Carton Adam ECIME 07

7/25/2019 Carton Adam ECIME 07

1/13

Integration of Enterprise Applications and Functional fit: the Role of ERP in SupportingManagement Objectives

Fergal Carton, Frdric AdamUniversity College Cork, Ireland

[email protected]@afis.ucc.ie

Abstract:Companies have been investing in integrated enterprise applications such as !"#$ forover a decade, %ithout firm evidence of a return from these investments. &uch research hascenteredonthe factors %hich %ill lead to a successful implementation pro'ecteg( )olland and*ight, +- hanks and eddon, /000$, but to date there appears to be little research on thelonger term impact of !"# systems on the organisation )eili and 1inck, /002$. Although thegreater level of system integration brought on by !"# has meant that there is more operationalinformation available to managers than ever before, the information stored in !"# applicationsre3uires much off4line manipulation in order to be meaningful to managers. 5he data held in !"#databases originate in physical processes that evolve over time, and thus inevitably a gap opensbet%een the !"# system, and the reality it is designed to capture *ee and *ee, /000$. 5akingthe evaluation of management performance against organisational ob'ectives as research

domain, and focusing on a case study in the pharmaceutical sector, this paper looks at thefootprint of a global !"# system in the day to day decision making of managers both at amanufacturing site level and at )ead3uarterslevel. Although the !"# implementation resulted inma'orimprovements in data integrity at an operational level , resulting in improved visibility ofcosts and traceability of transactions for head office, many of the benefits associated %ithe6ploiting the information thus collected have been compromised by the need to rely on non4integrated tools for certain specific functions. 5hus, for decision making purposes, managersmust still do%nload data to spreadsheets, %here they are manipulated and combined %ith datafrom other, non4integrated systems. 5hus, this paper e6amines the role of !"# systems insupporting management activity in a manufacturing environment, highlighting the gap bet%eenmanagement performance and the informational and decisional support provided by the !"#.

e!"ords: !"#, decision making, data integrity, organisational goals
mailto:[email protected]:[email protected]:[email protected]:[email protected]


2/13

Introduction

In streamlining the administration of day to day manufacturing activity, !"# systems areimplemented %ith the aim of reducing the potential for error in information handling in theprocurement, production and distribution processes. 7y incorporating efficient data capturetechnology %ith po%erful net%orking capacity, the vital transactional information is captured atsource by users as an integral part of their %ork processes, and then made availableinstantaneously for management 8alakota and "obinson, /00+$. 5his theoretical ideal makes theconcept of !"# an enormously attractive one for large firms faced %ith on4going struggles toreduce costs and report detailed results to shareholders in a timely fashion )olsapple and ena,/00/$.

)o%ever, the time saved by the organisation in automating administrative processes does notnecessarily imply faster decision processes or better decisions Chang and 9able, /00/$.&anagement decisions are generally based on the ability to compare actualdata often providedby the !"# system$ %ithplanneddata fre3uently managed in less integrated systems, andsub'ect to much fluctuation$. 5he !"# system is a vital link in this process of managementcontrol, being the uni3ue source of the :as4is; situation for procurement, inventory, customerorders and fulfilment "o%e, +$. Crucially, as a centralised repository of company data, bothsite and head3uarter level management have e3ual access to transaction status and history.

)o%ever, there appear to be several constraints concerning the value of !"# systems tomanagers in e6ploiting these systems to their full potential.

Firstly, %hen selecting and implementing an !"# system, companies are looking at a good fit %ithoperations, not %ith the more abstract decision processes. econdly, management is highlyconcerned %ith the efficient use of all company resources, including people and manufacturingcapacity. 5he data model of !"# applications is inventory centric, ho%ever, and therefore lacksthe scope to be able to support managers in decisions that involve trading off the costs related tothese resources. Finally, an !"# implementation implies a certain number of assumptions aboutthe company and ho% it operates at a specific point in time $ suggests, it is important in looking at the impact of information systems ondecision making to differentiate bet%een the decision process itself and the information re3uiredto support that process. It %as Ackoff?s contention, %ell before the age of global !"# systems,that most managers suffer not from a lack of relevant information, but rather from an over4abundance of irrelevant information.

9orry cott &orton +>+$ described the characteristics of the information re3uired byoperational, management and strategic levels in the firm as significantly different. perationalcontrol activities re3uire information that is detailed, real4time and based on the actual use ofinternal resources. &anagerial control, on the other hand, re3uires more summary information,not necessarily real4time and includes e6ternal sources of information.

