maximize chances bi success

8/10/2019 Maximize Chances BI Success

1/21Forrester Research, Inc., 60 Acorn Park Drive, Cambridge, MA 02140 USA

Tel: +1 617.613.6000 | Fax: +1 617.613.5000 | www.forrester.com

Best Practices: Maximize Your ChancesOf Business Intelligence Successby Martha Bennett and Boris Evelson, July 24, 2013

For: ApplicationDevelopment& DeliveryProfessionals

KEY TAKEAWAYS

For BI Initiatives To Succeed, They Need To Align With Business GoalsIn most cases, starting a BI initiative thats not clearly linked to specic KPIs or business

goals or aimed at addressing a specic pain point is sowing the seeds o eventual projectailure. Tis is why its crucial to have the processes and mechanisms in place to establishwhat the desired outcome would be and to start rom there.

Agility Is The Most Important Characteristic Of All ProcessesProcesses and rameworks are vital project ingredients, and BI initiatives are noexception. What makes BI projects different is the need or agility. But this doesnt meanhaving to adopt an Agile methodology as such. Te important point is not to stick totraditional water all methodologies, which are unsuited to the majority o BI projects.

Change Management And Training Must Be An Integral Part Of Any BI ProjectIntroducing new BI capabilities typically involves change. Te implications are ofen ar-reaching, requiring people to work in different ways; in many cases, they also result inprocess and organizational change. Failure to address change management and trainingright rom the beginning may put the success o the entire BI initiative at risk.
http://www.forrester.com/


2/21

2013, Forrester Research, Inc. All rights reserved. Unauthorized reproduction is strictly prohibited. Information is based on best availableresources. Opinions reect judgment at the time and are subject to change. Forrester , Technographics , Forrester Wave, RoleView, TechRadar,and Total Economic Impact are trademarks of Forrester Research, Inc. All other trademarks are the property of their respective companies. Topurchase reprints of this document, please email [email protected] . For additional information, go to www.forrester.com .

FOR APPLICATION DEVELOPMENT & DELIVERY PROFESSIONALS

WHY READ THIS REPORT

What can you do to make sure that your business intelligence (BI) initiative doesnt end up on the scrapheap o ailed projects? Seeking answers to this question isnt unique to BI projects but there is an addedsense o urgency in the BI context, given that BI-related endeavors are typically difficult to get off theground; there are horror stories aplenty o big-ticket investments that have not yielded the desired benet.Tis report provides application development and delivery (AD&D) pro essionals and their peers with asummary o the most important ingredients that contribute to BI success, as well as a review o the mosttypical pit alls that can derail BI projects and how to avoid them.

Table Of Contents

Treat Business Intelligence As A CoreCapability Of Your Business

Successful BI Projects Are Business Projects

Best Practice No. 1: Put The Business IntoBusiness Intelligence

Best Practice No. 2: Be Agile And Aim To

Deliver Self-ServiceBest Practice No. 3: Put A Solid GovernanceFoundation In Place

Best Practice No. 4: Select The Right Tools

Best Practice No. 5: Seek Help When YouNeed It

Best Practice No. 6: Make ChangeManagement And Training Integral

WHAT IT MEANS

Times Are Changing; Your Approach To BIMust Change Along With Them

Notes & Resources

Forrester interviewed business intelligenceleaders and experts at six user companies ofdifferent sizes and geographic distributions.We also spoke with BI practice leads andsenior consultants at Column Technologies,HCL, IBM, Infosys, KPMG, PwC, TataConsultancy Services, and Tian Marketing.

Related Research DocumentsEstablish Ongoing Processes To Assess YourBusiness Intelligence MaturityJanuary 4, 2013

Craft Your Future State BI Reference ArchitectureNovember 1, 2012

The Future Of BIJanuary 10, 2012

Best Practices: Maximize Your Chances Of BusinessIntelligence SuccessProcesses: The Business Intelligence Playbookby Martha Bennett and Boris Evelsonwith Holger Kisker, Ph.D. , Michele Goetz , and Sarah Bookstein

2

12

5

13

7

15

10

16

18

JULY 24, 2013
http://www.forrester.com/go?objectid=RES84621http://www.forrester.com/go?objectid=RES84621http://www.forrester.com/go?objectid=RES82341http://www.forrester.com/go?objectid=RES82341http://www.forrester.com/go?objectid=RES61273http://www.forrester.com/go?objectid=BIO733http://www.forrester.com/go?objectid=BIO1737http://www.forrester.com/go?objectid=BIO2566http://www.forrester.com/go?objectid=BIO5224http://www.forrester.com/go?objectid=BIO5224http://www.forrester.com/go?objectid=BIO2566http://www.forrester.com/go?objectid=BIO1737http://www.forrester.com/go?objectid=BIO733http://www.forrester.com/go?objectid=RES61273http://www.forrester.com/go?objectid=RES82341http://www.forrester.com/go?objectid=RES82341http://www.forrester.com/go?objectid=RES84621http://www.forrester.com/go?objectid=RES84621http://www.forrester.com/


3/21


Best Practices: Maximize Your Chances Of Business Intelligence Success 2

2013, Forrester Research, Inc. Reproduction Prohibited July 24, 2013

TREAT BUSINESS INTELLIGENCE AS A CORE CAPABILITY OF YOUR BUSINESS

Tere is no doubt that companies understand the importance o business intelligence (BI) tosupporting the efficient and effective running o the business. Whether its optimizing processes,improving customer service, increasing the accuracy o marketing initiatives, breaking into newmarkets, or seeking ways to get ahead o the competition, rms recognize that getting the rightdata to the right person at the right time is a key prerequisite to business success. Ever-increasing

volumes o internal and external data, shortening business cycles, and challenging economicconditions make it imperative to deliver a complete, up-to-date view o the business quickly andefficiently. It is also clear that companies recognize the value o data and analytics when it comes toimproving decisions and outcomes, as BI has been a top implementation priority or organizations

or a number o years now (see Figure 1 and see Figure 2).

