Strategic Product Development...Project Management, defined the seven deadly sins of project management: • Mistaking half-baked ideas for viable projects. • Dictating unrealistic

Project Management Institute South Africa (PMISA)‘Regional African Project Management’South Africa, 3 – 5 November 1999Edited Conference PresentationsDocument Transformation Technologies

STRATEGIC PRODUCT DEVELOPMENTA strategic approach to taking software products to market successfully

*J van Zyl, **A J Walker

*Rubico, South Africa

**Software Engineering Applications Laboratory,Department of Electrical Engineering,

University of the Witwatersrand, Johannesburg, South Africa

Summary: How can innovative quality software products be built on time and within budget? - aquestion frequently asked by management and product delivery specialists. Product features areknown but competitors are hard on one’s heels and the computer software industry is movingrapidly. Different, but integrated, standards are needed in the pursuit of building capableprocesses for delivering products. Project management principles can be applied, but not inisolation. Strategic Product Development is an approach that uses a number of industry standardsin building an organisation capable of delivering commercially successful products. StrategicProduct Development integrates the best of project management, product management, qualitymanagement, product positioning and marketing standards. The capability of processes within theapproach are assessed and improved by using the ISO/IEC TR 15504 standard.

1. Introduction

Typical questions asked throughout a productdevelopment life cycle include: Where to start?What exactly to build? How will concepts beintegrated? Will the product perform? Whatare the dependencies and how will they bemanaged?

Product managers are faced with difficult, butchallenging tasks when it comes to productinnovation, concept development and productcommercialisation. It is not always possible tolisten to the customer and deliver exactly whatis expected. The circumstances involved inobtaining the needs and wants for products areinfluenced by individual situations. Productinnovation must take place in order to createproducts and services that potential customersdo not yet know they need.

This paper presents an integrated approach thatties many concepts together in assisting theproduct development team to deliver world-

class software products. It also addressesindirectly the fundamental skills that areneeded to make this possible.

Section 2 discusses the problems and solutionsaround product development.

Section 3 introduces two concepts that canassist with the product strategy developmentprocess.

Section 4 introduces the product developmentfunnel. It shows how ideas and other strategicinputs are used to build products that canbecome commercially successful.

Section 5 discusses portfolio management as atechnique to assist organisations in building theright products at the right time.

Section 6 shows how product lines are formedand managed. Project management principlesare used to ensure project delivery.

Section 7 discusses some of the basicdeliverables and work-products that should be


produced in software related projects. Whatdoes it mean to deliver a product to market?

Section 8 discusses what is needed onceproducts are delivered to ensure long-termcommercial success of the product. Somebasic measurements are used to evolve productdevelopment organisations.

Section 9 brings a number of concepts together.The learning model is used as a medium tomature and evolve the product developmentbusiness as a whole.

Finally some statistics are presented andRubico’s technology life cycle is discussed.

2. Problems and solutions to productdevelopment

Many books have been written, discussionsheld and debates organised concerning thereasons products fail. The public has beenconditioned to accept that software projects areoften associated with high business risks.

Discussing problems and failures, as discussedin this document, will only skim the surface ofthe vast amount of material available on thistopic. It is not intended to be comprehensivebut qualitative. These areas were selected inthe light of product-based management.

What does it really mean when we say that aproduct or project failed? A definition mightbe: When a product, after being defined andcommitted to, did not give sufficient or anyreturn on investment, or was terminatedprematurely, then it is considered a failedproduct.

This definition obviously relates strongly to thecommercial world, where capital (financialinput) is needed for future developments.

Does this mean that all product and/or projectfailures are bad? Not at all - the learnings fromfailures are invaluable to an organisation,obviously assuming that learning took placeand that there is a way to disseminate theknowledge gained from new experiences.

Failures can be classified into two categories[1] namely:

1. “Stupid, uncaring failures” in which theindividual/s responsible need to becensured and

2. “Calculated risks or honest mistakes”,where the individual/s learnt valuablelessons but took calculated risks.