&oreover, the :nimbleness; of large integrated systems has been 3uestioned in the light of therapidly changing conte6t of managerial decision processes. /$ felt that the notion ofe6perts designing for a f irm one :completely integrated supersystem; %as :absurd;. Changes in


3/13

available technologies have clearly made this overly ambitious target more realistic less:absurd;B$, but


4/13

How does ERP provide informational support or hindrance! to managers in the differentactivity domains identified in Research "uestion #? How can the value of an ERP systembe represented in the map of managerial decisional activity?

5hese 3uestions %ere investigated during an in4depth field study of a large multinationalmanufacturing organisation. ver an +2 month period, over D0 intervie%s %ere carried out %ithmiddle managers in all functional areas of the firm, covering a number of locations at local and

head3uarters level. 5he ne6t section provides the background to the case studied.

Profile of the case study

5he 8ey #harma Company, 8#C real name %ithheld to allo% more detailed reporting$, is aleading manufacturer of pharmaceutical products, %ith a highly successful product portfolio inconsumer healthcare, prescriptions drugs and vaccines. ith annual sales of nearly G0 billion,and a "< budget of GD billion, 8#C is in a dominant position in its marketplace.

5he manufacturing organisation involves more than /0 autonomous plants %orld%ide. 5heseplants are involved in different stages in the production process. Active ingredients are producedat bulk manufacturing sites, and this material is subse3uently used by multiple formulation sites inthe production of finished product. A vast net%ork of commercial organisations, distribution sitesand rdparty licensees is then responsible for %orld%ide distribution.

&anaging the supply chain to efficiently satisfy demand is e6tremely comple6. ith over 0,000tock 8eeping Units 8U?s$ or lines of product, any one of the =00 sources of demand could beordering 004=00 8U?s each.


5/13

%able &: sources of data 'intervie"s and (uestionnaires)

An original facet of this research method is that intervie%ees themselves %ere asked to definetheir o%n areas of organisational focus goals$ and then describe their related decisionprocesses. 5his ensured a rigorous sampling of the decision processes studied and a tightcoupling %ith organisational ob'ectives. 5hus, our study %as pertinent to the organisation as a%hole.

5he ne6t section presents the case study according to the three research 3uestions outlined inthe introduction above.

Findings of the case stud!

Research Question 1:

What are the goals of the organisation? How are these goals operationalised, andtherefore what are managers expected to deliver?

An intense focus on 3uality and the critical :compliance; standards that stem from legislativeconstraints on the pharmaceutical industry as a %hole are ever present concerns for managers inthe company. &ore recent legislation regarding financial practices such as arbanes 6eley isalso becoming onerous in terms of reporting and documentation.

5he primary goal in terms of materials management is to ensure availability of sufficient ra%material to satisfy the planned level of demand for finished goods availability$.


6/13

#roduction planning at 8#C Cork also needs to incorporate, in addition to the fluctuating forecastdemand coming via the


7/13

procure the right materials, at the right place, at the right time

use manufacturing resource efficiently manpo%er, utilities, materials, L$

5he key decisions related to these goals are e6plored in the ne6t section.

Research Question 2

What decisions do managers make on a day to day basis, and what are the chief issuesin making these decisions?

5he key management decisions at the 8#C Cork manufacturing plant and at the above siteorganisation at head3uarters are concerned %ith ho% to orchestrate the supply chain to meetdemand effectively, %hilst keeping costs at a minimum. 5his ob'ective can be broken do%n into Hprinciple 3uestions, %hat to plan, %hat to buy, %hat to produce and %hat to deliver.

What to plan#lanners in the manufacturing plant face conflicting demands in committing their local capacity toproduction plans, in addition to the variability of that demand as described above in research

3uestion +. 5he principle issue for the planners at the bulk manufacturing site is the granularity ofthe individual 8U level orders as they filter back through the supply chain from the markets, asthere is no 7ill of &aterials to relate these 3uantities back to a re3uirement for bulk material.

5hree further factor complicate greatly the life of the planners. Firstly, the :bays; or physicalresources buildings and e3uipment$ that are used to set4up manufacturing processes are notdedicated to one type of process, they are instead shared by different processes. 5hus differentproducts share the same plant capacity. 5he time it takes to configure and de4configuremanufacturing bays for different processes is a matter of days, and therefore must be taken intoaccount in the scheduling of production runs. In general, this %ould mitigate to%ardsconsolidating the demand for a particular active ingredient into one longer campaign, rather thanincurring the setup time inherent in fre3uent de4configuration and re4configuration. 5he necessityto carry out regular maintenance further adds to the comple6ity of scheduling production in anoptimally efficient manner.

econdly, the manufacturing process for any particular active ingredient %ill involve severalstages, and each stage must %ait for the completion of the preceding stage in order to progress.5his implies that planners cannot plan for the production of one batch of active ingredientfinished product$ in one go, each stage in that process %ill be the ob'ect of an individual processorder and material batch record.