Figure 1 Companies Recognize Te Importance O Data And Analytics

Source: Forrester Research, Inc.95961

Implement a bring-your-own device strategy 13% 4%

Create a comprehensive mobile andtablet strategy for employees 17% 6%

Shift spending from core systems to applicationsdriving engagement with customers 18% 6%

Create a comprehensive cloud strategy 18% 7%

Develop smart product APIs that improveproduct and service capabilities 19% 7%

Create a comprehensive mobile and tabletstrategy for customers or business partners 19% 8%

Cut overall IT costs due to economic conditions 20% 7%

Reorganize or retrain IT to better align withbusiness outcomes and drive innovation 26% 7%

Help the organization better manage andintegrate its partners and suppliers 27% 8%

Improve IT budget performance 27% 8%

Develop new skills to better support emergingtechnologies and business innovation

28% 9%

Improve IT project delivery performance 32% 11%

Improve the use of data and analytics toimprove business decisions and outcomes 37% 18%

High priority Critical priority

Base: 3,616 business decision-makers from rms with 100 or more employees(multiple responses accepted)

Source: Forrsights Business Decision-Makers Survey, Q4 2012

Which of the following technology initiatives will youor your team ask IT to prioritize over the next 12 months?


4/21




Figure 2 Business Intelligence Is Te op Implementation Priority


Note: We rst included industry-specic software in the Q4 2008 survey; we rst included nance and

accounting in the Q4 2009 survey.Source: Enterprise And SMB Software Survey, North America And Europe, Q3 2007, Q4 2008, and Q4 2009;Forrsights Software Survey, Q4 2010, Q4 2011, and Q4 2012Source: May 27, 2011, Forrsights: The Software Market In Transformation, 2011 And Beyond Forrester report

7

6

5

4

3

2

1 Business intelligence

Customerrelationshipmanagement

Collaboration software

Finance andaccounting

Industry-specicsoftware

Enterprise resourceplanning

Human capitalmanagement

2008(N = 1,158)

2009(N = 1,021)

2010(N = 455)

2011(N = 913)

2012(N = 1,092)

2013(N = 1,631)

The top seven software applications in rms adoption plans by year

Business Intelligence Success Remains Elusive For Many

Recognizing the importance o data and analytics is one thing. Actually putting in place theprocesses and tools required to deliver data and analytics in the most efficient and appropriate wayto meet the needs o business decision-makers is a different matter: Te majority o organizationscontinue to struggle to get the desired value out o their BI investments (see Figure 3). Business

users complaints about their ability to access the data they need to support a decision when theyneed it rom their enterprise BI applications include:

Insufficient or inaccessible data. Tis is one o the most requent gripes about BIimplementations. I know we have this data; why cant I get to it? is a re rain that echoesthrough many offices, in particular when new visualization and analytics tools have shown whatmight be possible.


5/21




Excessive or wrong data. Users may have a variety o reports and/or dashboards available tothem, but they are either being overwhelmed by the sheer quantity o data or eel that the datathey get isnt quite what they need. A marketing specialist in a global consumer goods rm toldus that she has access to all o the data she could ever wish or but its delivered in different

ormats by different systems, and she has no idea how it all ts together.

Untrustworthy data. Tis is a common concern in organizations where a new set o reportsor analytics capabilities has replaced tried and trusted data sets. Te efficiency improvementsthat rms hope to get rom new BI tools are easily negated. Tis ofen happens when businesspro essionals either dont use the new system or seek to replicate results using their tried-and-true spreadsheet models.

Unacceptably long turnaround times or new reports or other BI capabilities. It doesnt

matter i the delay is days, weeks, or months; i a business pro essionals need has changed by thetime a new report or dashboard appears, its o little value to her.

BI tools that arent right or the job. Business pro essionals ofen grumble that they simplycant do anything with the in ormation presented to them. Examples include static reportsthat dont allow the user to drill down into the detail, let alone run what-i scenarios; report

ormats that make in ormation difficult to digest, such as endless tables o numbers instead o visualizations; and tools that only a hand ul o extensively trained specialists can use.

Tere are many reasons or this long list o grumbles rom business users, and why so many othem create their own analytics environments using spreadsheets or procure data discovery and

visualization tools without involvement rom their colleagues in I . Te most important are:

Lack o governance and business ownership. Governance rameworks are necessary to ensurethat you use data to support business decisions in the best possible way and without compliancebreaches, regardless o whether its how you capture, store, analyze, or use data; who has access todata and or what purpose; how you dene metrics and key per ormance indicators (KPIs); orwhat reports and dashboards you make available and to whom. While I has responsibilities as thecustodian o data and governance processes, it can only discharge them effectively i the businessside takes ownership o the decision on what data it needs and what rules apply to that data.

Overreliance on traditional I approaches. Water all sofware development methodologiesarent suited to the majority o BI projects; many organizations spend more time and effort ongetting data into a data warehouse (DW) than making it available or use. Uncompromisingstances when it comes to a single version o the truth; overly controlling approaches, such asinsisting that all data used in analytics has to go into the DW rst, ollowing lengthy duediligence processes; and not supporting end user sel -service are also key inhibitors o BI success.


6/21




Not having a knowledge management and collaboration environment. Many organizationshave more than one BI plat orm, and each BI environment may have thousands o reports.Ten theres all o the data that resides outside o the existing BI environment. Tis makes it very difficult or business pro essionals to nd what they need, whether its a report, data, or acolleague with relevant expertise.

Figure 3 Companies Struggle o Achieve Teir BI Goals


Base: 603 business intelligence users(multiple responses accepted)

Source: Forrsights BI/Big Data Survey, Q3 2012

Gaining competitive advantage

Expanding the scope of data we leverage to include

more data and more extreme scale

Ensuring compliance and reducing risks

Improving customer interaction and satisfaction

Monitoring and improving businessperformance across departmental silos

Improving and optimizing process performance

Monitoring process performance

Achieving better business transparency

Improving data quality and consistency

Improving business planning

Making better informed business decisions

22% 6%

25% 6%

26% 7%

30% 6%

32% 6%

30% 8%

31% 8%

34% 10%

36% 10%

39% 9%

43% 11%

How successful has your rm been so far in achieving the following outcomes since adopting BI?