Some problems and failures that have occurredin the computer industry over the last few yearsare discussed next. Some key concepts that areneeded to solve these problems are presented insections 2.2, 2.3 and 2.4.

2.1 Problems in context

The discussions in this section are not acomprehensive representation of the computerindustry, but provide some backgroundconcerning the challenges facing productdevelopment.

In Glass [1], Kapur, president of the Center forProject Management, defined the seven deadlysins of project management:

• Mistaking half-baked ideas for viableprojects.

• Dictating unrealistic project deadlines.

• Assigning underskilled project managers tohigh-complexity projects.

• Not ensuring solid business sponsorship.

• Failing to break projects into manageableentities.

• Failing to institute a robust project processarchitecture.

• Not establishing a comprehensive projectportfolio to track the progress of ongoingprojects.

Other issues include:

Stock markets and venture capital are notsufficient to succeed: Radical turnarounds areneeded to pull companies back on track, for


example, Sapiens, Borland, Seer, Dynasty,Wang, IBM, 3Com, Oracle [1].

Remember the basics: Galileo International is aprovider of electronic global distributionservices for the travel industry. They ran aproject to implement a new product calledAgile for many months over its originalestimated time frames. As management tookcontrol it became apparent that the basics ofproject management were overlooked. Fixeswere introduced and the project delivered latewithout being cancelled [1].

Not listening to customers: Technical peopleare not skilled at listening to customers. Theyfocus on the technological achievements andchallenges and forget that customers need touse their inventions. Business processes needto be in place that will facilitate communicationchannels.

Complexity causes confusion and costs:Building overly complex products early in theproduct’s commercial life cycle can causefailure. More highly skilled people are neededas products become more complex. Start smallto test the market and then evolve the product.

No definition of “complete”: Does a productever reach a state of completeness? In the everevolving world of improvement, few productsreach a state of being fully complete. Marketdemand and potential influences the level offunctionality that will be provided throughoutthe product development life cycle.Completeness is normally not defined beforeproduct development commences causingproblems in coming to a close. “Deliver lessmore often”, should be a slogan usedthroughout software delivery projects.

Emotional attachments: Teams becomeemotionally attached to their products. As timegoes by and energy pours into development,individuals become tied to products and fail tomake objective decisions. Sometimes lossesneed to be cut and direction changed.

2.2 Product management excellence

Product management is a strategic issue.Product development starts with ideas and thejourney commences to commercialise theseideas into tangible products. Productmanagement is defined to set the scene of whatis meant by “excellence”.

Product management can be defined as [8]:“The product manager is responsible for aproduct, a product line, or several distinctproducts in a group. The areas the productmanager has to deal with include packaging,branding, research and development,engineering, and production. The manager hasto continually appraise the product’sperformance in terms of growth targets, marketshare, working capital targets, and return onassets managed”.

In the light of this definition, the keycharacteristics required by productmanagement include:

• Being responsible for a product: Thisrequires total commitment and dedicationto the product, and intrinsic accountability.

• Packaging and branding: How will theproduct be presented to the outside world?What makes people look at the features inlieu of satisfying their business problems?

• Research and development: Managing thefunctional and non-functional [9] attributesof the product. It might be that releases ofthe product are produced purely to increaseperformance or usability, for example, andnot necessarily functionality.

• Engineering: Developing the product basedon sound software engineering principles[5,10].

• Production: Developing the product to arobust commercial state.

• Appraising the product: Accepting andpromoting continued feedback on theproduct. Providing sufficient feedback intoresearch and development forimprovements to take place.


Some management practices needed tocomplete the product manager’s role are [11]:

• Planning and tracking of products andrelated processes.

• Achieving the right organisational structurebased on current and future projects.

• Effective funding to get products to market.

• Building and communicating businesscases to relevant stakeholders.

• Training and education of the teams.

• Organisational risk management.

• Operations management to ensure effectiveproject integration and resource allocation.

• Customer and supplier interfacemanagement.

• Launching and institutionalising productlines.

• Technology forecasting and trend analysis.

