BI Publisher (Good)

8/8/2019 BI Publisher (Good)

1/48

A UK Oracle User Group publication

AN INDEPENDENT PUBLICATION NOT AFFILIATED WITH ORACLE CORPORATION

Issue 34 | Summer 2008 | 10.00 | www.ukoug.org

OracleScene

Building a Benefits-driven Business Case

PL/SQL Code: Auto-generation and Obfuscation

BI PublisherSeven Small Steps for a Consultant,One Giant Leap for Userkind


2/48


3/48

to Issue 34 of Oracle Scene!

Thank you all for your positive feedback on

the continued changes to Oracle Scene asalways, we are listening to your commentsand suggestions and will continue to developthe magazine in line with our readersrequirements, (I should add that the newContents page is one of my particularfavourites).

Whats in OSThis edition we have a great range of articles,as always.

I am really pleased to have an unusual, but key,B&M article to start with Building a

Benefits-driven Business Case. This is an area of constant struggle in any IT organi-sation how do you really define the true benefits in a business case, and how doyou measure these at the conclusion of the project? Derek Hancock takes us throughthe details, giving an excellent framework for those embarking on business cases.

Regular contributor Satnam Brar then takes us through Training for ERP profession-als in The Training Game always a contentious topic for any manager, and talksthrough the shifts in attitudes to training, and what that might mean in 2008.

Simon Tomey made me smile with his BI Publisher article, with a scarily realisticstoryline for our fictional BI hero! Another smile was generated by Tracey Bleakleysattention-grabbing How Excel Can Kill your system in 5 years!, where she tacklesthe bane of every ERP professionals life.

For our Siebel users, a very helpful article, outlining The Top Ten Problems ofSiebeland How to Avoid Them by Duncan Scattergood offers some practical advicefor new and existing users, whilst for everyone there is the second in a series of UPKarticles from Larmer Brown, with Considerations for Planning Content Development.

Moving into our Technical arena, long time contributor Lakshman Bulusu partnerswith Bulusu Rama to provide the interestingly titled PL/SQL Code: Auto-generationand Obfuscation (the latter word providing much debate among the editorial team,resolved only by the Oxford English Dictionary), followed by our Top Tips fromfavourites Tim Onions and Jonathan Lewis, covering Performance and Data Cluster.

In the Community News area, we have an interview with Lawrence Clark (JDEdwards Director, Oracle), and UKOUG continues its celebration of 25 yearsserving the Oracle community.

Last, but by no means least, a wonderful And Finally from our Apps deputy editor,Gio, with special reference to me via Hamlets upgrade to R12 dilemma (I have tosay, I do notice a similarity between Hamlet and my users, particularly in his inabilityto get to the point!!).

Digital Oracle Scene

You will all know by now that the first digital edition of Oracle Scene was recentlylaunched as a direct result of feedback from UKOUG members. The digital version isexactly the same as the print edition, but offers additional benefits by enablinginstant access, being easy to read, navigate and share, containing active links tofurther information and being searchable.

The feedback received following the launch of the first edition has been overwhelm-

ingly positive. We will be developing the digital edition further over time, so watchthis space to be the first to find out about any enhancements to this new offering.

We hope you enjoy your digital edition of Oracle Scene. If you have any comments,please send them to [email protected]

Ophelia Dodds, [email protected]

Welcome

OracleScene Issue 34 Summer 2008 03

Oracle Scene Editorial Team

Editor

Ophelia Dodds, [email protected] Editor, ApplicationsGiovanni DAlessio, [email protected]

Deputy Editor, TechnologyNeil Jarvis, [email protected]

Communications ExecutiveBrigit Wells, [email protected]

Advertising & SubscriptionsFor advertising/subscription information,seewww.ukoug.org or [email protected]

UKOUG Directors

ChairmanRonan Miles, [email protected]+44 (0) 7917 025 435

Deputy ChairDebra Lilley, [email protected]

DirectorPeter Phillips, [email protected]

DirectorPeter Robson, [email protected]

DirectorCarl Dudley, [email protected]

DirectorJonathan Lewis, [email protected]

Director

Jane McCullough, [email protected]

DirectorAnne Power, [email protected]

DirectorRoy Bailey, [email protected]

Director PeopleSoftDavid Kurtz, [email protected]

Director JD EdwardsDavid Rowntree, [email protected]

DirectorTracey Bleakley, [email protected]

UKOUG Office

UK Oracle User Group, User Group House,591-593 Kingston Road, Wimbledon,London, SW20 8SA.

Tel: +44 (0) 20 8545 9670Fax: +44 (0) 870 9000 335Email: [email protected]

Web: www.ukoug.org

Production

Designed and produced byAnne KotanTel: +44 (0) 1273 833 923Email: [email protected]

Printed byWoking Print & Publicity

Tel: +44 (0) 1483 884 884

Web: www.wokingprint.com

EDITORIAL


4/484 A UK Oracle User Group publication

03 Editorial

Business & Management

05 Building a Benefits-driven Business Case

Derek Hancock, Whitehouse Consultants Ltd

10 The Training GameSatnam Brar, Maximus

15 How Excel Can Kill Your System in 5 Years!

Tracey Bleakley, Edenbrook

Applications

12 BI Publisher: Seven Small Steps for a Consultant,

One Giant Leap for Userkind

Simon Tomey, BeLife Ltd

19 The Top Ten Problems in Siebel Projectsand How to Avoid Them

Duncan Scattergood, Customer Systems plc

23 Oracle User Productivity Kit (UPK) Considerations forPlanning Content Development

Peter McClintock, Larmer Brown Consulting

Technical

26 PL/SQL Code: Auto-generation and Obfuscation

Lakshman Bulusu, Genex Technologies inc,

& Bulusu Rama, Orissa Engineering College, Bhubaneswar, India

Top Tips

18 Top Tips Part One

Tim Onions, TOdC

37 Top Tips Part Two: Good Performance, Bad Performance and

Statistics Level ALL

Tim Onions, TodC

38 Data Cluster

Jonathan Lewis, JL Computer Consultancy

Blogspot

39 Attaining Approval for a Services Oriented Architecture Project

Mark Jermolenko

09 Oracle News10g Upgrade Companion

40 UKOUG NewsHappy birthday to you, happy

Simon Corbett, Jargon PR

42 Community NewsInterview with Lawrence Clark, JD Edwards Director, Oracle

PeopleSoft Conference Call for Papers

Debras Diary

44 Calendar of Events

45 Event NewsUKOUG Back to Basics Database Administration Event A Review

OUG Ireland 2008

46 And Finally

OracleScene UK Oracle User Group Ltd

The views stated in Oracle Scene are the viewsof the author and not those of the UK OracleUser Group Ltd. We do not make any warrantyfor the accuracy of any published informationand the UK Oracle User Group will assume noresponsibility or liability regarding the use ofsuch information.

All articles are published on the understandingthat copyright remains with the individualauthors. The UK Oracle User Group reserves theright, however, to reproduce and article, in wholeor in part, in any other user group publication.

The reproduction of this publication by any thirdparty, in whole or in part, is strictly prohibitedwithout the express written consent of the UKOracle User Group.

Oracle is a registered trademark of Oracle

Corporation and/or its affiliates, used underlicense. This publication is an independentpublication not affiliated or otherwise associatedwith Oracle Corporation. The opinions, state-ments, positions and views stated herein arethose of the author(s) or publisher and are notintended to be the opinions, statements,positions, or views of Oracle Corporation.

Contents


5/48

This article argues that a comprehensivebusiness case should be a detailed cost/benefit analysis showing the investmentimpact of the project and setting out thebusiness benefits and how they are to beachieved.

In the past, as far as many people wereconcerned, producing a business casemeant setting out the reasons for doingsomething, stating the potential effects ifit was not done (sometimes these were thesame) and itemising the costs involved.

Given that this approach is no longersufficient (if it ever was), the purposeof this article is to outline what, in thecurrent business environment, is needed ina comprehensive business case and to offerguidance on how to prepare and achievethe best results.

What is a business case?

According to the website of the Office ofGovernment Commerce, The businesscase documents information necessary tosupport a series of decisions. These deci-

sions, over time, increasingly commit anorganisation to the achievement of the out-comes orbenefits possible as a result of a par-ticular business change. (Italics are mine.)

This is interesting because it talks about aseries of decisions, that over time,increasingly commit an organisation to theachievement of the outcomes of aparticular business change. In other

words, the business case should be writtennot to seek a single all-or-nothing, yes/nodecision to go ahead but to support aprocess of gradually increasing commit-ment over time. It follows that, at anytime, the decision could be to discontinue.It emphasises not the action to be taken,but the benefits possible or results of theaction. This is a sound approach and

enables the company to take account ofchanging circumstances rather than fixingthe case at a moment in time.

