Getting Your ERP Implementation Back on Track

Project Auditors, LLC

Frank R. Parth, MS, MBA, PMP

Joy Gumz. CPA, PMP

March 27, 2003

Déjà vu, all over again.

Well, it’s happened again. The state of California, or at least a part of it, decided to replace an outdated, bureaucratic, and inefficient system with an off-the shelf software solution. Despite the best intentions, the implementation of PeopleSoft by the California State University (CSU) system has become mired in schedule delays, cost overruns, second-guessing, a strongly worded State Auditor’s report, and politics. Is this unique? Certainly not. It has happened repeatedly.

The state has a history of spending millions of dollars on poorly implemented or failed technology solutions. Witness the $44 million failure in the Department of Motor Vehicles to implement high tech solutions or the $40 million technology failure of the Youth and Adult Correctional Agency in 1997. By the middle of 2002 Governor Davis had fired four of his top aides for major failures in the state’s computer system. The variety of software we call Enterprise Resource Planning (ERP) systems, which includes PeopleSoft, SAP, Oracle, and others has its own history of failed implementations. Companies have spent millions of dollars on installing these systems and have thrown then out when they realized it was not going to give them what they wanted.

In some cases companies have spent so much money on failed implementations they ended up in bankruptcy. A high percentage of ERP implementations have resulted in litigation and in seriously damagedreputations of the major consulting firms that implemented them. Many of the technology solutions implemented by state governments end up overrunning their budgets, delivering late, and not providing the results expected. These are just symptoms of the root causes, however, and we will discuss the causes themselves later. Combine the complexity of the environment with the complexity of the technology and its an almost guaranteed disaster. Cutting through all the hype, the politics, and the newspaper attacks, Is PeopleSoft really a greedy company selling bad software at outrageous prices? No. Is the state government so bumbling and inept that it can’t manage a software installation without major failures and bad publicity? Again the answer is no.

It’s A Complex EnvironmentThe results of a survey done by theStandish Group in 1996 illustrate that30% of technology projects done bygovernment were failures and werekilled before they were completed.Another 52% finished but were overbudget, behind schedule, or deliveredfewer features than were originallycontracted.Why is this? What causes the failurerate to be so high? As people at alllevels of government will tell you it is avery complex environment. There aremultiple agencies, each with their ownpriorities, budgets, personnel, andstrong determination to ensure that theirrequirements are met.Any time there is competition for limitedresources political battles are to beexpected. When you’re managing one ofthe largest economies in the world witha shrinking budget, any tactic that canget your department more money isconsidered fair, even if it comes at theexpense of someone else’s program.Another reason is the technology itself ismoving faster than the cumbersomeprocesses for purchasing andimplementing it. A state processdesigned for buying mainframecomputers can easily break down whenit’s asked to install a 3-tier client/serversystem at twenty five separate fieldoffices and tie them together with adedicated network.

It’s the Technology, Stupid!While the environment may be highlycomplex and politically-driven, thetechnology of ERP systems itself is socomplex, and crosses so manyorganizational boundaries, that it oftenfails even in a less complexenvironment. Coca-Cola spent $10million on its SAP implementation andthen threw it out when it failed to deliverthe expected benefits.In a highly publicized case, ClevelandState University threatened to suePeoplesoft when it realized it could notprocess more than half of the 4,000+financial aid applications filed bystudents in 1998. The universitycontinued the implementation althoughthe final cost exceeded $15 million,$10.8 million over the original budget.The decision to forge on was madeeven when it was realized key businessrequirements in the financial aid areawould not be implemented until near theend of the six year adoption period.At Ohio State University the budget forPeopleSoft's human-resources systemsoared from $53 million to $85 million.At University of Minnesota, the originalbudget allocation was increased from 38million, then 53 million, and finally 60million.In January 2000 seven Big 10 schoolswrote a joint letter to PeopleSoftdetailing their frustrations with thesystem and asking to work more closelytogether to alleviate them.Commonly cited reasons as to why

ERPs don’t meet project objectives
include:1)

Unclear business objectives2) Lack of engaged, visible leadershipat executive level3) Poor communications4) Lack of project methodology, or pooradherence to the methodology used5) Resistance to change within theorganization6) Failure to prepare the organizationfor change, including inadequatetraining7) Failure of user departments to takeownership8) Lack of experience of project team9) Incomplete requirements definition.

The Usual Suspects

Based on the accumulated experienceof thousands of projects the field ofproject management can call up astandard set of answers.Is your project behind schedule?• You need to do a better job up frontestimating the tasks involved in theproject.Did you not deliver what was expected?• You should have defined yourrequirements better.

Over budget?•

You should have used one of thebudget estimating tools.And so on. These are easy answers thatreflect the fact that most technologyprojects fail for the same reasons.But answers that are so easy to give arevirtually impossible to put into practice ina real world environment. Just asaerospace and defense contractorslearned decades ago how to tailor theirproject management to the complexitiesof dealing with high risk federalgovernment contracts, so mustconsulting companies learn how todeliver solutions in the complex andchanging environment of stategovernments.Are there solutions being found? Yes,but slowly.