Finally, certain ra% materials, solvents in particular, are shared across many different processesand products. 5heir consumption, as %ith the yield of active ingredient, %ill be sub'ect of somevariability. 5his introduces comple6ity particularly for the buyers, discussed in the ne6t section, butalso represents a constraint for production scheduling, as it %ill only make sense to launch acampaign %hen all the necessary resources and materials are kno%n to be available.

5he above site organisation suffers the conse3uences of these constraints in that it operates as aremote valve in the supply chain( in situations %here demand cannot be met by campaigningmanufacturing sites, head3uarters must intervene to come up %ith a solution, and be thearbitrator of priority in the event that production schedules are re3uired to change.

#roduction planners at the local plant level are re3uired to periodically %ork to%ards specific shortterm goals that %ill drive a certain type of decision making. An e6ample %ould be a year4endstock re4evaluation, %here the normal rules for buffer stock may be abandoned in favour offinishing the year %ith the lo%est possible levels of inventory.


8/13

ther short term decisions might involve scenario planning for specific situations. #lanners areasked to evaluate the impact in the event of particular occurrences eg. shortening cycle times,reducing door to door times, implementing 8an7an, stock4out simulation, defect rates, productionbay sharing bet%een / competing products, a po%er outage, increasing effective capacity, L$.

5he re3uirement to carry out :%hat4if; analysis including all the possible parameters for suchscenarios means that planners could not %ork %ithout their spreadsheets, not%ithstanding theimplementation of the costly and all covering !"# application.

What to 1uy

#rocurement of ra% materials using an &"# model is not feasible for / reasons. Firstly, 8#C?sproducts are characterised by the lack of a 7ill of &aterials relating the demand for finished goodsto their constituent active ingredients, and the active ingredients to their constituent ra% materials.econdly, the variability in demand discussed earlier also militates against an &"# type system(

Well . mean you could, if you wanted to, if you used the system to sort of, you can ringthe suppliers today and say . don&t need this load till next Wednesday, and tomorrow orthe day after your plans change again, you&d be bringing it forward again, you know,you&d be continuously changing'

#rocurement participates in the se3uence of events outlined for production planners( customerdemand, coming from the #roduct "evie% &eeting, %ill update the re3uirements for eachproduct. 5his is then fed into the site?s rolling 3uarterly revie% of annual volumes. mall demandshifts may not mean any change to the plan. In any case, updates to the plan drive changes tothe purchasing schedule. 7uyers are therefore sub'ect to the same constraints as productionplanners %ith respect to the vagaries of the demand picture(

the whole element of planning, and the actual adherence to that plan by productionimpacts us to an extent, we&re continuously re2planning, revising plans on an almost dailybasis, certainly weekly basis

Furthermore, actual consumption of ra% materials, and therefore the level of on4hand available,

%ill not be kno%n till the end of any particular campaign, and planners therefore %ill allo% for ayear end campaign to use up the outstanding ra% material.

*ead times for products can be anything up to = months, and some ra% materials aremanufactured specifically for 8#C.

What to make#roduction is organised in shift teams of operators and these teams %ork around the clock, onlystopping for planned maintenance, including an annual plant shutdo%n during the summermonths.

&anufacturing is organised locally in :campaigns;. A campaign is the result of the orchestration ofplant capacity vessels, filters, pumps, L$ %ith labour and materials to a schedule such that a

certain 3uantity of bulk active ingredient %ill be produced. 5his 3uantity is %hat must be set inadvance by the planners. Actual yield may vary from the planned 3uantity for a variety of reasons,related to any of the resources involved. Chemical reactions may affect yields in unforeseeable%ays, breakdo%ns or maintenance issues may affect the physical plant, and sickness M holidaysmay affect the labour availability.

At the end of any given %eek, the production team %ill prepare its %orkload for the follo%ing %eekby printing off all the necessary batch records and related documentation, and they %ill ensurethat the dispensary have physically issued the necessary materials related to those batch


9/13

records. 5he e6istence of the batch records in the !"# system is sufficient for the %arehouse tobe a%are of the re3uirement for material.

Follo%ing the go4live of the ne% !"# system, and despite e6tensive training for all operators, it%as agreed that the :data maintenance; activities described above %ould be carried out byspecific resources %ithin the production teams.