5 = Very successful4 = Successful

SUCCESSFUL BI PROJECTS ARE BUSINESS PROJECTS

o understand what distinguishes a success ul BI implementation rom one that struggles or even

ails completely we spoke with 15 BI specialists and carried out extensive secondary research.Our interviews included executives in charge o BI programs in a variety o industry sectors andcompany sizes, as well as BI practice leaders or their equivalent in a number o global consultingrms and smaller specialist providers. We also explore the topic o BI success or ailure in ourongoing dialogue with clients, as well as discussions during con erences and workshops. Whilethe best practices we identied may seem obvious, they are what differentiates those who struggle


7/21




with their BI initiatives rom those who succeed. Methodologies, tools, and processes certainly helpdeliver success ul BI projects, but cant make the difference between ailure and success on their own.BI leaders who want to maximize their chances o success are strongly encouraged to:

Put the business into business intelligence. Ensure that all BI initiatives have a businesssponsor and align with the business strategy, goals, and KPIs. Put the appropriate processes inplace to ensure close collaboration with the relevant business execs and subject-matter experts(SMEs), both throughout a project and afer its completion. Make sure that project success is ashared responsibility.

Be agile and aim to deliver sel -service. Whether or not you ormally adopt Agilemethodologies, the ocus must be on short review cycles and requent deliverables. Developa process that works or you and stick to it agile doesnt mean ad hoc or ree- or-all.Troughout the process, aim to deliver as much BI capability as possible in the orm o sel -service tools.

Establish a solid oundation or your data as well as your BI initiative. Knowing what datayou have and how you capture, store, and process it is just as important as having rules andguidelines in place to ensure that any reporting or analytics unctions yield reliable resultsand dont give rise to compliance issues. O equal importance is having governance principlesand processes or your BI environment itsel . At the same time, its important to ensure thatgovernance rameworks and processes dont become so onerous that they get in the way, andthat you develop them in small steps with tangible deliverables.

Select the most appropriate tool set. ake into account current and uture requirements romthe technical and business perspectives, and dont neglect to assess the current skills base in theI and business departments. Develop a set o criteria based on use cases to help you decidewhen to adhere to corporate standards and when exceptions make sense. Tere is rarely anoccasion when one size ts all.

Seek external help i needed. Understand where your limits are and involve external expertsas required. Tis applies to all stages and aspects o a BI initiative, including requirementsgathering, technology selection and deployment, and change management and training.

Make change management and training an integral part o any BI initiative. Be aware othe nontechnology implications o BI initiatives, whether its replacing a single tool orintroducing a whole new way o doing things. Put appropriate support mechanisms in placeright rom the beginning.


8/21




BEST PRACTICE NO. 1: PUT THE BUSINESS INTO BUSINESS INTELLIGENCE

Tis may sound like a oregone conclusion, but it isnt. Some BI projects ail at an early stage;others go right through to nal delivery and signoff be ore someone declares that they dont meetexpectations. Te most commonly cited reason in all cases is a lack o involvement rom thebusiness, either at the executive or SME level or most ofen both. But its about more than justhaving a business sponsor and a set o requirements its about having the relationships andprocesses in place that ensure collaboration with business execs and SMEs. For BI projects to besuccess ul, it is paramount that:

Business and I sit at the same table right rom the beginning. Tis means having aconstructive working relationship between business and I that revolves around the business,not technology. It also requires a business sponsor whos prepared to champion a BI project atthe executive level and take the lead where necessary to surmount roadblocks along the way.

BI aligns with business strategy, goals, and KPIs. Tis applies to each individual BI projectin the same way that it applies to an organizations overall BI strategy. I you cant express thedesired business outcome o enabling very specic business decisions based on the availabilityo particular data, dont even start the project. Yes, there are cases where business execs orSMEs can only articulate the desired outcome in potentially vague, qualitative terms when theyarent sure yet what kind o results theyll get a classic chicken-and-egg problem and a greatsupporting argument or Agile BI, which enables real-time prototypes and proo s o concept.And even i requests are very specic in terms o what new reports or dashboards are required,but its not an optimal starting point unless you link them to specic business outcomes. akethe discussion right back to the business goal or issue.

A process to ensure that alignment is in place. Whether the process centers on storyboardingtechniques, checklists, scorecard methodologies, or some combination thereo is a matter ochoice. What works best or an individual organization depends on a number o actors, suchas company culture and structure, BI maturity, and existing processes. Most important is theoutcome: mapping what the business is asking or to a well-dened, existing KPI or businessgoal. Tis can include addressing a pain point thats clearly part o the overall business strategy.

Te end o a project isnt the end o the business/I partnership. While BI implementationstypically consist o discrete projects and ongoing support, no rm should ever regard a BIinitiative as completely nished. Requirements or new reports or dashboards are boundto sur ace, usability issues are certain to arise, and new data sources or tools with greatercapabilities are likely to appear the list is endless. o prevent todays BI success rombecoming tomorrows legacy application, business and I executives alike need to have anongoing dialogue, even afer the success ul delivery o a project.

Te business has (or takes) ownership o BI usage. Aligning the BI objectives and prioritieso the business and I organizations is ofen quite challenging. Some o the challenges are dueto disconnects between business and I program drivers or example, the business goal o


9/21




revenue generation and risk mitigation versus the I goal o cost savings, and the businesspriority o selecting the best tool or the job versus the I priority o using enterprise-standardBI tools. Dont ght such potential disconnects leverage them. AD&D pros should do what

they do best: control, manage, and plan, handling the data preparation layers o the BI stack likedata integration, metadata, and data warehousing. Ten business users and business processpros can get more involved and own the data usage layers o BI, writing their own reports, adhoc queries, and dashboards and analyzing the data that has been integrated and standardizedaccording to corporate policies (see Figure 4).