The state of Utah has an e-governmentcommittee that approvesand priorities the initiatives. Californiaitself has learned to break large projectsdown into smaller pieces that are moreeasily managed.Recently the Quality Assurance Institutehonored North Carolina with its "Best ofthe Best" Award for its third partyindependent quality assurance reviewprocess. NASCIO, the NationalAssociation of State CIOs, sees this asa future trend for state government.The University of Wyoming is currentlyupgrading their Peoplesoft system toversion 8.x. In the process they foundthe changes to the underlying system,architecture and technology was morecomplex than originally anticipated.They scheduled additional training midwaythrough the project and are nowbeginning to feel more comfortable withthese technical challenges.We cannot change the complexity of thestate government, nor can we makecomplex software trivial to install. But wecan modify the process to identifyproblems early. Part of the process mustbe a completely independent audit ofthe project throughout its delivery cycle.

The project manager cannot audit theproject because he or she is far tooclose to the details and cannot give anobjective assessment. The consultingcompany, no matter how big, can notprovide resources that are trulyindependent of the project.Similarly, the state cannot audit theproject in a truly independent fashionbecause they have a vested interest inthe outcome of the audit. While the CSUPeopleSoft project was initially justifiedand approved on one set of financialrules, it was audited under another,different set of rules resulting in a StateAuditor’s report that claimed the projectwas significantly over budget.Only a truly independent, objectivereview of the project can identifywhether the project will deliver theresults expected, or recommend anunbiased action plan. Although thefacilitation and planning processes forsuch a review may be done internally,the quality assurance activities often arebest performed by expert resourcesexternal to the project.These expert resources should befamiliar with performing quality reviews,whether technical reviews or userreviews. They should be independent ofthe project’s organization and report toupper management, usually the Boardof Directors, the Executive sponsor forthe project, or to the state auditor or ahigh-level committee.

Recommendations
1) Problem

Everyone involved in a project, both onthe government side and on thecontractor’s side, has a vested interestin the outcome of a project review.Unfortunately those interests nevercoincide. Projects rarely fail due to onlyone party being incompetent and theother blameless. Both participants in theproject are typically responsible forproblems. But any audit done by oneside or the other is going to have a bias.SolutionThe initial contract with theimplementation partner and/or softwarevendor should include a section allowingindependent evaluation, inspection, andtesting to address performance. Thisclause many never be exercised but itprotects the customer by alerting thesupplier that their work will be potentiallysubject to a quality review by anindependent third party. In the event thepurchaser elects to invoke this option,the contract sets expectations of bothparties regarding the quality review andsubsequent action plan.

2) Problem

Because the environment of the stategovernment, or even a part of it, is socomplex any simplistic recommendationto gather requirements or to plan betteris simply not practical. Each part of thegovernment is behaving the way itshould – it is looking out for its own bestresult. However, the end result is anenvironment where no single solution islikely to make everybody happy.SolutionThe best solution for this is to set up agovernance committee composed of asingle representative from each groupaffected by the system being installed.The representative should have thepower to commit their organization tothe solution. They don’t necessarilyhave to be happy with the solution, butthey must be willing to live with it.

3) Problem:

Too many projects are so large that thecomplexity of the projects almostcertainly guarantees failure. Breakingthem into smaller independent projectsjust means managing more projects anddoes not reduce the overall complexity.You’ll just have more project teams anda wider variation on the results.Solution:Break projects into phases, with eachphase delivering a subset of the overallfeatures. The project will have a longerplanned schedule before the full featureset is delivered, but will deliver usablefunctionality earlier and the end resultwill be completed in less time becauseless rework and integration will have tobe done.

4) Problem:

People who are assigned to the projectoften don’t have all the knowledge andskills they need, both on theimplementer’s side and on theorganization’s side. The results are thatproject plans don’t reflect the realities ofimplementing such complex systemsand the end users are oftendisappointed in what they get.Solution:Provide adequate training for thecontractors implementing the solutionsas well as for the users who will have tosupport the project and live with theresults afterwards. The training shouldinclude: project management training,change management, process design,and ERP-specific training.

About the AuthorsFrank R. Parth, MS, MBA, PMP is thePresident of Project Auditors, LLC, a projectmanagement, training, and auditingcompany headquartered in southernCalifornia. Mr. Parth has over 25 yearsexperience in technology development. Heis a published author and an internationalspeaker in project management and teachesproject management courses for localuniversities. He obtained an MBA from thePeter F. Drucker Graduate ManagementInstitute and is a certified ProjectManagement Professional by the ProjectManagement Institute.

Joy E. Gumz, CPA, PMP, is a Director atProject Auditors. She is an acknowledgedexpert on enterprise software andtechnology implementations with in-depthexpertise in commercial off-the-shelfsoftware for financial and human resources.She has held senior consulting positionswith Peoplesoft, Plumtree Software, andErnst & Young, and has workedinternationally. She has published articles inPM Network, a Project ManagementInstitute publication. Ms. Gumz is also areviewer for Cost Engineering, thepublication of the Association for theAdvancement of Cost Engineering, Intl, andComputing Reviews, the publication for theAssociation for Computing Machinery.

About Project Auditorshttp://www.projectauditors.com/Project Auditors LLC offers advancedproject management training andservices to organizations seeking toimprove their business processes andproject successes. Our online projectmanagement dictionary is recognized asthe leading reference tool for projectmanagement professionals and our listsof project management tools andorganizations is a favorite bookmarkamong many companies worldwide.Project Auditors is dedicated to servingthe project management needs ofFortune 500 companies, U.S. federalgovernment agencies, andentrepreneurial businesses in NorthAmerica and Europe. We provide projectmanagement training, project qualityaudits, and organizational projectmanagement maturity model (OPM3)assessments. A privately held Limited LiabilityCorporation headquartered in southernCalifornia, Project Auditors was foundedin 1994.