A good business case helps the organisationto determine exactly what it wantsto achieve and how it will do so, ratherthan just being used to obtain funding.It should support sound investmentdecisions that are based on strategic fit,adequate options explored, affordability,improving efficiency, achievability andappropriate commercial arrangements thatare value for money, informed by more

reliable estimates.

Obviously the series of decisions, theperiod of time and the size of the business


Building a Benefits-drivenBusiness Case

by Derek Hancock, Whitehouse Consultants Ltd

These days it has become an accepted requirement that all new projectsshould be justified with a comprehensive business case. There aregood reasons for this. Everyone has heard horror stories about IT projectfailures: massive overspend, systems way overdue, absence of benefits,poor user acceptance. It is not the purpose of this article to discuss indetail all the reasons why projects fail, but some of the reasons areconnected to the subject matter and those will be examined.

case should be in proportion to the scale ofthe project in terms of cost, duration orimpact on the business.

Irrespective of size, any business caseshould contain information coveringfive key aspects:

Strategic fit

Options appraisal

Commercial aspects

Affordability

Achievability

Again according to the OGC, a concise,well-written business case should alloworganisations to:

Make the right investment decision

Deliver the changes

Realise the benefits of those investments

Once again, the realisation of the benefits isa key point.

So, the purpose of a business case is:

To document the justification for theundertaking of the project, assessed onthe estimated costs of development andimplementation and the risks versus theanticipated business benefits and savingsto be gained.

To gain Management commitmentand approval for investment inbusiness change, through rationale forthe investment.

To provide a framework for planning andmanagement of the business change.

To enable you to monitor and measure therealisation of the benefits of the project.

We can see from this summary that thebusiness case is a touchstone throughout

the life of the project. It documents thejustification, it gains commitment, it pro-vides a framework for project planning andit monitors the realisation of the benefitsonce the project is live and the system is inproduction.

A good business case helps theorganisation to determine exactlywhat it wants to achieve and how it

will do so, rather than just beingused to obtain funding.



BU

ILDINGABENEFITS-DRIVENBUS

INESSCASE

Why produce a business case?

Justifying the investment Once upon atime, IT investment was seen as a goodthing per se, but in this highly competitiveage, as with any change in the business,investment in IT has to be seen to offerbenefits to the business: results such asreduced transaction costs, improvedcustomer service or increased margins. ITinvestment must compete on an equal foot-

ing with any other business investment.

Reducing the risk of failure Due to thehigh profile of many projects and theassociated bad news whenever a failureoccurs, management is much more awareof the potential risks of IT projects, and weall want to avoid being associated with afailed project. If a sound, logical, justifiedcase can be put forward, showing benefitsto the business, it is possible we can reducethe risk or at least put the project on asound footing.

Increasing the likelihood of approval Apart from the above, as IT professionalswe want to increase the chance of a projectbeing approved and thus preserving ourlivelihood, so it makes good sense toproduce a business case anyway.

A project is 60 percent more likely to beapproved with a cost justification andbusiness case.Gartner

More than 82 percent of IT decisionsnow require a cost versus benefitanalysis.

Information Week

Return on investment is king, andprojects with a quick and clear cost

versus benefit are much more likelyto get funding in todays uncertainbusiness climate.The Industry Standard

In all the above statements, the idea ofbenefits comes through, either implicitly ina cost justification, or explicitly.

Another interesting statistic:

In a survey of over 700 sites by CSC,82% of respondents said that their ERPimplementation was successful, but only22% said that they had achieved theanticipated benefits.

There are three questions here.

1 If only 22% could say they achieved

anticipated benefits how could theother 60% know that their project

was successful?

2 How do you measure success if not bybenefits to the business?

3 Why could only 22% say they achievedthe anticipated benefits?

This is where reasons for failed projectsrelate to the subject of this article. Thereare many reasons why a project can fail.

Failure may be a case of poor project gov-ernance (by governance, I mean the wholeproject structure including staffing,scoping, planning, initiation, project man-agement, risk management, quality), or itmay be poor preparation for change withinthe organisation and therefore poor take-upof the system by the users. No matter howgood the system, if the users cant or wontuse it, it is a failure.

In many cases it is that the business casewas not properly prepared. Perhaps it wastoo superficial, or not aligned to thebusiness drivers, or it only considered costsrather than cost versus benefits. Equally, itmay be that the case was well prepared butno monitoring of benefits was carried out.Sometimes, even though a project hasseemed successful it is viewed as unsuccess-ful, perhaps because of poor perception bythe users. In this case, documented resultsthat prove success through comparison tothe identified expected benefits can oftenchange these perceptions. Without abusiness case properly setting out the rea-sons for the change, the strategy to beadopted, the benefits to be achieved, andgaining management buy-in, a project is offto a shaky start.

How should I prepare for thebusiness case?

You will need to do a lot of preparatorywork before you can produce your case.For example, you will need to understandthe decision making process; who isinvolved and what motivates or intereststhem, who has the power and who has theinfluence, what decision-making criteria arethere and what else is going on.

Who expects what?

What is the process for making theinvestment decision?

What are the primary business drivers?

Why do it now?

What are the boundaries? What is thescope of the intended change?

What are the intended outcomes

and benefits?

What options are there for making theintended changes? What products orservices would be required?

Who is the executive sponsor?

Who is involved in the decision-making?

Who cares about what?

What are the decision criteria that willbe used?

Are the changes achievableand affordable?

With what other possible projects areyou competing?

What should be in abusiness case?

A business case typically contains thefollowing sections:

Management summary

Outline scope, reasons, strategic fit

Options appraisal. Benefits expectedand quantified. High level cost/benefitanalysis of (ideally) at least threeoptions for meeting the business need.Include analysis of soft or intangiblebenefits that cannot be quantified infinancial terms.

Recommendation, preferred option withreasons for choice

Costs & timescales. High level plan forachieving the desired outcome, with keymilestones and major dependencies(e.g. interface with other projects).

Without a business case properlysetting out the reasons for the change,the strategy to be adopted, thebenefits to be achieved, and gainingmanagement buy-in, a project is off toa shaky start.


7/48OracleScene Issue 34 Summer 2008 7

Investment appraisal, calculation detail,cash flow, payback, ROI

Risks and mitigation. Contingency plansfor non-delivery

Assumptions

Source information, proof, validation

Benefits realisation plan

Here is the first mention ofhow to realisethe benefits a benefits realisation plan.For many this is a new concept but if youare going to justify a project based on sav-ings, you better be clear how you are goingto achieve them in reality. If done correctly,this should be a great help when the projectgoes live and you want to measure theresults. It may also be the differencebetween a perception of success or failure.

Once you understand the playing fieldand have gathered your information; start

putting your case together.

Get it right!!! Be sure of your facts;check your calculations. Nothing under-mines your presentation like an incorrectcalculation spotted by a director Go away and come back when youhave got it right.

Make it relevant. Gear it to what youknow about the decision makers and thecompany drivers.

Focus on the top five costs, but do not

overlook other costs. Directors will wantthe headline costs but not the minutiae.Aggregate the smaller costs. Beware ofhidden costs.

Provide explanation, document well.It has to be a standalone document thatreaders can follow. Remember, you willhave to come back to it later and under-stand what you did.

Define formulae and variables. Showhow you arrived at results. Do not leavegaps where questions or doubts can arise.

Obtain metrics and calculate. Evaluatethe expected benefits using data from thecompany. Be realistic. Use measures thatare important to the company.

Make metrics easily measurable, ifpossible. When it comes to assessing

whether benefits have been achieved youwill want to be able to do so without toomuch trouble.

Its OK to estimate. In some cases youwill have to, but be realistic and explainyour logic.

Include assumptions. Ensure peoplecan understand how you arrived ata conclusion.

Determine proof needed and provideevidence. Back up your assumptions andestimates with relevant evidence, usingindustry data if possible and anecdotalevidence as well. E.g. Company Bincreased their throughput by 15% whenthey implemented this system.

How big is a business case?

This may seem an odd question, but manypeople have not been exposed to a seriousbusiness case exercise so have no feel forappropriate size. It is no good trying to

justify a multi-million pound project withone page of A4; conversely, 20 pages maybe overkill for a new server, so the size isimportant.

The size of the business case should be pro-portionate to the size of the projectand its impact on the business.(5 15 pages excluding appendices).

The management summary should beabout 10% of the document.