Additional areas where product managementmust focus include:

• Direction setting and aligned of theorganisation.

• The selection, training, and development ofpeople to ensure a capable workforcewhereby the vision can be realised.

• Through people, creating, shaping andinfluencing how work gets done.

Product stakeholders need to react whenproblems are identified in the development ofproducts. It might be required that teams arere-aligned through horizontal movements,where team members are moved to otherteams, or re-construction of the entire team,where the team is re-formed with new roles andresponsibilities. Failure to act or react quicklyto problematic situations can cause delays,possibly leading to product failures. It is betterto make a decision, be wrong, and then fix thesituation than to make no decision at all and letthe world pass by.

Product management should ensure thatreleases are delivered more often in the earlydays of the product’s life cycle. Great careshould be taken that the architecture is soundand a platform is built for future development.This requires a focus on all product issues inthe early days of delivery.

Some of the key skills required by strategicproduct direction setters include: beingfinancially astute, having high level contacts,and technology skills.

2.3 Leadership roles

Leaders need to lay the essential productmanagement foundations early on in thedevelopment cycle. To ensure properexecution, follow up reviews must beconsistent and disciplined. There are “pulling”and “pushing” roles. Ideally leaders shouldspend most of their time by pulling (leading).

Pulling roles:

• Direction setting: the vision, mission andbusiness strategies need to be translatedinto a “picture” shared and understood bythe entire team for effective execution.

• Product line architect: Building the right setof products is vital to keep the organizationcompetitive.

• Portfolio manager: There must be a clearunderstanding of the products beingdeveloped and to be developed. Themanagement of people, skills andknowledge is another portfolio thatinfluences product success.

• Ultimate ownership lies with the leader.The decisions around which projects shouldcontinue and which not, become one of themost important aspects of a leader’s role.

Pushing role:

• Launching teams involve the approval ofindividual projects and versionresponsibilities in a product line.


• There must be a continuous source ofpositive energy from the leader. Skills andknowledge shortages must be identifiedopenly and corrective plans put in place.

• Projects do not reach completion unlessthere is absolute commitment by the teamand team members. Building commitmentrequires strong leadership and peoplecapabilities.

• Finding and fixing problems that relate toprocess, people and technology can resultin drastic improvement exercises. Theleader needs to know when to changedirection.

2.4 Management influence on productoutcome

Organisational structures must be designed insuch a way that the senior manager stays inclose contact with the product development lifecycle all the way up to pilot development.People such as Bill Gates, the Chairman andCEO of Microsoft, still become involved withthe creation of ideas and then how the ideasmust be developed.

Figure 1 Management Influence in ProductDevelopment [2]

Management should become involved early inthe process by building the right culture andcapability to deliver innovative products. Cleardirection needs to be provided all the waythrough the life cycle. The above figure showsthat management normally moves into fire-

fighting mode as problems occur when moneyand time run out near the end of initial pilotprojects. By then it is too late to influence theoutcome significantly.

3. Concepts to assist with strategy

3.1 Building a capable organisation

The ISO/IEC TR 15504 [5] standard has acapability dimension that assesses the ability ofprocesses to perform. Mature organisationsuse this as a method of benchmarking theircapabilities against others in a certain industry.Each level has a brief reference to theindividual and team abilities needed toperform. The five capability levels are:

Level 1: Performed: Processes are performedand people generally understand what needs tobe done. Work products are produced.

Level 2: Managed: Processes produce workproducts and deliverables according to plan.Work products satisfy required quality andperformance levels. Individuals working inthis environment understand the implicationsof working with a defined process in assistingthe delivery process. Teams are still unsure ofhow the final deliverables will be met.

Level 3: Established: Processes are performedbased on sound software engineeringprinciples. Processes are tailored to providethe best fit to what needs to be delivered.Defined process outcomes are achieved at thislevel. Individuals have obtained a clearunderstanding of the abilities needed toproduce software products based onengineering principles. Teams operate withinthe limits of strategy and have a goodunderstanding of what is needed to deliverpredictably.