One client, a global manu acturer, recently told Forrester that it staffed its new BI competencycenter entirely with employees who were shadow I BI experts rom various lines o business.Tese experts were superusers who knew exactly where to get the data and how to use it butdid it only or their own workgroups using homegrown desktop applications. Te leadershipchallenged them to put their money where their mouth was, stop complaining about the lacko I support or enterprise BI, come out o the shadows, and become the development andsupport center or enterprise BI. It was and still is a hugely success ul endeavor.

Figure 4 BI Success Depends On Balancing Responsibilities And asks


Communication Collaboration Agile methods and tools Change management

Infrastructure provider Custodian of data and rules Data preparation Technical standards Training and support

Strategy Goals, objectives, and KPIs Data and rules owner Report and dashboard

creation Data usage

BusinessIT


10/21




Pitfalls To Avoid: Taking An IT-Led Approach May Seem Easier

A BI project can still end in ailure even i the business was the driving orce behind it. Te riskis even greater i business executives abdicate their responsibilities to I . From an I perspective,this may eel like the proverbial easy way out: Just ocus on delivering on time and to specication.Ultimately, though, the business will blame I or not delivering the desired benet with the project.Hence, I executives should re use to embark upon, or continue with, a project when any o the

ollowing occur:

Business requirements are not aligned with goals or KPIs. It doesnt matter whether youdene requirements in terms o specic desired outcomes, data items, new reports, dashboards,or tools i the business pro essionals concerned cant link their wish list to the businessobjective or KPI that this new intelligence or capability will support, raise the warning ag. Youmay need your entire port olio o diplomatic skills here, but this is the time when the onus

really is on your colleagues in the business areas concerned to clari y how having certain data orcapabilities is likely to contribute to business success.

Te business sponsor hands over responsibility to I . It happens quite ofen: Business andI specialists work together on requirements; once theyve reached agreement, the businesssponsor signs off on the budget and sometimes also specically on which tools I will deploy.And then theyre gone, expecting I to deliver whats been agreed, on time and within budget.Its debatable whether any technology project can be success ully delivered in this way; BIprojects certainly cant, even when theyre on time and within budget. Needs may change alongthe way; but more importantly, needs ofen only get claried once the project is underway andthe business reviews the interim deliverables.

Strategic and tactical initiatives are out o balance. I typically thinks in strategic terms enterprise DW, a single version o the truth, a centralized plat orm whereas business tendsto take a more tactical approach (I need to measure these KPIs). Unless rms approach thesestrategic and tactical aspects in parallel, rather than sequentially, project ailure is likely.

I takes the BI lead. Tis sounds rather cruel, given that I executives not only have a uniqueinsight into the data a business holds and has access to, but also have an end-to-end view oan organizations processes. And theres nothing wrong with I being the catalyst or a newBI initiative; there are many examples o I introducing the business to the art o the possible.However, i its not possible to convince the business to become an active participant in theproject, such initiatives rarely succeed; they either become bogged down in endless argumentsover scope and budget, or the resulting deliverable isnt quite what the business needed aferall. Teres only one exception to this rule: As one o the systems integrators we interviewedemphasized, there are situations where a purely technical re resh is required be ore I can claimits seat at the business table to discuss the potential or new BI capabilities.


11/21




BEST PRACTICE NO. 2: BE AGILE AND AIM TO DELIVER SELF-SERVICE

Even the best-planned and -supported BI deployment will not achieve the desired result i thechosen development methodology isnt suited to delivering a BI initiative. In particular, traditionalwater all SDLC methodologies are not appropriate or the majority o BI initiatives, which tend torevolve around needs and requirements that typically arent ully dened at the start o the project.o maximize the chances o BI project success:

Use agile methods where possible. Note that this doesnt have to be Agile in the strict senseo ollowing a set methodology such as Scrum or Kanban. For the majority o experts we spokewith, its about being agile and taking an iterative approach, with the emphasis on breaking allelements o a project into the smallest possible chunks, working collaboratively, and reviewingtangible deliverables requently. How it works in practice tends to be quite organization-specic, but its important to note that agility and exibility dont mean doing things in an

ad hoc manner. eams still need to agree on and codi y processes albeit lightly, ratherthan setting them in concrete. Be ore you can break rules in the interests o moving a projectorward, you need to have them in the rst place whether theyre about allocating roles in the

project, setting meeting schedules, having lists o go-to people, or deciding how to documentrequirements and processes. And dont orget about project and program management: Makeuse o the tools and rameworks that your organization is likely already using.1

Put a methodology in place to classi y BI initiatives and process steps. Tere aresome exceptions to the use o agile. Te key ones are major, enterprise-grade BI plat ormdeployments; well-dened and largely xed BI requirements, as or tax and regulatory reporting;and data preparation. As one executive put it: When it comes to data, there are times when Areally must come be ore B, and B be ore C.

Focus on requent deliverables. Frequent (sometimes daily) review sessions involving businessstakeholders and demonstrations o early prototypes to business pro essionals not only serve toshow progress, but also to ensure that you can make adjustments be ore its too late. Breakingprojects into the smallest possible chunks acilitates the delivery o something tangible at theearliest possible stage; this not only helps keep business pro essionals engaged, it typically alsoresults in ideas or uture development.

Consider colocating team members or parts o the project. While virtual teams can be veryeffective, especially with the right support processes and collaboration tools in place, there aretimes when teams can achieve the most effective results most quickly when its people literallywork side by side. Tis applies equally to smaller rms, where its simply a matter o migratingto a different oor or a period o time, as it does to larger rms or when external partners areinvolved, when considerable investment may be required to bring people together. All o the BIspecialists we spoke with had stories o projects that reached an impasse; they nally made therequired breakthrough by colocating the relevant team members or at least as long as it took tosolve the issue.