For small projects, you may only need toproduce the business case in one stage,

with confirmation of prices, fundingavailability and other relevant material.

For larger projects, you should consider amulti-stage approach:

Preliminary business case (or Strategicoutline case) to confirm strategic fit and

business need, typically no more than oneor two pages.

Outline business case indicativeassumptions to support the preferred wayforward (including procurement strategy,

where applicable), variable lengthdepending on the scale of the project.

Full business case validated assump-tions to support the investment decision,

variable length depending on the scale ofthe project. The level of detail required ateach stage depends on organisational

standards and the scale or complexity ofthe project.

The full business case should include abenefits realisation plan.

To give you an idea of the size of theexercise, it is recommended that you shouldspend up to 3% of the projects potentialinvestment on building the business case.(So, on a half million pound project, thats

15000 or 15 30 days depending onyour costing.)

Documenting the businessprocesses that will be affected

One useful exercise that can be carried outfor a relatively modest outlay is to map thebusiness processes operating in the affectedareas. This serves several purposes:

It encapsulates the scope of theproject for which you are developingthe business case.

It provides a graphical description of thescope for inclusion in the business case.

It enables you to understand how theprocesses work.

BUSINESS&MANAGEMENT

The level of detail required at eachstage depends on organisationalstandards and the scale or complexityof the project.



BU

ILDINGABENEFITS-DRIVENBUS

INESSCASE

It enables you to capture the knowledgethat normally resides in peoples heads.

It allows you to see processes thatare broken.

It provides a sound basis for processimprovement when the project getsunder way.

It forms the basis of your testing and

training materials.

It assists with benefits realisation byallowing you to monitor the improvedprocesses.

Find a specialised tool that enables you tomaintain the process maps together with allthe associated information and data on aserver, so that you can easily maintain themas the processes change. Word andPowerpoint are frequently used for this sortof thing, but the results are static and tend

to get filed away and forgotten.

Find a consultancy company who,equipped with and trained in the use ofthat tool, have the skills to help you carryout the initial exercise of identifying anddocumenting the processes.

Involve the users. With the right tool andconsultants, you can carry out interactive

workshops with the users and get theirinvolvement and buy-in. This gives you animproved understanding of the processes,leading to a more thoroughly researchedbusiness case.

Producing the results

Before you can produce your businesscase, you need to gather and record all theinformation, carry out the calculations andanalyse the results.

The key part of the business case is theinvestment analysis. This is the foundationfor your conclusions and their justification.Different companies use different measuresand some measures mean different thingsto different people, so make sure you know

what are the preferred measures in yourcompany. As with many things in life, it isnot always simply what the numbers say,but a combination of factors that decidethe best course of action; however, makesure the numbers back you up.

Identify and describe the benefits thatthe proposed project should provide tothe company.

Record intangibles.

Identify the metrics to be used thosemeasures that are relevant to the projectand important to the business.

Quantify the metrics in the current

situation using company data wherepossible, or industry standard data if not.

Document your reasoning, yourassumptions, the costs, the metrics,the calculations, the evidence.

Estimate the improvements based onwhat you know about the new system.E.g. reduces time to process a claim fromthree minutes to two. Try not to over- orunder-estimate.

Quantify the metrics in the new situation

based on your estimated improvements.

Calculate the benefits.

Compute final results (IRR, NPV, ROI,payback period, whichever is used inyour company).

When you have produced the numbers,you will then be able to derive yourconclusions and recommendations.

Evaluate each benefit in turn.

Identify themes. Several minor,apparently unrelated benefits may add upto a particularly relevant one.

Check for compelling benefits fordifferent levels of the organisation.

A percentage increase on the return oncapital employed may not mean much tothe Procurement Manager but it might tothe CFO.

Match results to acceptance criteria.Obvious really!

Include charts & diagrams. Some ofyour audience may not have time toread the verbiage but may take in agraph at a glance. You know what apicture is worth.

Determine recommendations. Analysethe numbers, decide which appears tobe the best option(s) and make sureyour decision is a result of reason, notemotion. By this I mean, dont workbackwards to justify a pre-conceivedconclusion.

Use a worksheet layout to present theresults, it imposes order.

Keep it simple one or two pages any more and people will lose interest.

Presenting the business case

When you have pulled it all together,analysed the results and drawn yourconclusions, ask yourself the followingquestions:

Is the business need clearly stated?

Have the benefits been clearly identified?

Are the benefits consistent with theorganisations strategy?

Is it clear how the benefits willbe realised?

Is it clear what will define a successfuloutcome?

Where there is external procurement isthe sourcing clear?

Is it clear why this is the preferredsourcing option?

Are the risks faced by the projectexplicitly stated?

Are the plans for addressing those risksexplicitly stated?

Is it clear what the preferred option isand why it is preferred?

All of this sounds like it is a pretty labourintensive and manual exercise. This is trueand I would suggest that for a large exer-cise you put a team together to share theload. Some people are better than others at

writing, some are better at numbers andsome like doing the investigation but notthe analysis. If the exercise is going toinvolve 3% of project expenditure on alarge project, it needs a leader and it needsto be planned. Its execution will also verylikely have to tie in with board meetings.

One way of mitigating the effort is to finda supplier with a tool that is designed toassist with producing the investmentanalysis. Whilst you could do it yourself inExcel, a purpose built tool that allows youto record information about the factorsor benefits, hold related metrics andcalculate results, i.e. quantify the benefits inmultiple scenarios, can save you a lot oftime and effort.

With the right tool and consultants,you can carry out interactive work-

shops with the users and get theirinvolvement and buy-in.



BUSINESS&MANAGEMENT

With a purpose-built tool you can:

Explain each calculation

Use it at any level

Create as many scenarios as you want

Be optimistic or conservative

Include set-up costs and running costs

Include intangibles

Refer back to it

Repeat the exercise with different values

Show savings over a selected numberof years

Show IRR, NPV, ROI, cashflow &payback period

Provide reports and charts

Use it to monitor results

ORACLENEWS

Conclusion

For IT projects of any size, a business caseis a must. You must be clear about the busi-ness reasons for undertaking the projectand about how it will benefit the business.

The investment analysis is a key part of thecase and should show quantified businessbenefits versus costs. The understanding ofhow the decisions will be made, who willmake them and with what other require-ments for funding you are competing is

as important as the business case itself.When you put all this together to producea benefits-driven business case, yourproject will stand a better chance of beingapproved and of being successful.

For more information about building abusiness case, refer to the web site of theOffice of Government Commerce:http://www.ogc.gov.uk/delivery_lifecycle_business_case_management_.asp

With acknowledgement to the Office ofGovernment Commerce.

About the AuthorDerek Hancockhas spent 23years in theApplicationsSoftwarebusiness, joiningMSA in 1985.Following seven

years in sales andservices with Marcam, he joinedJD Edwards as a Client Managerin 1997, staying there until thecompletion of the Oracle take-over inMarch 2005. Whilst at JD Edwards hespent three years running a CustomerOutreach program aimed at helpingcustomers get more value from theirsoftware. He has worked in theJD Edwards marketplace ever since,spending time with Differentia andCincom Systems. Now withWhitehouse Consultants, the onlyOracle Certified Advantage Partnerfor JD Edwards in the UK, he stillconcentrates on working with theexisting customer base to ensurethey continue to get value from theirsoftware investment.

The Oracle Upgrade Companion helps you upgrade yourOracle Database from Oracle9i Release 2 to OracleDatabase 10g Release 2 (10.2). The guide is not anautomated tool but instead, provides guidance for pre-upgrade, upgrade, and post-upgrade steps. This docu-ment is continually updated as new information becomesavailable. Please check back prior to your upgrade. The

latest version of the 10g Upgrade Companion (v. 2.00)was released on February 13, 2008.

This current version includes information about:

Changes in the behaviour of Oracle Database 10g Release2 that will affect customers who upgrade

This section documents important changes in the behaviourbetween Oracle Database Release 9.2 and Oracle Database10g Release 2 (10.2). It also focuses on behaviourchanges that require a DBA to make an informed decisionin order to minimise the risks that may be introduced bythe changes.

Best practices for pre-upgrade planning, upgrade opera-tions, and post-upgrade analysis

The Best Practices section is not a replacement forthe Oracle Database Upgrade Guide but rather acompanion document emphasising and elaborating on

recommendations and requirements. Oracle technical staffderived the Best Practices which is an accumulation of realworld knowledge and experience obtained whilst workingwith our customers.

Critical and recommended patches, by platform

This section will help you prepare for an upgrade byproviding a list of patches that are recommended for yourplatform. This list of patches is based on the most commonissues encountered by customers on version 10.2.0.3 ofthe database. Carefully review each patch and only applythe patches specific to the components that are currentlyinstalled on your database.