Level 4: Predictable: Defined processes areperformed consistently to achieve goals andobjectives. Predictability is high as the processis quantitatively managed. Individuals candeliver product after product predictably. Theyknow when a process is out of control. Teams


can decide actively when to terminate a processunder execution based on their capabilities.

Level 5: Optimising: The processes arerepeatable and can be optimised to achievebusiness goals. Process innovation drives theability of the organisation to become morecompetitive. The dimensions of people,technology, and process are all consideredwithin the context of the organisation’spurpose. Teams operate in the context ofstrategy and purpose and changes are dealt withproactively.

3.2 Building a “stage 4” business

The strategic role of teams in the businessplays an important role in productdevelopment. Teams need to be balanced sothat each product area can be a source ofcompetitive advantage. Teams require theenvironment and direction to perform at theiroptimum. The risk is that teams can be pushedto perform at a level before they are ready [3].An evolutionary approach is required wherepeople and teams are educated and preparedover time to operate at level 5 (ISO/IEC15504).

Weelwright and Clark [2] define four stages ofevolution that a team goes through beforeachieving a strategic role:

Stage 1: Minimise negative potential, beinternally neutral.

Stage 2: Match competitors, be externallyneutral.

Stage 3: Back the strategy, be internallysupportive.

Stage 4: Pursue a distinctive, sustainableadvantage, be externally supportive.

Each stage is dependent on the previous, and isneeded to facilitate essential learning. It is onlyonce people are fully ready that they canoperate purely within strategy as being the onlydirection-setting framework.

3.2.1 Stage 1: Minimise Negative Potential -Be Internally Neutral

The team is put into a narrow operating rolewith little or no input into strategy. It isrecognised that the team is necessary but is notessential. These people are not involved withanything outside of their environment andfunction inside of a tactical role.

3.2.2 Stage 2: Match Competitors - BeExternally Neutral

The teams understand competitors and start tobenchmark their performance. Outsideassistance is still required although theyoperate at a higher level than stage 1. Theystill have a limited strategic role. The only waythat management can allow teams to operate atthis stage is when a managed level is reached.

3.2.3 Stage 3: Back the Strategy - BeInternally Supportive

The teams can make decisions in the light ofbusiness strategy. They are actively supportingkey business objectives. The function isdirected and based on their abilities to satisfythe business objectives. People involved at thisstage have input into the business strategy.Little outside influence is needed for this teamto function. This stage will only be reachedonce the capabilities of the team are wellunderstood, and they operate at an establishedlevel.

3.2.4 Stage 4: Pursue a Distinctive,Sustainable Advantage - Be ExternallySupportive

Customers and other stakeholders recognisethat the team is a source of competitiveadvantage. The capabilities of the teamdifferentiate the organisation from itscompetitors. Team leaders understand theirstrategic role and identify long-term trends andopportunities to build capability in anticipationof business strategy. The capabilities arereplicated into other areas of the organisation.Delivering products consistently in line with


organisational strategy requires at leastpredictable processes.

In order to build a stage 4 product developmentorganisation the following items should beconsidered:

• The capability of the organisation as awhole in order to achieve predictableoutcomes. Strategic plans must be in placeto make all within the organisation aware ofthe intention of the organisation to becomeworld-class.

• A learning framework to assist in theaccelerated growth of teams and individualsin teams. The culture of the organisationmust dictate the fact that life-long learningis needed.

• People maturity and capability in dealingwith the evolution as different teams can beat different stages of development. Cleardevelopment plans need to be in place toassist in the selection of people to performroles.

4. The product development funnel

Product conceptualisation and product strategicplanning begins with strategic planning atBoard level. The following diagram illustratessome of the stimuli that influenceorganisational strategy.

Figure 2 Business Environment

Organisations operate in an environment thatinfluences how products are built and targeted.It is not within the scope of this paper todiscuss the implications of the macroenvironment on product delivery.

Products and services are targeted based on anumber dimensions (not a complete list):

• Targeted customers: What need does theproduct satisfy? How does the customerbuy products of this kind?