#lanners build their production schedules manually, that is, planned orders are converted intoprocess orders for e6ecution$ only on a %eekly time horion. 5he process orders for any given%eek must take into account the actual production and any related contingencies$ from theprevious %eek. #rocess orders also termed material batch records$ are entered into A# %ithoutthe associated sales orders, planners use the :daily tables; screen to bypass the link tocustomers and drive production activity directly through the %eekly launching of batch processes.

Another variable in the campaigning of production processes is the ability of the customer toabsorb the active ingredient. Although an annual re3uirement might be set in advance, thecustomer can only consume it at a certain rate, therefore bulk sites must bring this into thee3uation %hen laying out the campaign plan.

What to deliver

5he multiple formulations produced in 8#C Cork are shipped to multiple commercialorganisations in countries %here 8#C have a commercial presence$, or to multiple distributionsites %ho sell product in those local markets on 8#C?s behalf$. 5he site might make +04+Dshipments on average per day, and manufactures about the same number of active ingredients.

5he 8#C plant in Cork, Ireland, is part of the global &anufacturing and upply organisation. Corkships H,000 batches of goods per year +,000 batches of finished goods$. >0J of shipments areby air giving rapid door to door delivery times eg. U8 %ithin / days, U %ithin > days$.

Firm orders from customers are received via the


10/13

(he fact of life is that what we have at the moment is like a patchwork 3uilt of all sorts ofsystems, all sorts of structures and all sorts of plans, and what they are basically doing isthat we are going to carry on aligning those and improving them, and improving the flow,streamlining the whole thing'

In adidition, managers have access to the A# 7usiness arehouse system pre4configuredreports from A#$ and financial planning tools AdaytumN$.

In 8#C Cork, it %as recognied early on in the !"# pro'ect that cross4site comparison %as one ofthe key driving forces in rolling out a centralied !"# system to all manufacturing sites. &anagersat head3uarters use 8#I?s to drive performance in the supply chain, but e3ually overallperformance of the net%ork is important, rather than performance comparisons bet%een / nodes(

4s a supply chain, what&s their performance to their customers, rather than talking aboutan individual site&s performance to their customers

5hus, although the !"# system is used to drive local inventory control and financial reporting,that is, as an integration tool at a local level, it has not been used to integrate bet%een head officeand sites. 5his seems regrettable given the scope for improvement in this area.

*ocal planners at 8#C Cork carry out their planning on their personal spreadsheets. =0J of the%orkload is scenario planning, responding to contingencies and events, both e6ternal andinternal. #rocess orders called &aterial 7atch "ecords$, corresponding to the launch of aparticular production batch, are launched manually they are created in A# one by one$ basedon the spreadsheet4based plans. hen a batch has run, and there are redundant process orders,these can hang around in the system until someone specifically deletes them.

In summary, the implementation of !"# for 8#C Cork %as not an option, it %as an imposedcorporate solution to address potential compliance risks and there %as only limited scope to %orkon a good fit bet%een local needs and the application configuration. n the other hand, the moveto a ne% integrated system has had its advantages(

we have better traceability on our customers, we have less write2offs' We have morecontrol over manufacturing

In addition, the !"# system has revealed :stress points; in the organiation %here the process isin need of optimisation(

but ERP is only highlighting what is wrong with our current system, there is nothingwrong with ERP, it&s /ust saying, look, the reason this is difficult is because you are notdoing it right anyway, it&s showing up issues'

+essons from the case

5he enhanced data integrity brought by the !"# has improved management decision making in /

%ays. Firstly it has removed the time %asting arguments about data integrity, and secondly it hasgiven managers more time to focus on more value4adding activities, such as investigating thecauses of variances in operational performance.

It is clear that systems cannot automate the managerial decisions relating to variance analysis fora manufacturing plant, but if, on the other hand, they can provide a clear picture of actual vs planperformance, across a number of operational metrics, plus the ability to look at historic actual vsplan performance, plus the usability and sophistication for managers to create their o%n :%hat4if;scenarios, then it could be argued that the return on systems investment can be 3uantified in


11/13

terms of improved decision making speed and accuracy being the most important decisionattributes impacted$. 5his seems to concord %ith )olsapple and ena?s /00/$ survey findings.