The document is completed through a reference list aboutrelevant Oracle documentation. It is available in printableand interactive versions.

Please review MetaLink Note 466181.1 for details:https://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=46

6181.1 (MetaLink login required.).

10g Upgrade Companion



Jason Wilcox, an Oracle trainer whofocuses on finance, procurement and HR,argues that the best training comes from acombination of approaches. I think youneed something like 50% classroom based

work, supported by 50% follow-up onspecific issues with small groups andindividuals. he says. If you ask anyone

who has gone through some form of ERPtraining they will always tell you that thesecond element was the most useful, butthats really only because they have alreadyundergone the first part of the process.

In my view its vital to have a clearoverview of how Oracle works before youget down to the nitty-gritty.

Another trainer, Tahir Hussain, whospecialises in Oracle Financials, takes asimilar view. Lets face it, Oracle is a com-plicated product and its users will alwaysneed formal training to leverage maximumbenefit from it. The classroom is the idealplace to deliver that formal training but, inmy view, its important to get the numbersof each teaching group right. Ideally youshould be looking at a group of somewherebetween eight and sixteen, but some organ-isations, particularly in the public sector,cant resist the temptation to cram more into save time or budget. What formal train-ing does very effectively, if you get it right,is to show people where they fit into theOracle process and why their part of it isimportant. Done properly it addresses thehearts and minds aspect that actuallymakes Oracle work.

According to George Harvey, who is bothan organisational development trainer withOpen Limits and a career managementspecialist at Amaze Associates, addressingthis human element is a benefit that oftentakes second place to pure technical train-ing but is essential if IT specialists aregoing to accomplish the British ComputerSocietys stated aim of recognition along-side more established professions such as

law, accountancy and medicine. If IT isto deliver on its real promise then it needsto become more and moreaccessible and relevant tothe needs of the businessesit serves, he says.And this means that ITspecialists need to developand employ a much moresophisticated range ofcommunication and inter-personal skills. However,because of the way that

people in our industrydevelop in the early stagesof their careers, focusingon the key technicalaspects of the profession, teaching theseskills is not necessarily easy. Most ITspecialists are brought up in an environ-ment which is very process driven andtotally logical an environment where thefirst recourse is not to have a conversation.Consequently, the biggest challenge is oftento get them out of their usual way of think-

ing, perhaps by running scenarios whichfocus on experiences away from the ITsector, such as recalling a particularly bador good piece of customer service and thenrelating it back to the problem at hand.

Not surprisingly, therefore, the trainersseem to think that they are definitelyadding value to the industry, but whatof the organisations which actually haveto pay for their work? Phil Wilson,consultancy director at inOApps, an Oracleimplementation and training specialist,

believes that training has a vital part to playin the development mix, and one whichcannot be replaced with simple on the job

learning. There is training being done outthere by trainers who understand the sub-

ject on a theoretical basis, but who have noreal life experience of using these skills inthe workplace. Equally, there are highlytechnical and experienced individuals whostruggle to convey the skills in a structuredand understandable manner. You also see

The Training Game By Satnam Brar, Maximus

The ERP sector and the IT industry as a whole have come a long wayover the past twenty years in their attitudes to formal training.No longer content to rely on the command to RFM or read the, ahem,manual, organisations and individuals alike now spend millions of poundsevery year on an increasingly wide range of training packages that covereverything from open boot camps to the highly bespoke. But howeffective is much of this training and what sort of return on investmentdoes it provide to its purchasers? Does it really make companies and ERPspecialists more productive and, if so, is there a best practice method ofdelivering it?

...its vital to have a clearoverview of how Oracleworks before you get downto the nitty-gritty.


11/48

a lot of on the job training which can oftenlead to people picking up bad habits andacquiring a very blinkered view of what thetechnology is capable of. If you really wantto get it done properly, you have to beprepared to pay for someone who has thetechnical knowledge, hands-on experienceand the ability to communicate this in aneffective manner. Tailoring or customisingthe training to meet your current skilllevels, working environment andrequirements also improves the outcomeand benefits of training and provides best

value for money.

Wilsons view is largely shared by anotherpurchaser of training, Paul Gillott of OracleHRMS implementers, Symatrix.Learning on the job will always be aninvaluable part of developing as an ERPspecialist but it has to be supplemented byformal training delivered by trainers whotruly understand their subject and candraw on solid commercial experience.

Without it you continuously run upagainst the we do it this way, because

weve always done it this way mentality so

you never move forward, never reallyimprove. Getting in professionals whohave real command of their subject, arecompletely up-to-date and who know howto teach may look expensive initially, but itcan be a very worthwhile investment.

Organisations are, of course, only onegroup of purchasers of training. IndividualERP specialists also dig deep into theirpockets to join the growing number ofopen courses on the market. RamnikPattni, an experienced Oracle Applications

Technical consultant, recently spent asignificant amount of his own money onthe five day course Extending and BuildingOA Framework Applications. Whilst thecore development technology and toolsetshave remained more or less the same overthe past eight to ten years, in recent ERP

versions there has been much more of ashift to Java based end-user self-serviceapplications, he says. The technologyshift will become more visible in futurereleases of the Oracle ERP product andI was keen to keep ahead of developmentsrather than trailing along behind them.

At first I thought Id be able to do this onmy own, but its just not possible whenyou have the day-to-day pressures of workso I decided to get some formal training atone of the UK Oracle training centres.It wasnt cheap, but it was a very goodinvestment of money and time andI learned a lot, both from the trainer and

from other people on the course. Aftersuccessfully completing the course Ivebeen able to put the knowledge gained intoaction and develop self-service solutions formy current client. Developing real-worldsolutions gives you greater confidence in

what youve learned and, I think, makes meimmediately more marketable.

But does formal training really make youmore attractive to potential employers?In over ten years as a recruiter in the ERPmarket, I can think of very few occasions

when a client has specifically asked fora particular qualification hands on experi-ence has always been the key criterion inboth permanent and contract roles.

However what training and qualificationsdo supply is a valuable icing on the cake

which can help individuals to stand outfrom the crowd. And they can also haveless predictable but no less valuable bene-fits. When it comes down to hiring, Imless interested in the technical content of atraining course, than I am in the fact thatthe individual has put themselves throughit in the first place, says Paul Gillott.It shows a degree of organisation, com-mitment and self-belief that makes a goodconsultant, but even with all these qualities,no course is going to take the place ofgood, old fashioned coalface experience.

Anecdotal evidence and industry surveysboth suggest that the appetite for trainingis growing, but it will be interesting to seeif it will be sustained if the much heraldedeconomic downturn really does start to biteas 2008 progresses. As Jason Wilcox putsit, Training is always one of the thingsthat gets cuts first on over-runs, whilst

Tahir Hussain admits that for many organi-sations, training is still regarded as a niceto have rather than as an essential.However, George Harvey believes that wemay have entered a new era where ITdirectors have begun to see training, not


BUSINESS&MANAGEMENT

as a rather unfortunate and expensivenecessity, but as a tool to boost their ownprofessional credibility. CIOs and CTOsare very keen these days to demonstratethat their departments are adding real valueto the business in order to stop the trendtowards outsourcing. And training is play-ing an increasingly important role in this asit helps IT professionals to plug directlyinto commercial reality. If organisationscome to see IT specialists as valued advisorsand implementers rather than back-roomtechies then everyone will benefit, and for-

mal training is one very effective way ofmaking it happen.

About the AuthorSatnam Brar isthe founder andmanaging directorof specialistOracle recruitmentconsultancy andOracle partner,Maximus.The firm works

with individuals and organisations onpermanent and contract assignments,

both across UK and in key marketsaround the globe such as the MiddleEast and continental Europe.

He can be contacted [email protected]

Developing real-world solutionsgives you greater confidence in whatyouve learned and, I think, makes

me immediately more marketable.


12/48A UK Oracle User Group publication12

BI PublisherSeven Small Steps for a Consultant,

One Giant Leap for UserkindBy Simon Tomey, BeLife Ltd

Oracle BI (XML) Publisher is a rich new piece of functionality inE-Business Suite Financials and one of the easiest to implement.You can forget the text based reports you have been strugglingwith and delight your users with clean, attractively formatted documentsand spreadsheets.

This is a story (fictional) about how yourorganisation could apply BI Publisher togood effect to its reports. The article gives

an outline as to how you could do the sameand concludes with where to get moreinformation.

Delighting your users withBI Publisher...

Whats the point of paying all this moneyfor a system, if we cant even get output wecan send to our customers?, said Colin thecredit controller.