• Suppliers: What are the influencing factorsthat will inhibit or enable product delivery?

• Competitors: Are the products positioned insuch a way that there are cleardifferentiators?

• Opportunities: How can an organisationtake advantage of the current environmentin which it operates? New marketopportunities and diversification can takeplace to generate opportunities.

The product development funnel is aimed at themicro environment. There are number offacets in the environment that must beconsidered to set direction and implement thefunnel effectively [4].

ProductDevelopmentIdea

Gene-ration

ProjectDefinition

Filter:�Market & bus strategy�Technical feasibility�Market potential of idea

Stimulae :

Strategy

Research

Knowledge

Wants

Needs

etc.

MarketLaunch

ConceptDevelopment

Filter:�Functionality limited

by time-to-market

Filter:�Advert functionality�Project planning

Filter:� Technical readiness� Market integration

Figure 3 Product Development Funnel

Strategic planning and placement directlyinfluence the product delivery cycle, asillustrated.

The funnel is used to take relevant inputsrequired to deliver products to marketcontinuously. Inputs are accepted from:

• Existing customers: Inputs are providedbased on current products and problems.These are normally the “needs” that makeproducts work in the field.


• Research: Basic and advanced researchprojects are used to explore new conceptsand generate innovations.

• Market trends: Trends are analysed toassess the ability of the organisation tocapitalise on opportunities.

• Strategy: This is the interpretation andformulation of where the organisationneeds to go based on the analysis ofavailable information.

Concepts are assessed throughout the funnel toobtain a clear understanding of the commercialrealities around taking the final product tomarket.

Filters are used to determine the inputs into theeventual product. Products are classified intoproduct types. This is done to have optimalmethods applied based on the type of productthrough the evolutionary and revolutionaryproduct development process. Commercialimpact is assessed on an ongoing basis.Predictability is important as the teams candraw from previous experiences.

All concepts and/or products under researchmust be channelled through a productdevelopment funnel. Organisational strategistsare responsible for funnel integration and teamformation in order to deliver products tomarket.

4.1 Product types

Products can be classified into four maincategories. Products evolve over time andnormally start with a breakthrough. Thesebreakthroughs are used as platforms for otherrelated developments. Platform products areused to build derivatives. Derivatives aresupported over time up to the point ofretirement.

Teams operating at stage 4 (section 2.4) canmake decisions as to which product features areimplemented when. They own the entirefunnel and development that are required tocommercialise the product.

4.1.1 Breakthrough products

Breakthrough products are products that aresignificantly different to existing productofferings. It normally means that thearchitecture has changed or components of anexisting incomplete architecture wereinnovated. New value propositions can bepresented to existing, but more likely, to newcustomers.

The life span is normally long and is seen to bea product that will be accepted only by earlyadopters. It could mean that the business isentering a new market.

4.1.2 Platform products

Platform products are built using breakthroughproducts. This is where the real customer valueis realised. Customers will see that there issomething innovative that can be used. Newtechnologies and concepts are generallyintroduced with platform products.

Product families are formed to group thedifferent technologies and concepts together. Itcould mean that the way the product isdistributed needs to change.

4.1.3 Derivative products

Derivative products are built to leverage theinvestment in breakthrough and platformproducts. Extensions are added in aneconomical way but still within the qualitylimits. The product families become morestable as derivatives are created and deployed.

4.1.4 Support products

Support products are the small extensions thatare needed to ensure product quality. They canalso be used by the sales force to presentadditional features on products that are nearingthe end of the life cycle.

4.2 Process standards

The ISO/IEC TR 15504 and ISO 12207 [5]standards can be used as a basis for processesin software related organisations. These mustevolve with all facets [4] in the business to be


able to operate at a level where innovations canbe commercialised consistently.

Processes must enable the business to moveforward and not be a hindrance. Tailoring isrequired throughout the product delivery cycleto obtain effective results. One of the biggestrisks is that management’s understanding is notsufficient to lead their staff. Predictableprocess behaviour is required to delivereffectively.