12/21




Seek to deliver sel -service capabilities whenever possible. Forrester advocates that businessusers build about 80% o reports and dashboards themselves to ensure that they make optimaluse o their organizations BI tools and skills. Provided that an appropriate ramework is inplace to govern the use o tools and data, sel -service helps break the never-ending cycle oun ullled requirements.2 Tat said, not everyone is ready to build their own apps or embarkupon a voyage o data discovery. In one o the organizations we spoke with, it was a major stepto move experienced business pro essionals away rom their existing set o static reportsdelivered in a variety o ormats to a new reporting tool. Providing this set o users with sel -service tools at this stage would have been a step too ar; most likely, they would have rejectedthe new BI environment.

Pitfalls To Avoid: Choosing Too Rigid A Process Or None At All

Regardless o the method and process chosen, the key to success ul BI delivery is nding the rightbalance between control and exibility. Tings go wrong when:

Process adherence is too rigid. Following agreed-upon processes and procedures is necessaryto prevent the kind o chaos and con usion that ultimately lead to project ailure. But always

ollowing methodologies or processes to the letter can also be detrimental. As soon as there aresigns that something isnt right with a project or example, progress is too slow or under-the-radar subprojects start appearing its time to review whether its appropriate to make anexception or tweak the process. Tis applies specically in the context o BI projects, whererequirements may change quickly or only become clear through rapid iteration or trial-and-error approaches.

Tere are no processes in place, or everyones ignoring them. While exibility and agilityare the keys, dont use them as an excuse not to put a set o rules in place that govern how BIinitiatives are run. I such rules are absent or no one ollows them things will go wrong:duplication o effort, missed requirements, use o the wrong data, or ailure to agree on themetrics or a report. Sooner or later, chaos will reign, costs are likely to get out o control, andproject ailure is certain. Te only exception: early-stage, experimental projects involving a smallnumber o people. Te key is to recognize the point at which such projects turn into mainstreaminitiatives requiring a process ramework to ensure their success ul continuation.

Project leaders ail to distinguish between scope creep and adjusting requirements. Byapplying agile methodologies, those involved in a project can make sure that they adjustrequirements as necessary and that the end result is an application thats t or purpose ratherthan an expensive mistake. But the line between rening requirements and scope creep is ofena ne one. Seeing the rst report, dashboard, or application typically gives business users ideas

or more which you should welcome in principle. But make sure that the ocus remains onthe original objective o the project and examine each new idea or request accordingly; i a newrequirement doesnt map 100% to the project goals, it must go into the queue or later.


13/21




BEST PRACTICE NO. 3: PUT A SOLID GOVERNANCE FOUNDATION IN PLACE

In the context o BI, there are two key aspects to governance: data governance and governance othe BI deployment itsel . In order to use data to improve operational efficiency and gaincompetitive advantage, organizations rst need to understand what data they have available rominternal and external sources and the value that this data has in terms o process improvement,decision-making, development o new products and services, and so on. Tey also need to be clearhow they dene each data item and what internal and external rules apply to the capture, storage,and urther processing o such data. BI governance covers rules and processes pertaining to reportcreation, ownership, distribution, and usage, as well as prioritization o must-have and nice-to-have capabilities. Tats why all BI projects including experimental ones ying under the bigdata ag have as their prerequisite:

Te recognition that data management and quality are important. For most I pro essionals,

this is a given; they understand what it takes to provide the required support structures andprocesses and are able to make it happen. Teyre also aware o the risks posed by inadequatedata management mechanisms.

A willingness to address data governance shortcomings and break down silos. Even thoseorganizations with well-established data management programs and data governance structuresofen put them in silos; urthermore, the advent o new (ofen external) in ormation sourcesmeans that data o unknown quality and sometimes doubt ul provenance enters the mix. Tismakes it all the more important to have organizationwide governance structures in place; closecollaboration is needed between those involved in BI projects and those who are responsible ordata governance overall.

An understanding that data and BI governance arent I responsibilities. echnology isessential or putting effective governance mechanisms in place and maintaining them and to helporganizations comply with applicable laws and regulations. But rst and oremost, governance isabout a clear understanding o whos responsible or what, and having processes in place togovern things like permissions and entitlements and making decisions on what metrics andreports the company needs and how it uses them. Tat typically means having the business takeownership to dene the rules and processes pertaining to data as well as BI provision.

Pitfalls To Avoid: Treating Governance As An Afterthought Or Overdoing It

Many organizations that dont already have governance rameworks in place try to skimp on thedata governance step and avoid BI governance altogether. In particular, when it comes to projectsinvolving data discovery or bringing in external in ormation sources, the temptation is just to get onwith it and think about the boring bits later. Te three most common mistakes are:


14/21




Hoping that the issue will just go away. It wont. I anything, itll get worse, because sooner orlater somebody will want, or need, to understand the data that ed the model or algorithm thatbecame the basis o a particular decision. Te old adage garbage in, garbage out continues toapply, except in amplied orm: Big garbage in, even bigger garbage out. Tis can have seriousconsequences in terms o misguided business decisions or compliance breaches. Or, in thewords o a telco exec, I you neglect data quality and governance, youre toast. Equally, withoutBI governance, you may end up with report and dashboard proli eration or nd that the wrongpeople are using the reports, dashboards, and tools or the right people using them but not ortheir intended purposes.

Putting I alone in charge o governance. In many cases, I might actually be quite capableo taking care o it. However, that doesnt mean that the I specialists have the requiredbusiness understanding to make optimal decisions. For example, the risk o compliance ailure

can remain high because I may not understand all o the nuances o local legislation. Or, atthe other extreme, costs can increase unnecessarily because I applies the same principles osecurity and integrity to all data, even where this is not warranted.