I had implemented receivables somemonths ago and we used third party soft-

ware to format the text based dunningletters from Oracle. When I designed thecustomer number, it seemed like a goodidea to base it on an existing reference

which gave a length of 12 characters. I hadchecked the length of the field in Oracleand it had worked very well with the restof the team. Only now did I discover thatthe text based output only printed the first11 characters.

The director of marketing is going tomake an issue of this one, continued Colinin his inimical way its KOLRC! (Kicking

off left, right and centre).

I never really liked the idea of using thirdparty software. We had to buy anotherprint server to host it, the set up didnt

really work and no one else knew how toconfigure it. I remembered a conversation

with Oracle support where they men-tioned This thing we have called XMLPublisher . I wonder if I could get that to

work. Its integrated and its supported.

That lunchtime I went tothe library to get somebooks on XML. XML foridiots. Well, I felt like anidiot at that moment; what

was I going to do about thecustomer number on the

dunning letters? My wifeGilly called that evening;the kids are desperate tosee you.

I have to stay a bit later tonight, some-things come up, I replied, distracted bymy thinking, XML is all very interesting;a very flexible, self defining language withtags, but I just dont see how it fits withOracle the files dont even work withouta schema.

Gilly interrupted my thoughts. We need

to decide about the extension, we cant

have three children in one bedroom muchlonger, but the builder said its going tobe expensive.

Yes, of course, we need to do somethingsoon I replied. Goodness if I cant sort thisout, that something soon might be goingdown to the Job Centre.

The next day I went to the finance meeting.Surprisingly for a finance meeting, I actually

looked forward to these meetings. Jack, thehead of finance, and Sarah were enthusiasticusers of Oracle and they always appreciated

what I could do, which made me feel better.I was trying to make sense of the XML userguide, so got to the meeting a little late.

The new girl, Janice, was in full flow...When I worked for the Big Corporation,they had SuperFinanceSoftware and it wasgreat. In my last company, you could down-load everything you wanted from the GLstraight into Excel... I dont understand why

weve got Oracle. The reports dont printproperly and when you copy them into Excelthey go all funny. While the RX reports are

better, it is a hassle to run them and thedescription that we type into the GL gets cutoff at only 15 characters so I cant get all theinformation. Clare chipped in Yes, Ive gota problem with reports too... Marketing needa list of payments by invoice paid, but Idont know how to get this. The meeting

was getting worse; it was not what I waslooking forward to, to cheer me up...

Later that afternoon, Jack took me to oneside. Jack was very good with people andknew how to encourage others when they

were struggling. Perhaps he had news on mypromotion and I could tell Gilly to go ahead

with the extension to the house.

The meeting was gettingworse; it was not what I was

looking forward to...


13/48

Michael, Ive had a chat with John.Not only was John the Finance Director,but he had influence at GroupHeadquarters. It seems a number of usershave been going over my head and com-

plaining about the issues with the Oraclereports... John has given me a deadline ofthe end of the month to get this sorted, orhe will be looking to outsource the Oraclesupport. We need to deliver something forhim, or Im afraid were going to have toput you on the re-deployment register,

Jack continued. Im a bit worried abouthow hard youve been working recently.Im supposed to be going to the UKOracle User Group Financials SIGtomorrow, but perhaps you want to taketomorrow off to attend in my place.

The rest of the conversation disappearedinto a mist as I tried to maintain my com-posure. Well at least a day out of the office

will be a good chance to think. Perhaps Illget some sympathy from others at the SIG unlike my users, they understand the dif-ficulty of the standard Oracle reports.Perhaps Ill get some leads on other

jobs going.

Financials; Special; Interest: theyrewords that dont go together easily...I joked with the taxi driver as he droppedme off. My interest was piqued by one of

the agenda items, Transforming yourreports with BI (XML) Publisher possibly the richest and easiest to imple-ment new piece of functionality. I won-dered, could this help me?

The presenter seemed to have great enthu-siasm for the subject and opened by sayingUsing BI (XML) Publisher isnt easy, butits easier than it appears. He asked for ashow of hands Who is actually usingBI (XML) Publisher? (very few hands)...And who thinks theyd like to imple-ment it but have been put off because theydont know how to start? (over 50 peopleput their hands up).

What I hope youll get today is enoughknowledge to be able to get BI (XML)Publisher working on at least some of yourGL reports, but Ill also be usingReceivables customer open balance letters

as an example.

I jotted down notes from the presentationas follows:

The Seven Steps toTransforming your Reportswith BI (XML) Publisher.

Step 1 Choose a report and savethe output

Choose your first report, run it and savethe text based output somewhere. Make a

note of the report code (see the log file).For example, customer open balance letterin receivables (ARCOBXL).

Step 2 Change the output formatto XML

Log in to systems administrator,concurrent programs, define and changethe output format from text to XML.

Step 3 Rerun the report and savethe XML file

Go back into the responsibility as before(e.g. Receivables Manager) and re-run thereport. The report will now be produced as

an XML file. Open the output and save itas an XML file somewhere where you canfind it again (tip: put the report code in thefile name to make it easier to find).

Step 4 Create an RTF template

Use MS Word (or similar program that cancreate RTF) to create a template. Its easiest

with the BI Publisher add-in, but you cando it manually.

1. Set up your template so that it looks likethe output you want. Include formattingand logos, watermarks, signatures, etc.

2. Put something in which will remind youwhere you want items to be filled withlive data. E.g. put address line1 whereyou want address line1.

3. Open the XML output you saved in step3 and put the tag names into the docu-ment where you want the live data to be.Surround the tag names by e.g.

.

4. Save the file as RTF (not *.doc!)

Step 5 Link the Oracle output toa data definition

Login to XML Publisher administratorand create a data definition (home, datadefinitions, create). Fill out the boxesmarked by asterisks and save the XMLoutput file (saved earlier) as preview data.Chose a sensible name (free text field),type the report code into code (this must

be exactly as per the report), application isyour module (e.g. receivables) and a startdate of today or earlier.

Step 6 Link the Oracle output, thedata definition and your RTFtemplate to a XML template

Stay in XML Publisher administrator andcreate a template (home, templates, create).Fill out the boxes marked by asterisks andsave the RTF file (created earlier) as file.Choose a sensible name (free text field),type the report code into code again (this

must be exactly as per the report and astyped into the data template), application isyour module (e.g. receivables) and a start


TECHNICAL



date of today or earlier, language isEnglish, territory is United Kingdom(assuming you are in the UK and commu-nicate in English). Click apply and waitfor the message saying that your templatehas been created. Preview what you havedone by clicking the preview icon.

Step 7 Run the report

Go back into the previous responsibility

(e.g. receivables manager) and runthe report. The output should now beproduced as per your template (and as perthe preview). If you get a message sayingunable to find the published output forthis request and that no output existsfor the request nnnnnn, dont panic its just that this report needs to be pub-lished via the two step process. (See myslides (number 87) or see Metalink NoteID: 364547.1).

I set aside the next day to see if I couldreplicate the presentation on my test

environment. It took me most of the

day to get the hang of it, but once I haddone that I tried the same principles ondunning letters.

These dunning letters are great, saidColin once I showed him the test output.When can we implement to live?

Not only were the letters able to display thefull content of the fields, but the lettersappeared much clearer and include logos,tables and signatures. I used the standard

dunning letter output, so the text could bechanged using the normal maintenancescreen, and the formatting is easily modi-fied via the RTF template. Best of all, theycan be produced as PDF and emailed to amailing house to print and send.

Encouraged by the success with thedunning letters, I set GL account analysisreport to run with the Oracle pre-loadedtemplate. Instead of having to copy pagi-nated text into Excel and fiddle with it toclean it up, users could now paste the GLaccount analysis directly into Excel. Best of

all, the full field content was displayed, so

BIPUBLISHER

:SEVENSMALLSTEPSFORACO

NSULTANT,ONEGIANTLEAPFOR

USERKIND

For more details on the seven steps to setting up your reports in XML publisher youcan download a recent presentation from UKOUGs January E-Business SuiteFinancials SIG. See http://www.ukoug.org/calendar/show_presentation.jsp?id=8167

or http://www.belife.co.uk/OracleR12_XML_Publisher.html (username OracleScenepassword Summertime).

Note: BI publisher used to be called XML publisher and is part of the functionalityavailable from the whole Oracle BI Enterprise package. You do not need the wholeBI package to transform your Oracle Financials reports, as the requisite parts areintegrated into Oracle Financials (and other applications).