Strategies should be in place wherebycontinuous cycles of education, experienceacquisition and evaluations can be performed.Each cycle will represent a step towardsbecoming a world class software provider.

4.3 Project management standards

Having a good process does not guaranteeproduct development success. The basics ofproject management must be in place toexecute the process. Project management isused to deliver a certain set of deliverableswithin the product development funnel.

Some of the basic processes [6, 12] that mustbe in place in order to manage projects are:

• Scope management: Involves all aspects ofscope definition and change management.

• Time management: Involves the planningof activities as defined in the process andincludes the estimation of duration,schedule development and control.

• Cost management: Planning and control ofproject related costs.

• Quality management: Deals with theprevention, detection and tracking ofissues.

• Human resource management: At projectlevel staff planning handles all aspects ofteam selection and development.

• Communication management: All channelsare defined and managed.

• Risk management: Risks are identified andclassified. Risk management might involveperforming research studies.

These processes will be controlled at productdevelopment funnel and delivery cycle levels.The macro product development plan shows allthe product lines and interdependencies withregards to deliverables and resources. Productcategories are kept together and innovations inthe area of platform products are balanced.

4.4 Architecting the product line

Breakthrough products normally presentsignificantly new and reworked architecturaldesigns. It means that there are untested andunexplored areas that represent many risks, asthe products that use the architecture mature.Architecture is the most vital area that needsstabilization early on in the product life cycle.Costly rework is minimized if there is acomplete and stable architecture for a productfamily or product line.

Architectural snap-shots are needed to evaluatethe appropriateness relating to productdirection:

• Products available today.

• Products under development.

• Products of the future.

Considerations when working with productarchitecture include:

• Products of the future.

• Position: price/performance, market andcompetitors.

• Type: market need, contingencies.

• Timing: introduction, pace of change,market perception.

• Relationships: products for optimalintegration and feature overlap.

Product architecture must be designed in such away that software component and service-object reuse rates are high. This involves thetight integration of project deliverables


amongst product lines. The Engineeringprocess category, as part of the ISO/IEC TR15504 standard, has as part of its designprocess an architectural design base practice.

5. Portfolio management

The entire portfolio must be managed againstbusiness goals. Goals must include aspects thatcover innovation in order to stay competitive.Radical ideas must not be excluded early on inthe product funnel. These ideas can form thecore of future products and present solutions,when implemented, to existing customerproblems.

There are three macro-goals [13]:

• Value maximisation: there must be tangiblevalue being generated by the portfolio.Some of the elements include: long termprofitability, return on investment, the rateand likelihood of success and value createdfor customers that consider current andfuture potential.

• Correctly balanced portfolio: theorganisation needs to define what balancemeans. Some of the elements that wouldbe considered are: long-term vs. short-termprojects, high risk vs. low risk, quickreturn, markets that the business operates invs. future markets to compete in, differentproduct types and different project types.

• Business strategic direction: the business isultimately run based on its strategy that willinclude: financial and funding aspects,product direction, customer and targetmarket, use of suppliers and partners andthe tools and equipment that will be used.

5.1 Value maximisation

Expected Economic Value is a method used tomaximise the commercial worth of theportfolio. It takes into account factors such asthe:

• Expected commercial value of the product,

• probability of commercial success,

• probability of technical success,

• development costs of the projects that makeup the product,

• commercialisation costs that includemarketing and launching the product.

• and the nett present value of the product’sfuture earnings.

It is not always feasible to follow this approachwhere up-front financial calculations determinea product’s future. There must be times whenmanagement will continue with projectsbecause of an intuition. Many organisationshave built world class products even thoughthey were seen initially as being a looser.

Organisations should build their own scoringmodels whereby the factors that are mostimportant to them will be considered. Achecklist might look like this:

• Expected economic value, including theassumptions that accompany the ratios andvalues.

• Core competencies in the organisation thatcan be exploited in such a way thatinnovations will occur that can yield majorreturns.

• Trends in the market can play a major rolein deciding what to build and when to buildit.

• Competitive advantage and marketperception influences all decisions.