Regarding data governance as a necessary evil. For many organizations, data governance isa box-ticking exercise aimed mostly at ensuring regulatory compliance. Tis in turn makes italmost inevitable that the business will see it as a cost and an overhead, rather than an essentialbuilding block o a data-driven business strategy. 3

Being too rigid in the approach to governance. Tere are cases when good enough really is good enough. Agility and exibility are crucial in the governance context as well; too rigid a

governance ramework will not only get in the way, but someone will circumvent it sooner or later. 4

BEST PRACTICE NO. 4: SELECT THE RIGHT TOOLS

echnology alone doesnt deliver a good BI solution, but all BI implementations have technologyat their heart in one orm or other. Selecting the right tool can make the difference between a BIproject that struggles to get off the ground and one that people embrace with open arms. Tis is whyits important to:

Select the most appropriate tools rom a strategic as well as tactical perspective. As is usuallythe case, the 80/20 rule applies. Can the tool do most o what most o its potential users want itto do? How easily can it accommodate potential uture requirements? Who are those users BIpro essionals in the I department, power users on the business side, or occasional users on thebusiness side? Are there real-time (or near real-time) processing requirements? How large and/or diverse are the data sets? Whats the likely li e span o the solution? Does the sofware complywith corporate standards?


15/21




Find the right balance between en orcing standards and permitting exceptions. From acost and support perspective, it makes sense to try and standardize on a single plat orm andas small a tool set as possible. In practice, this is rarely easible, particularly in large, diverseorganizations. Such organizations should ocus on adhering to governance rameworksand inter ace standards; when it comes to the core BI plat orm, strategic considerationsand standards should take the upper hand. As ar as end user tools are concerned, it may beper ectly appropriate to select sofware that may only be in use or a comparatively short periodo time, be ore either it or the need or it has been superseded. Te key is to have processesin place that govern tool selection. Whether decision-making involves checklists or decisionow charts or other methods is a matter o choice, as are the criteria to apply when consideringdeviations rom corporate standards. 5

Emphasize the rationalization o data management plat orms. Concentrate your efforts to

rationalize plat orms, adhere to corporate standards, and save costs on data management, ratherthan on BI plat orms. I all data ows through a single logical pipeline and gets integrated,trans ormed, and cleansed along the way, it will be much easier or you to support multipleBI plat orms. I every BI plat orm accesses the same data sets, inherits the same data catalogs,structures, and other metadata, having different BI user inter aces will matter a lot less.

Pitfalls To Avoid: Allowing Technology Selection By Accident

Having said that technology alone cannot deliver BI success, making the wrong choices is likely tohave a detrimental effect on a project. At best, your costs will go up and youll have to extend yourtime scales; at worst, the entire initiative may ail. More than that, it can make it more difficult to get

uture BI efforts off the ground, as everybody ocuses on the ailure. Te most common mistakes are:

Buying everything rom your incumbent provider by de ault. Tere are many reasons whyit makes sense to acquire additional sofware rom a vendor whose solutions you already use

or other business unctions, such as enterprise resource planning. However, it shouldnt be aoregone conclusion or you to do so, even i the license terms appear more avorable to begin

with. I a tool doesnt provide at least 80% o the unctions you require, or poses serious issues interms o usability or training requirements, its most likely worth looking elsewhere.

Not carrying out technical due diligence. Caveat emptor still applies; beware o vendorsmaking promises. Tis may sound like a rather condescending warning to I pro essionals,but we continue to see companies signing substantial deals without putting vendor claimsto a thorough test. It always gets pain ul when a buyer doesnt properly assess integration orskills requirements; its even worse when, or example, they dont realize that the individualcomponents o a vendors complete BI suite dont all have the same technical heritage, or thatthe so-called sel -service tool can only be used by extensively trained power users.


16/21




Allowing cost considerations to be the prime driver o technical decisions. Sofware is onlyone o the cost elements in a BI initiative, and typically not even the largest but, as its theeasiest to quanti y, its also the easiest to squeeze. Tis is usually a mistake, as hoped- or savingscan pale into insignicance compared with the additional costs incurred by choosing thewrong tool. Such additional costs take many orms: extra training needs, more complex (andhence costly) integration requirements, the need or add-on sofware to overcome unctionalshortcomings or incompatibilities the list is endless.

BEST PRACTICE NO. 5: SEEK HELP WHEN YOU NEED IT

Depending on the type o BI initiative youre embarking upon, you may realize that you need toaugment internal expertise with external resources. Tis realization may dawn at any stage romright at the beginning, when you sense that its difficult to get stakeholders to agree on anything, to

near the end, when it becomes clear that you dont have enough resources to support the rollout othe new tool. External experts can also add value via the methodologies and processes they bringwith them. o maximize the chances o BI project success:

Decide what kind o assistance you need and choose providers accordingly. Many differenttypes o external support can come into play during a BI initiative, whether its a majorcorporate program or a specic project. For example, you may eel that external acilitation ismore likely to ensure alignment with business goals; you would like assistance developing themost appropriate processes to support your BI initiative; your BI experts in I require technicaltraining; or maybe you just need extra resources on a temporary basis to clean up and integratedata. Another common scenario is the lack o in-house expertise in areas such as advanced

analytics or big data technologies. Whatever your need, exercise care when choosing a supplier. 6 Dont just take into account the port olio o offerings an individual provider may bring to thetable make sure that its the right cultural t or your organization. You and your team aregoing to have to work very closely with whomever youve chosen to assist.

Adapt existing vendor selection processes i required. Tis isnt an invitation to rideroughshod over existing corporate processes. However, the hurdles and unexpectedrequirements that can arise during a BI project require a pragmatic and exible approach; i alengthy, bureaucratic process delays partner selection to the degree that the project is at risk, itsprobably time to change or even ignore that process.

Be clear about how you want your external partners to work with you. Be ore starting tonegotiate, its important to decide what you expect external partners to deliver, and how. Do you just need extra bodies temporarily to help with specic tasks? Do you want the external partnerto continue providing support indenitely afer the project goes live, or do you want to makesure that it trans ers skills to your own people during the course o the project?


17/21




Decide i youre ready or BI sofware-as-a-service or even BI business process outsourcing. Several o the organizations we spoke with ound that it was quicker and easier to introducenew BI capabilities by making use o cloud services. ypically used as an adjunct to existingBI environments, the as-a-service approach may not only provide a aster path to answeringbusiness questions and address skills shortages, but also help introduce new ways o thinkingabout BI delivery in organizations that have become set in their ways. Cloud-based services mayalso be more suitable when external data sets are involved.