About the AuthorSimon Tomeyis a charteredaccountant andOracle Functionalconsultant.Simon has signifi-cant experience ofOracle BI publisherand is an enthusias-

tic advocate of its practical application.He would be delighted to hear fromanyone with questions.

[email protected]

whatever was entered into the journalwould appear directly in the Excel report.

The next day John called me into hisoffice. Michael, Head Office has heard

what we have been doing with XMLPublisher and is impressed. They are talk-ing about setting up a team to migrate

what you have been doing to all the othercompanies in the group. It seems that themoney which we will save on maintaining

our old Oracle reports will easily fundanother two posts, not to mention thetime we are saving in our division in nothaving to populate our MI reports. Ourcustomers are impressed by the presenta-tion of the documents we send out, theboard members like the reports we givethem and were well ahead on convertingour old reports for the planned R12upgrade. Would you consider leading theteam it will mean more responsibility,but of course that will come with theappropriate rewards.

The End ... (or is it just the beginning?).


15/48

For those companies already in thissituation, there is an easier way to fix thisthan to re-implement every time a combination of change management,benefits realisation, and project planningare needed to rectify the systems problems.

These three strands are all interconnectedand the timings are critical. This article

will suggest a methodology for gettingback on track.

When spreadsheets are dangerous

Excel is a fantastic and invaluable tool forbusiness. It is easy to use, universallyunderstood, and makes calculations easy todefine and repeat saving lots of time andeffort. This article does not suggest thatExcel should be banned from an Oracleenvironment far from it!

The distinction (and where spreadsheetsbecome dangerous), is where they becomede-facto databases, holding informationoutside of the system, replicated in different

ways in different departments. Despite thewell-known number of errors in the averagespreadsheet, to the user thesesmall databases become theirdata-master. By stealth, thepre-project state of multipledatabases can becomereplicated or even worsened,and through a predictable seriesof steps, trust and use of thesystem can become so damagedthat the only option can seemto be to re-implement.By understanding how thishappens, we can assess if this isreally inevitable, and hopefullyavoid constant and costly reim-plementation cycles.

Can we get back on track, or do we need tore-implement? Well, this will depend onyour individual set of circumstances. Whatis likely, though, is that by understanding

how the situation happened, a moreinformed decision on the way forward canbe reached.

How do the problems start?

The problems can start right from go-live.Users need to pick up a huge amount ofinformation in a very short period of time(usually through training and UAT).

Whilst common processes might beembedded, exception conditions andinfrequently used processes may be quickly

forgotten. Few companies have an ongoingtraining plan post go-live, expecting astable system and user-base to transferknowledge to new hires.

Post go-live there is pressure on the help-desk/support system and users are likely toprefer working things out for themselvesrather than ask the How do I questions,hence workarounds and sub-optimalprocesses begin to emerge simply becauseusers are keen to get on with the job withthe minimum of fuss!

When a project goes live it is very unusualfor all requirements to be met and all bugsto be fixed. Usually there is a prioritisationprocess, with a gap and snag list left over.

Without allocating budget at the start ofthe process to fix these issues, they areoften forgotten. It is very hard to subse-quently raise a business case for gap andsnag projects, as the projected benefitsshould have been realised by the previousproject and the incremental differences arehard to define. It is sensible to set aside anallowance and timescale in the original

budget to rectify these issues, withoutwhich the projected project benefits maynever be realised.

Also, the push to minimise customisationsand take vanilla processes (whilst sensible),can lead to manual processes outside thesystem being required if applied too rigor-ously. Even standard functionality has somegaps (examples including prepayments andaccruals) meaning that most companiesneed some manual processes from day one.Consideration should be given to the costsof maintaining these spreadsheet-based

processes when evaluating the TCO (TotalCost of Ownership) of customisations,especially with SOA (Service Oriented

Architecture) making customisations easierto support.

The problem with workarounds

Implementation projects rarely focus on themanual process workarounds needed wherestandard functionality is not a direct fit.It is unusual for manual processes to bedefined and tested prior to go-live, as thefocus is on the system. Therefore, Excel-based processes are often developed atspeed, and different departments maygenerate different spreadsheets usingdifferent information for the same task.

Already, there is no standardisation.

If it is the case that the manual workaroundis replacing an existing process orspreadsheet, then this will often becomethe standard, with the new system needingto work around the existing Excel report.Immediately the fit will be worse thanbefore, leading to dissatisfaction withthe system.

Even if there is a genuine appetite for asnagging phase, the demand may disappearonce the manual processes are in place.Remember, the users will be more comfort-able with Excel than the new system (theyhave used it for longer!), and will be more


How Excel Can Kill YourSystem in 5 Years!

By Tracey Bleakley, Edenbrook

Does your business rely on multiple spreadsheets to support yourOracle system? Is Oracle no longer the single source of informationfor the business? Do you have days of reconciliations and manual accrualsat month-end? Is reporting an onerous manual Excel driven process?Does the business even trust the Oracle system when so much manualwork needs to take place before any results can be published? How didwe get here?

...trust and use of thesystem can become sodamaged that the onlyoption can seem to beto re-implement...



likely to trust it, despite the averagenumber of errors being proven to be muchhigher. Users can at least see all the dataand they know how it is being manipulated(using a new system often feels as though

data is disappearing into a black hole!)Already this means that reconciliationsbetween system calculations and resultsand those generated from Excel-basedprocesses are needed, taking time andeffort. Users are more likely to trust theresults from Excel and rarely have thesupport or expertise to investigate anydifferences. In addition, if users do nottrust system reports they are more likely togenerate their own in a spreadsheet.

Changing business needs

At some point it is inevitable that thebusiness will begin to change from itsoriginal state when the system wasdesigned. Perhaps this happened beforego-live, perhaps it was several years afterimplementation.

If the business changes are accompanied bythe time and budget to change or updatethe system, then many changes can beaccommodated. New reporting models canbe built in using OFA (Oracle Financial

Analyser) or Hyperion, for example.The Chart of Accounts can be extended ifnecessary by using spare segments (do notbe afraid to use these thats what they arethere for!); additional modules can be usedfor new business areas; new industryspecific systems can be integrated usingSOA; multi-org can be used to divest andintegrate new companies. Also, SOA canbe used to integrate new business front-ends (such as employee portals), so that thebig systems integration work can beaccomplished in a more sensible timeline.

However, the danger is when the changesare undertaken without giving notice to ITor the systems users, or they happen subtlyand incrementally, meaning that the userstry to accommodate the changes usingmore and more manual processes andspreadsheets. By using ITIL processes(change management), a process can be putin place to assess business and systems

changes and provide governance over howchanges and additions should be managed.It is easy for a system to be perceived asinflexible, when it is more the case that dueconsideration has not been given to the

options available, through lack of time,forethought or budget. Again, the businesscase should take into account the real costsof not making the systems changes in termsof the costs of implementing yet moremanual processes.

Another issue is the Fat GL, where thebusiness needs more information and theGL is expanded to hold this at the detailedlevel. By doing this, the system becomesmore inflexible to subsequent changes.

The GL was not intended to hold businessspecific detailed reporting data,

and this is often best held in asubledger or data-warehouse.

System maintenance

System administration is anarea that is often overlookedduring a project. Whilst theremay be some training onsystem administration it mayfocus on the more immediatetasks such as managing respon-sibilities, security, audit, print-ers etc. Often the processesnecessary to keep system data clean andefficient are ignored, such as:

Chart of accounts values maintenance

Closing purchase orders

Maintaining reports

Accuracy of core data (such as suppliers,customers and employees)

Allowing journals generated bysubledgers to be updated in the GeneralLedger

Allowing journals to be entered in theGeneral Ledger using accounts used forsystems generated processes (leading toincreased reconciliations)

HOW

EXCELCANKILLYOURSYSTEMIN5YEARS!

Allowing journals to be entered in theGeneral Ledger that should be processedthrough the subledgers

By controlling these areas from day-one,it is possible to keep data and the system incheck and prevent the system errors andproblems that lead to reconciliationdifferences on the systems side. An exampleis where accounts are created outside theparent, leading to reporting errors. Another

is where two accounts are created indifferent ranges for the same item, leadingto confusion and misleading results.

A third is where multiple accounts are usedfor systems based processes (such asaccruals), leading to manual processes toallocate amounts, lack of drilldown and ahigh reconciliation overhead.

Again, as more and more off-systemreports are generated and morereconciliation processes are necessary,the data starts to become more and morefragmented, stored amongst a number of

spreadsheets and departments.Data becomes duplicated, fragmented andprone to error, making accurate reporting(even across spreadsheets) difficult.

Of course, all these extra manual processesand reports also take time to prepare.