)o%ever, there appear to be several barriers to site level managers e6ploiting the richness ofinformation held %ithin the !"# system(

*ack of skills both a detailed kno%ledge of the business processes and !"#$

*ack of access to data access to production data is more or less proscribed$

Incompatibility of transaction data %ith aggregate management 3ueries

#lant level users are left on their o%n to master the information held in the !"# system, and yetvery fe% have the re3uisite level of access to be able to e6plore the possibilities of using the datato support management 3uestions. tandard reports are more or less ignored, by both users andI5, and there seems to be a commonly held belief that :if it?s not in a spreadsheet then it?s notacceptable;, %hich is parado6ical in an !"#4enabled business %ith a strong compliancere3uirement.

5he data itself is 'ealously guarded, and at best a select group of users %ill have access to 3ueryrights on a copy of the production database. In 8#C, the number of people %ith access to specific3uite po%erful$ 3uery responsibilities is the strict minimum( / people. 5his is for security and forguaranteeing high availability, but it also has the effect of making it practically impossible for auser to get hold of data in a discovery mode Adam and #omerol, /00>$.

Finally, !"# transactions are coded in a %ay that %ill provide linkages to other elements in thedata model, and also in a %ay that is applicable to the %idest range of businesses. For thesereasons, the ra% data very often does not have the necessary attributes to allo% it to be used toans%er management 3uestions. 5hese additional attributes or metadata$ are usually added in adata %arehouse application, %hich refreshes from the !"# production database at pre4definedintervals eg. once a day$.

o %hat %ould be re3uired is a more fle6ible investigative approach to data analysis, empo%eringmanagers to investigate the root causes of changes to particular patterns or phenomena. 5his isa pure 3uestion of competence, in the same %ay that if an organisation decided to implement a

lean approach to manufacturing, it %ould to recruit or develop the skills in4house eg( =igma,etc$ to lead the initiative.

It %ould certainly appear that the comple6 structure of !"# discourages managers fromdeveloping the skills to run 3ueries and establish root cause of symptoms that are e6perienced.5his is particularly true in the area of costs, %here !"# systems are capable of providing greatdetail on current and historic trends, assuming a thorough understanding of the data structureused inn the application?s database.

At 8#C, 8#I?s are the means for driving local plant performance, and management are veryfocused on the 3uantifiable goals that they are given. here the data comes from to drive these8#I?s is immaterial to managers, they simply re3uire data that they can trust, such that they canbe made a%are early of any potential variances such that root causes can be investigated and

action taken before the issue is picked up at an above4site level.

)ere is a key contribution of !"# applications( in the 3uantitative analysis of %ork processes%hich they facilitate and in the cultural change they engender. 5hus, !"# could help drive theimplementation of 8#I?s. )o%ever, every time there is an interface to another system, andcertainly %hen there is manual manipulation of the data to generate the 8#I, the process forcalculating and distributing 8#I?s moves a%ay from a :repeatable M sustainable; model, relyinginstead on manual intervention. 5his in turn raises the issues of specialied resources, skills andlatency, an important one %hen processes face time constraints %hich is fre3uent in moderndays supply chains$.


12/13

5his research highlights the gap bet%een the information re3uired to put together these metricsand the information available automatically in the !"# system. 5he !"# system represents an:actual; picture of the use of corporate resources in terms of inventory, cash and other operationalmeasures. )o%ever, :planned; elements and greater representational fle6ibility are re3uired inorder to make these metrics more meaningful for management decision making.

Further research

Further research should be oriented to%ards frame%orks for modelling and actively managing thegap bet%een business processes and the enterprise system, such that ade3uate resources canbe deployed to address the gap before the benefits of a centralised I5 infrastructure are frittereda%ay in a blossoming of local data %arehouse solutions and personal spreadsheets.

5his research should address the means the organiation needs to put in place to plan for theeventual divergence of !"# systems from reality, and to mitigate against this gap %ith skills andtools. A non trivial starting point %ould be to design a frame%ork for the tools, skills and resourcescapable of reporting on the gap, such that the gradual tendency to%ards system obsolescence isvisible, and can be rectified on the fly. Arguably, if !"# applications are not gear to%ards suchevolutionary patterns, the task of organiational actors %ill continue to be very harduous.

References

Ackoff, ". *. +=>$. &anagement &isinformation ystems. 5anagement $cience, 1ol. +H, Ko. H,+H>4+D=

Adam, Frdric and #omerol, Oean Charles /00>$ +, Fall$. A Frame%ork for &anagement Informationystems. $loan 5anagement Review, 1ol. + Issue +$, pp DD4=2

)eili, O. and 1inck,


13/13

*ee, R. and *ee, O. /000$ :An !"# implementation case study from a kno%ledge transferperspective;, Oournal of Information 5echnology, +D, pp. /2+4/22.

&arkus, &. *., A6line, ., #etrie,

Documents

Carton Adam ECIME 07