Pitfalls To Avoid: Abdicating Your Responsibilities

While external partners can make the difference between BI success and BI ailure, there are limitsto what even the best provider can achieve. Youve sown the seeds o ailure i you:

Expect external parties to solve problems on their own. Unless youre working collaborativelywith your provider, its unlikely that youll achieve the desired result. Even with the best skills,no outsider can know your organization, its data, and its processes as well as you do.

Automatically award additional work to the same provider. A company that youve alreadyengaged to upskill your technical staff may not be the best place to acilitate a workshop withyour most senior business leaders to tease out clarication about certain KPIs or tackle thedelicate subject o process changes that the rm may need to make to get the most value out othe BI initiative thats underway.

Abdicate responsibility when using cloud services. Moving data management or analytics to anexternal cloud service provider does not change anything when it comes to taking responsibility

or data and the processes governing how the data is used. Failing to negotiate service-levelagreements that reect your compliance requirements or to make sure that your cloud serviceprovider can track and prove to you that its taken care o all regulatory, data protection, andsecurity requirements can lead to uncom ortable moments with auditors and regulators.

BEST PRACTICE NO. 6: MAKE CHANGE MANAGEMENT AND TRAINING INTEGRAL

Change management and training are critical aspects o any BI initiative. Most importantly, you needto include them right rom the beginning o any program or project. For the best chances o success:

Never underestimate the need or change management. People are the most critical elemento any BI project. Whether its I specialists or business pro essionals, regardless o senioritylevel, i theyre not com ortable with the tools youre introducing or the process changesyoure making, they can make the project ail not through malicious intent, but by simplynot engaging and by sticking to known ways o accomplishing their tasks. Tis is why itsso important to consider change management right rom the start and to put the necessaryprocesses and support structures in place.


18/21




Reect the skill and com ort level o business users when rolling out new BI tools. In someorganizations, business users may be eagerly queuing up to get new tools, dashboards, reports,or data discovery capabilities. Some may already be very com ortable working with BI toolsthemselves, even creating their own apps. But all organizations are different; depending onwhat business pro essionals are used to, they may eye the new with great suspicion. All othese scenarios require tailored approaches that take company culture and established ways oworking into account. For example, when aced with a skeptical user base very much set in itsways, one rm decided to start by simply recreating an existing set o reports in the new BI toolin order to give users the chance to become acquainted with the tool and gain condence in thedata it delivered to them. Once users overcame the initial hurdles and the value and potential othe new tool became clear, the requests or urther unctions started pouring in.

Dont skimp on training and support. For each end user- acing deliverable in your BI project,

assess what support mechanisms may be required, both or existing staff as well as newcomersto the organization or team. Whether you use in ormation observation techniques or ormalusability labs and ocus groups, make sure that you understand where the knowledge andskills gaps are and how best to address them. And dont leave it until the last minute; apart

rom anything else, you may nd that redesigning a report or dashboard may be a bettersolution than writing a 20-page training manual or it. Decide early on what mix o ormaldocumentation, eLearning, instructor-led training, and one-on-one tuition is most appropriate.Some companies also assess the potential ongoing training needs o individuals by analyzinghow they use the tools available to them and offering appropriate hints and tips (or remedialtraining, i required).

Create a knowledge-sharing environment. Te best tips and tricks ofen come romcolleagues whether its business pro essionals or BI specialists on the technology side whohave already solved a particular problem. Rather than risk having people work withsuboptimal solutions, reinvent the wheel, or simply struggle to solve a problem, put the toolsand processes in place to enable people to share their discoveries and seek help rom the mostappropriate expert. Unless this is already in place, it also requires creating a culture thatencourages and supports the sharing o expertise.

Pitfalls To Avoid: Focusing On Technology Development And Rollout

Many BI tools and capabilities ail to gain traction once theyre rolled out, even i the technology isas per ect as can be. ypically, the main reasons or this are:

Completely ignoring the need or change management. Organizations ofen underestimate just how much it takes to get people to change the way they do their jobs. Trow companyculture, politics, and vested interests into the mix and you have a per ect recipe or disaster, inthe shape o accusations that an I project has ailed to deliver yet again.


19/21




reating training as an aferthought rather than putting it together expertly. Ofen, thesame people who developed dashboards or reports or congured new tools are also the onesthe company asks to provide the training. But that doesnt make them experts in instructionaldesign. Also, given that they know their work inside out, theyre not best placed to putthemselves in the shoes o somebody whos never seen the new environment be ore.

W H AT I T M E A N S

TIMES ARE CHANGING; YOUR APPROACH TO BI MUST CHANGE ALONG WITH THEM

How companies use the data thats available to them rom internal and external sources isincreasingly going to be a key competitive differentiator, i not the key competitive differentiator.Applying the best practices that weve identied will help deliver the BI capabilities required to makeorganizations more data-driven in particular the importance o agility and business user sel -service. But it doesnt stop there. Whether you call it big data or not, increasing data volumes andthe variety o data types that rms must process pose a challenge to existing ways o delivering BIand analytics capabilities. Increasingly tech-savvy business pro essionals want to have a greater sayin the sofware tools and hardware devices they use as part o their job, and theres a greater varietyto choose rom than ever be ore. Firms need to adapt their governance rameworks, technologyin rastructures, and tools to strike the right balance between enabling the business to get on with its job and maintaining control where necessary to avoid cost overruns, duplication, and compliancebreaches. Firms need to adapt existing approaches to account or:

Te good enough principle. Many business decisions dont depend on a precise gure; they

may only need an indication o trends or a rough estimate. And speed is o the essence ormany data requirements. While it remains important to understand a data source and thelevel o trust that one can apply to it, rms do not need to treat all data with the same rigor.