This may result in more people being hired,or simply longer hours, low morale andincreased stress amongst the user base,especially at month-end or during requestsfor ad-hoc management reports.

Training is it really necessary?

At the same time, the lack of ongoingtraining will be starting to bite. The newusers will never have been trained in thenew system and will possibly be enteringdata without knowing why, meaning theyare even less likely to trust the system.

The old users will remember little of theinitial training and if user guides exist, it isunlikely that they will have been kept up todate. The business is now not in a positionto effect any improvements to the systemor processes through lack of knowledge,and the only option may be to bring inexternal help.

A well-trained and updated user base willbe in a position to effect improvements andupdates to the system and processes

...Often the processesnecessary to keepsystem data clean andefficient are ignored...


17/4817OracleScene Issue 34 Summer 2008

providing valuable assistance to the busi-ness. This is one of the key tenets of theshared services concept. When evaluatingtraining costs and the business case fortraining, these benefits should be taken intoaccount.

For the wider user base, it is essential tokeep up-to-date with how to use thereporting tools and self-service functionalityemployed. This not only reduces the load

on the core users and IT, but also sweatsthe asset (i.e. the system) to its fullcapability, and the system then becomes afull business partner.

Also, consider internal super-user groupsand informal networks. Training does nothave to be external or costly. There area range of options, and often using acombination works best. Consider onlinetraining content, especially for inductiontraining and refreshers or high levelintroductions; consider classroom trainingfor detailed topics, using external facilities

to get people out of the office environmentto increase concentration and generate newideas; and use internal training andnetworks to bolster and embed training,knowledge transfer and new ideas.

Without all of this, the data will oftenbe well and truly in silos. The Excelspreadsheets can almost become a rivalsystem, with users claiming to prefer andtrust their spreadsheets. However in realitythey now represent a huge number of data-bases, often using the same data in different

ways and with multiple different values.

This is possibly worse than the pre-implementation situation where multiplelegacy systems may have existed.

At this point, data integrity is completelycompromised.

At this stage, performance will also beaffected and no amount of hard work bythe users at month-end and criticalreporting times can mitigate this.System performance can be compromisedby badly managed data, proliferation ofreports, and lack of training, such as usersrunning reports without a start date or

without using parameters to restrict data.

At the same time, the hardware may bestruggling to cope with the growth in data

volumes that may not have been addressed.This becomes cyclical as users becomefrustrated by poor performance.

On top of this, the manual processes andreconciliations can take days at month-end,and reporting can be a nightmare ofmultiple spreadsheets, calculations andreconciliations. Worse still, any adjustments

may mean starting the whole cumbersomemanual reporting process again.

So what can be done?

At this point, all is not lost. A combinationof user training, data cleansing (treated assystems improvement projects), a hardwarereview and a reporting project(possibly to implement eithernew solutions, or new modelsand training within existingsystems) may be appropriate.It is often easier to build a

business case on thecombination rather than theindividual parts, as all of theseelements combine to give theoutcome bringing back theinitial business benefits of thesystem. As the system is likelyto now be a corporate issue,due to the time taken to compilemonth end and ad-hoc reports, as wellas user complaints, this is likely to be atimely solution.

Left unaddressed, the problems become

even more severe. At some point theaccuracy of reporting will becomecompromised and this cannot beovercome by simply adding more manualprocessing time. The same problems willmultiply, and if this is combined withbusiness change, the issues can begin tolook insurmountable.

However, there is still a solution. In addi-tion to the changes advised earlier, it maybe possible to effect a number of systemschanges to bring the system in-line withthe business. Again, it is recommendedthat the business case be built on the

combination rather than the individualparts, as all of these elements combine togive the outcome. If this is the case, thenit is recommended that the individualcomponents be implemented as aprogramme so that each element is fullyimplemented at the right time, in order tomaximise the benefit.

And if we do nothing?

If the problem is not addressed, then trustin the system can become irretrievablycompromised. As the problems escalate,re-implementation may start to seem likethe only option because trust in the systemis gone, and the likelihood of signing offadditional funds to resolve systems issuesis low.

Things can only get worse if not addressed.When standards in data entry are forgotten,then reporting can completely break down.In Accounts Receivable this can mean aninability to produce an aged-debt, inGeneral Ledger it can mean that reports

need to be generated using child values asthe parents are inaccurate opening up theprocess to more errors, and so on.

Performance will continue to be impacted,as will user morale. The system will start tobecome a serious management issue.

So what now?

Recovery is still possible, but must bescoped as an overall programme. Small,tactical one-off projects are unlikely tosucceed as the problem is spread across thesystem and users and management willneed to see quick-wins and constantimprovement.

The question of re-implementation versusimprovement will need to be seriouslyconsidered. If the business has changedsignificantly, if the version of the applica-tion is old and there are serious gapsbetween that and newer versions, if thechart of accounts is no longer valid, if theGeneral Ledger is full of detailed obsoletebusiness specific data, if the shape of thecompany has changed (requiring a newoperating structure), or if the hardware isobsolete, then re-implementation may be a

valid option.

If the answer to all of the above is no, thenimprovements as part of a consolidatedprogramme may be a cheaper and more

BUSINESS&MANAGEMENT

As the problems escalate,

re-implementation maystart to seem like theonly option.



pragmatic option (especially if the newlytrained user base can participate in theimprovements).

So in summary, how didwe get here?

By going live without all the requirementsand bugs being addressed and by notrevisiting these later, we deliver anincomplete system that cannot possibly

realise the projected benefits. If the initialtraining is not adequate or not planned asan ongoing effort, then the users will notbe empowered to maximise use of thesystem ongoing. Trained users, good sys-tem administration and governanceprocedures, and a focus (and budget) forconstant improvement, underpinned byITIL, are essential.

This may be expensive but considerthe costs and benefits as opposed to re-implementing every five years and theeffect on the business as the system

diverges in between!

If you think your system needs a re-implementation rather than a series ofimprovements to get back on track, thenconsider the following:

A Chart of Accounts and Calendar fitassessment

Fit between current business processesand system processes

Ability of the system to supportmanagement and statutory reportingrequirements, including consolidations,group reporting and tax

Future business flexibility required

Potential merger or divestment activity,or legal/market changes

The current level of customisations

The gap between the current and thelatest software release and functionality

Organisational changes such as sharedservices

And if we do re-implement,what should we do differentlythis time?

Identify target benefits upfront andcompare to progress, even after go-live

Budget and plan for ongoing trainingand system development

Identify an internal systems governance/administration role and processes

Review latest best-practice thinking andrevise business processes

Take advantage of new functionalityavailable (such as SOA) to increaseflexibility

Decouple data where possible: thinGeneral Ledger, standard reporting lines,correct use of subledgers etc

Archive old data where possible andimplement an archiving strategy

Take the opportunity to update supportarrangements

Spend time on conference room pilots,

UAT and training

Implement regular system reviews andprocesses this is a living corporate asset!

About the AuthorTracey Bleakley isHead of ChangeManagement atEdenbrook anda director ofUKOUG. She has12 years experi-ence of designing,implementing and

managing business transformationprojects supported by Oracle solutionsacross a range of sectors includingtelco-media, financial services, retailand professional services.

HOW

EXCELCANKILLYOURSYSTEMIN5YEARS!

Before I dive into the meat of this issues tip(s), I need to clear upsomething from last time and I also have one or two follow-ups toslip in. I am indebted to Des Browning who queried something Iimplied in my tip about getting the database server hostname:

I've never used a login.sql because when you switchbetween databases with logging it isn't re-executed to

refresh the prompt. Has this changed? I didn't see areference to it in your article.

This is a brilliant point which brings up two further (small) tips.

As Des correctly suggests, yes, the way SQL*Plus handles theLOGIN.SQL script has changed LOGIN.SQL is re-run, from the10g version if my memory does not fail me, whenever you usethe CONNECT command.

However, where the 10g version of SQL*Plus is not used, what Ihave done for years (well, since Oracle7) is to create and use ascript called CONNECT.SQL, place it somewhere withinSQL_PATH and call this rather than use the CONNECT commanddirectly (i.e. always use @CONNECT when changing users/ses-sions). My CONNECT.SQL simply issues a CONNECT commandbased on any command line parameter supplied and then re-runsLOGIN.SQL. So when in SQL*Plus you can use:

@CONNECT user/password@alias

and be confident that your prompt and variables are set appropri-ately for the database to which you are logged in. Granted, youneed to remember to add the @ symbol (or START command)

otherwise things will be out-of-date. I have found, however, thatafter a very short while this gets engrained in your psyche andrather than forgetting to use it, you will find yourself forgettingNOT to use it when you are on a client (or database server) otherthan your normal means of running SQL*Plus. Note, it does alsopay to just run @LOGIN manually if at any point you are unsure ofyour prompt.

