We waved |
| and informed, they usually improved the | | | | off the implementation date twice, |
| situation. Notable examples: | | | | because we wanted additional |
| • When it became apparent that | | | | preparation. By the third time, we knew |
| training was seriously deficient, | | | | what to look for, knew what we needed, |
| several VP's and the CEO took decisive | | | | and to some extent, even knew what it |
| steps to alleviate the problem. | | | | was we didn't know. We made a list of |
| • When we noticed that contingency | | | | the outstanding issues that had to be |
| planning was potentially deficient, the | | | | resolved before we would "go live" with |
| CEO made it a personal crusade and | | | | the implementation and worked them hard, |
| insisted upon a plan before he would | | | | as a team. A couple of months' delay is |
| approve the implementation go-live date. | | | | insignificant compared to years of pain |
| The Steering Committee was the principal | | | | from a failed implementation. Never |
| tool employed for executive engagement. | | | | implement until your people are |
| All of the executives whose departments | | | | confident and ready. |
| were most affected were on the steering | | | | Other |
| committee, along with the CEO, IT Mgr. | | | | A few additional hints-Don't put up the |
| (who was also the Project Manager) and | | | | latest version of the vendor's software, |
| an outside business consultant. As it | | | | no matter how they praise its virtues. |
| inevitably happens in situations where | | | | Major ERP packages are never |
| busy people have multiple priorities, it | | | | perfect-ESPECIALLY major new releases. |
| was a chore to get everyone to show up | | | | Let others debug it for you, even if it |
| at meetings, do critical reviews and | | | | means waiting a little longer for some |
| make decisions. When it was successful, | | | | bells and whistles. We took the |
| which it wasn't always, several people | | | | conservative route and avoided much |
| had to get together in advance to push | | | | agony. Make sure you have key |
| an agenda and line up participation. | | | | management reports in place before going |
| Thankfully, the new CEO, once he got on | | | | live. We didn't have them all and paid |
| board and saw how important this project | | | | the price. Make sure a critical mass is |
| was, proved invaluable with his support | | | | trained in report writing. We didn't, |
| when it was most needed, although we had | | | | but when offering training the second |
| to carefully choose our battles. | | | | time around, classes were "sold out" |
| BPR (Business Process Reengineering) | | | | immediately. Don't scrimp on hardware |
| should be done up front. | | | | resources. You'll get punished more for |
| The company elected to put the system up | | | | bad system performance than for spending |
| first and reengineer later. This was | | | | too much on hardware. Once we eliminated |
| primarily attributable to budget and | | | | a few bottlenecks in testing, |
| schedule reasons. The downside is that | | | | performance became a non-issue. |
| the system, as implemented, tends to | | | | Establish specific systems performance |
| perpetuate the status quo, although | | | | objectives, get commitments to them in |
| certain improvements came about just | | | | writing. |
| from installing it. Some minimal | | | | Reward contributors any way you can-with |
| reengineering was done where time | | | | praise, training, coveted team |
| permitted or where the old procedures | | | | assignments, promotions, etc. Your good |
| were almost totally unworkable. Now, | | | | people are your finest asset. |
| we're going back and working on the | | | | EPILOGUE |
| processes with the best improvement | | | | The new system went live on 4 January |
| potential, but not as fast as we'd like. | | | | 1999 and we never looked back. There |
| Objectives/metrics/issues should drive | | | | were only two serious problems. |
| the project. | | | | • Bugs in order entry slowed us down. |
| Budget and schedule drove this project | | | | Better testing and documentation of |
| and it showed. Metrics were developed, | | | | results would have mitigated this |
| but, key performance reports are still | | | | problem. |
| being brought on-line. In the instances | | | | • In addition, an administrative |
| where the objectives were stressed, | | | | problem resulted in a number of blanket |
| improvements were made. Examples: the | | | | vendor purchase orders not being put on |
| CEO and Finance stressed 1)- Pricing | | | | the system for weeks, screwing up |
| rationalization. Margin improvements and | | | | purchasing and inventory records for the |
| simplified order processing have already | | | | duration. |
| resulted. 2)- Inventory accuracy was | | | | We had our first major system crash in |
| stressed as a prerequisite to system | | | | March, 1999. Recovery procedures took |
| performance- improvements are gradually | | | | nearly a day, but restoration resulted |
| occurring. | | | | in zero lost data, using the task level |
| The team did do a good job making | | | | recovery features of the system. |
| business issues drive the project, | | | | We have lost some key trained talent and |
| towards the end, as their expertise | | | | are trying to improve our successorship |
| developed. One of the main benefits of | | | | and training management. |
| the project was the training of a cadre | | | | The reengineering effort seems to be |
| of people prepared to run a formal | | | | more of a gradual, continuous |
| business system. Now they're ready to do | | | | improvement approach. It needs to move |
| a great job on the next implementation! | | | | faster. This will take more emphasis and |
| Management is driving needed | | | | rededication. There are lots more things |
| improvements and forcing the system to | | | | that we can do to improve. Other |
| adapt. For example, they are addressing | | | | important projects seem to present even |
| customer service and inventory issues | | | | better opportunities at this time. |
| effectively. | | | | More importantly, "the vision thing" is |
| Effective training is needed at all | | | | back and sales and profits have been |
| levels. | | | | great. The system is working and helping |
| The first priority of key project people | | | | to provide a firm foundation for the new |
| is to be familiar with the overall | | | | century! |
| principles of ERP and the applications | | | | |