A different approach to data architecture. raditional layer-cake architectures sourcesystems at the bottom, integration and DW in the middle, and BI and analytics on top areproving ill-suited to the new environment o larger data volumes, greater variety o datatypes, and differing requirements or data quality and speed o delivery. By contrast, newarchitectures that involve canonical data models, data virtualization (AKA the in ormation

abric), and hub-and-spoke data management can help break down silos and provides amore exible environment better suited to catering or the ofen conicting demands posedby a data-driven organization. 7

New ways o delivering applications or not delivering them. One CIO we spoke withhas the strategic goal o providing data to the business via APIs, with the business decidingwhat tools and applications to deploy. While this approach may be a step too ar or manyorganizations (at least or now), the principle o decoupling back-end systems and data


20/21




preparation rom the delivery o capabilities to end users or example, by using newexible and personalized approaches such as internal app stores still applies.

ENDNOTES1 I you want to explore whether a true Agile methodology is right or you, you will nd a wealth o

in ormation and guidance in the Agile and Lean playbook. Forrester has also conducted an evaluationo the top vendors in sel -service BI plat orms. See the June 12, 2012, he Forrester Wave: Sel -ServiceBusiness Intelligence Plat orms, Q2 2012 report.

I you nd that your project and program management support structures and processes are no longer tor purpose, or arent suited to the context o your BI initiative, see the October 16, 2009, Making he Case

For he Next-Generation PMO report.2 Te key oundational principle or enabling effective sel -service is the separation o data preparation and

data usage within the ramework o an agile BI organization. See the January 25, 2013, Build An Agile BIOrganization report.

3 Forrester clients describe their data governance work as long on effort, but limited in success. Clientsdescribe an inability to sustain data governance beyond a project, eel it is too bureaucratic, or unableto show business value. Yet there is strong evidence that competency in data governance leads to bothoperational efficiency and tangible business outcomes. op-per orming organizations are using data

or competitive advantage and digging deeper into internal and external data. It is not technology thatseparates top per ormers rom the rest o the pack; it is data governance and the alignment o data processand business process that makes the difference. See the May 20, 2013, Data Governance Equals BusinessOpportunity. No, Really report.

4 For a detailed discussion o how governance models need to change, see the August 31, 2012, RearchitectGovernance For Sustainable Business Agility report.

5 For guidance on how to match BI tools to the most common use case, check out the ollowing Forresterblog. Source: Boris Evelson, Use Cases For Specic BI ools, Boris Evelsons Blog, June 28, 2012 (http://blogs. orrester.com/boris_evelson/12-06-28-use_cases_ or_specic_bi_tools).

6 For more details on the best practices and methodology to select a BI service provider, see the October 18,2012, he Forrester Wave: Business Intelligence Services Providers, Q4 2012 report.

7 For more details on data virtualization and canonical data model, see the January 5, 2012 he ForresterWave: Data Virtualization, Q1 2012 report. For more details on emerging hub-and-spoke dataarchitectures, see the June 12, 2013, Deliver On Big Data Potential With A Hub-And-Spoke Architecture report.
http://www.forrester.com/go?objectid=RES71902http://www.forrester.com/go?objectid=RES71902http://www.forrester.com/go?objectid=RES55456http://www.forrester.com/go?objectid=RES55456http://www.forrester.com/go?objectid=RES47315http://www.forrester.com/go?objectid=RES47315http://www.forrester.com/go?objectid=RES83342http://www.forrester.com/go?objectid=RES83342http://www.forrester.com/go?objectid=RES81161http://www.forrester.com/go?objectid=RES81161http://www.forrester.com/go?objectid=RES61392http://www.forrester.com/go?objectid=RES60746http://www.forrester.com/go?objectid=RES60746http://www.forrester.com/go?objectid=RES83303http://www.forrester.com/go?objectid=RES83303http://www.forrester.com/go?objectid=RES60746http://www.forrester.com/go?objectid=RES60746http://www.forrester.com/go?objectid=RES61392http://www.forrester.com/go?objectid=RES81161http://www.forrester.com/go?objectid=RES81161http://www.forrester.com/go?objectid=RES83342http://www.forrester.com/go?objectid=RES83342http://www.forrester.com/go?objectid=RES47315http://www.forrester.com/go?objectid=RES47315http://www.forrester.com/go?objectid=RES55456http://www.forrester.com/go?objectid=RES55456http://www.forrester.com/go?objectid=RES71902http://www.forrester.com/go?objectid=RES71902


21/21

Forrester Research, Inc. (Nasdaq: FORR) is an independent research company that provides pragmatic and forward-thinking advice toglobal leaders in business and technology. Forrester works with professionals in 13 key roles at major companies providing proprietary

Forrester Focuses On Application Development & Delivery ProfessionalResponsible for leading the development and delivery of applications

that support your companys business strategies, you also choose

technology and architecture while managing people, skills, practices,

and organization to maximize value. Forresters subject-matter expertise

and deep understanding of your role will help you create forward-thinking

strategies; weigh opportunity against risk; justify decisions; and optimize

your individual, team, and corporate performance.

ANDREA DAVIES, client persona representing Application Development & Delivery Professionals

About Forrester

A global research and advisory rm, Forrester inspires leaders,informs better decisions, and helps the worlds top companies turn

the complexity of change into business advantage. Our research-

based insight and objective advice enable IT professionals to

lead more successfully within IT and extend their impact beyond

the traditional IT organization. Tailored to your individual role, our

resources allow you to focus on important business issues

margin, speed, growth rst, technology second.

FOR MORE INFORMATION

o nd out how Forrester Research can help you be success ul every day, please

contact the office nearest you, or visit us at www. orrester.com. For a complete list

o worldwide locations, visit www. orrester.com/about .

CLIENT SUPPORT

For in ormation on hard-copy or electronic reprints, please contact Client Support

at +1 866.367.7378, +1 617.613.5730, or clientsupport@ orrester.com . We offer

quantity discounts and special pricing or academic and nonprot institutions.
mailto:[email protected]://www.forrester.com/mailto:[email protected]