A few months ago, a work colleague recommended to me a blogthat had just been started up by Oracles own OptimizerDevelopment Group. To my shame I only just very recently found

time to take a look but, when I did, I found the content there(and on the associated Oracle Database Real-World Performance& Scalability blog) very interesting and valuable indeed. Whatparticularly caught my eye was that at least two of the articlesprovide follow-on material for recent Oracle Scene articles: onecovers V$SQL_PLAN/DBMS_XPLAN and hence relates to an arti-cle I wrote a little while back; another deals with ANSI syntaxouter joins and hence builds on Tony Haslers article in the lastissue of Oracle Scene. So, if you have a technical, quizzical mindand want to hear direct from those in Oracle who should know(and from reading these blogs most definitely do know) thencheck-out:

http://structureddata.org/about/and

http://optimizermagic.blogspot.com

Top Tips Part Two can be found on page 37.

Top Tips: Part One By Tim Onions TOdC Limited



Siebel is an immensely powerful productwith an extremely wide set of functionalcapabilities. However, it is the way in

which the product is implemented thatdetermines whether the anticipated benefitsare delivered.

A project can fail for a variety of reasons.During our ten years working with Siebel

we have been called on to rescue manyprojects that have gone off track, and time

and time again the cause can be traced to abasic set of problems.

By bringing together the ten reasons why aSiebel project is most likely to fail, thisarticle examines the factors that separatesuccess from failure. We explore potentialpitfalls and how these can be avoided, andconsider how to bring problematic projectsback on track.

The Top 10

1 Unrealistic short term expectations

One reason why a Siebel project may beperceived to fail is that there is an unrealis-tic expectation of how quickly the benefits

will be seen.

Merely installing Siebel and making itavailable to users will not transform theeffectiveness of an organisation. Instead,realising the full capabilities of the applica-tion takes a lot of hard graft from technicaland business oriented teams. Time andeffort are essential, and not all the benefits

will be delivered at once.

For example, if a key benefit is moreaccurate forecasting of sales opportunities,that benefit will not be realised until theorganisations opportunities have been

actively managed through the new processand there is consistency in the data.

To avoid setting unrealistic expectations,care must be taken to understand at theoutset:

The actual IT costs of the project

The anticipated benefits and when theyare expected to be delivered

The prerequisite process steps, bothtechnical and business, to obtainingthe benefits

The necessary commitment by the busi-ness of time and resources to the project

If adequate consideration isgiven to these points

when developing aproject plan, it ispossible to celebratemilestones along the

way to the ultimategoal, rather thanbecoming disillu-sioned by how much

work has been doneand yet how far therestill seems to be to go.

2 Insufficient Siebel knowledge indesign team

Experts in the relevant business processes,or experienced business analysts who have

worked on other applications, can play vital

roles in the analysis and design team, butthey must be complemented by real Siebelexperts. A lack of Siebel expertise is likelyto give rise to the following scenarios:

Processes are designed based on hownon-Siebel applications work, leadingto many screens being required forsomething which could be achieved

with one or two button clicks. This leadsto significant and often unnecessary re-engineering of Siebel.

Particular ways of working are imposedon the application, when Siebel alreadysupports the exact requirement in aslightly different way. Recreating func-tionality that exists in the core product isunnecessary and expensive, both in termsof initial development, and of ongoingmaintenance, enhancements andupgrades.

The system fails to meet all the require-ments because it is perceived that certainthings cannot be achieved in Siebel when

in fact, with appropriate skill, they arefairly small configuration tasks.

The importance of engaging real Siebelexperts at the early stages of a project,

with both functional and technical knowl-edge about what is and is not achievable and how best to achieve it cannot beoverstated.

3 Separate technical team

In a traditional IT project, a businessanalyst speaks to users in order to gatherrequirements, before documenting themand passing them to a technical designer.

The designer translates them into atechnical design which is then implementedby a programmer before the system istested and deployed. This approach resultsin a large gap between the users with the

The Top Ten Problemsin Siebel Projects ...

By Duncan Scattergood, Customer Systems plc

O rganisations often spend huge sums on buying and implementingSiebel, and many see a fantastic return on their investment - increas-ing sales, improving customer service and reducing internal costs.However, others struggle and often become disillusioned with the wholeprocess after years of effort. Why do Siebel projects go off track, and howcan organisations ensure that theirs will be a success?

... and How to Avoid Them

Time and effort are essential,and not all the benefits will bedelivered at once.


20/48

original business problem, and the techni-cal team charged with solving it. In a moreextreme version, the development workmay even be carried out off-shore.

This model may still have a place in somesituations but it is not appropriate in aSiebel project. Siebel is different because askilled configuration expert can bend thecore application to meet a particularrequirement very quickly. This presents an

opportunity to run projects in a way thatgreatly increases the probability of thedelivered system meeting users needs,

while lowering project costs by reducingthe project team size.

The ideal model for a Siebel project is asmall, expert team where each individualis responsible for all activities relating toa particular functional area, fromrequirements gathering, through design,development and testing, to deployment.In larger projects, there may well be people

who specialise in design, build or test, orpeople who are involved in only one area,but the core team should remain through-out the whole project.

4 Insufficient user involvement

Ultimately, a Siebel implementation is abusiness project that must deliver businessbenefits. The project is for end users andtheir managers, so it is critical that they arefully engaged throughout the project toensure that the final system delivers whatthey need.

Often, other responsibilities mean thatusers have only limited time to give to theproject, but if they are only prepared tocommit very small quantities of time oreven no time at all then our experience isthat a project will not be as successful as itshould be.

Typically, the most important stages forsignificant user involvement are:

In the design phase, including workshopsand prototyping to verify that the userinterface will support what they need todo. At this stage, it is also important torun through each of the processes thatusers will perform, to check that nothingis missing.

In the user acceptance test phase, tocheck that what was finally built reallydoes do the job.

In the deployment phase, to assistin rolling the system and associatedprocess changes out into the wideruser community.

In addition, continual (if less frequent)involvement throughout the project,

including build, is advisable. This involve-ment really does help users to have a strongsense of ownership of the project, as well asallowing frequent checks that the project ison track, avoiding surprises during useracceptance testing.

When planning user involvement, a keyconsideration is who to involve. This variesfrom organisation to organisation, and alsodepends on the task to be performed. Of allthe user-centric tasks, design workshops arenormally the most critical to a projects suc-cess. Senior management often do not have

enough time to do the activity justice andmay not actually understand the finer detailof the processes that users will perform ona day to day basis. On the other hand, run-ning workshops with low level users canresult in the collection of information thatreflects the details of the current systemrather than the changes that senior man-agement are trying to implement.

Therefore, the best candidates for thesetypes of activities are often found at teamleader or middle management level.

5 Poorly thought-out integration

Many Siebel implementations involve com-plex integration with other applications.Even if each individual application thatsupports a business process is well designedand robust, if the interfaces between thoseapplications are flawed, then the resultingoverall system will not be solid.

Again, expert Siebel knowledge is essentialin this scenario. If the team handling theintegration has little Siebel knowledge, thefollowing often occur:

Inappropriate split of tasks betweenapplications: for example, somethingthat could be easily achieved in Siebel iscustomised into another applicationbecause no one knew better.

20 A UK Oracle User Group publication

Unresolved inconsistencies betweendata models: for example, a contact canhave many addresses in Siebel, but onlyone is supported in a particular legacyapplication.

Incorrect interface type selection: thewrong decision is taken over whichintegration technology (EIM, eAI, WebServices etc) to use. Integration routinesmay also be poorly developed, since this

is primarily a Siebel configuration jobrequiring expert Siebel skills.

Even when Siebel expertise is engagedin the design and build of integration,organisations often choose to subdividethe Siebel team along the line between con-figuration and integration. This is a mistakebecause decisions made at the design stageabout how the user interface should workcan affect the format, style and frequency ofintegration and choices made in integra-tion (often forced by legacy applications)can affect how information should be pre-

sented to users. Therefore, the two aspectsare by no means independent, and as aminimum, the team building the coreconfiguration should be expert in the issuessurrounding integration and vice-versa.

Integration must be considered in thecontext of the overall application design,and not dealt with as an afterthought.Of course, individuals with high levelsof expertise in the other applicationsconcerned should also be involved, andcollaboration is the key to successfulintegration projects.

6 The difference between sales andservice projects

A Siebel

Documents

BI Publisher (Good)