5.2 Balanced portfolio

It is not always possible to reach a balanceamongst product lines and projects.Information might not be available orenvironmental conditions might influencedecisions. Customers play a major role indecisions made about what gets built. Productmanagement needs to make decisions thatbalance existing customer value vs. what isneeded to generate new customers in the future.

Having a portfolio with breakthrough productsonly, for example, presents a major risk, as


there will be maturity and deploymentproblems. It is advisable to have a balance ofdifferent products at various stages of maturityin the product family.

A checklist with a weighting mechanism,defined by the product managers, might looklike this:

• Will customers buy the product that isbeing developed?

• Sales method and financial impact.

• Technology issues that influence thesatisfaction of existing customers.

• Will the product provide competitiveadvantage?

• Product type, maturity and the effortneeded to make it commercially viable.

5.3 Strategic direction

Strategy in the software organisation has onemajor component that everything revolvesaround. Software is conceptualised, invented,designed and built by people. There is a degreeof automation by using development tools, butthe thinking part is human. By far the mostexpensive and strategically critical item istherefore the human component.

Strategy should consider the following items:

• Invention vs. delivery ability of people.Direction should dictate clearly howinventions are utilised and commercialised.

• A “top-down” strategic alignment meansthe vision and mission set strategy.Management allocates money.

• A ”bottom-up” strategic alignment meansthat the company exploits its technologyability and shapes the business around it.When successful, the portfolio will lookafter itself.

• A “top-down, bottom-up” alignment meansthat the vision and mission sets the overallstructure, but innovations could changethat. Key people have input into thespending of money across portfolios. This

seems to be the best approach used byinnovative and successful softwarecompanies.

6. Product delivery cycles

The product development funnel filters all theproduct concepts up to the point wheredevelopment must start. This requires a cleardefinition of features that need to be developedover a certain time frame. Time-to-market, asdefined by product strategists, is used toprioritise the product development cycle.

A product line is made up of a number of codelines. Code lines are for specific technologiesand implement features for a specific release ofa product.

Architectures are realised over a number ofcode and product lines. The architecture mayeven change over the life cycle of a productalthough it is not likely. A product is builtonce the architecture is defined and forms thebasis of future releases.

Products are built over many releases. Eachrelease adds functionality and becomesincrementally bigger and more complete. Itdoes not mean that only one code line is used.Many technologies can make up one release ofa product, and usage is based on relevantstrengths and weaknesses.

Product Development

Product line, e.g.ROB (Rubico Obj. Broker)

Architectural Definition

Development Life Cycle:1-5 years

Code lines3.0.0

Java Java Java

C++C++C++

3.0.1 3.0.2 3.1.0Java

C++

DeliveryLife Cycle(2-3 months)

Productbaseline

Productbaseline

Technicalproducts

Managementproducts

Figure 4 Product Delivery Cycle


Components that are constructed during thevarious code lines for a product line are used assoftware assets [11]. These assets includeCOTS (commercial off the shelf) softwarecomponents and are all part of the final productfamily. Inter-product line dependencies aremanaged by having these well-definedcomponents under configuration managementearly on in the funnel.

7. Evolutionary or revolutionary cycle

Innovative products are normally seen to behigh risk when the time-to-market deadlinesare aggressive. Think of the funnel as arevolutionary product development life cyclewhere evolution is challenged continuously.Challenges could involve the rethink ofarchitecture resulting in an entirely newplatform product. The key to having productsaccepted by customers is to have more releasesearly on in the life cycle with a tight feedbackloop.

Specialists working on features in a releaseneed to understand the importance of customeracceptance and feedback. The followingdiagram is aimed at having a part of or theentire product delivered.

Technical and management products areneeded to establish a baseline. Baselines can beinternally (technical reasons) or externally(marketing and business) defined. It is vitallyimportant that the team understands theimplications of bringing a product to a close.Forward and backward planning techniques areused to set the actual “work time”. Forexample, two days are needed to get the teamready and another three days to prepare therelease.

Support Processes:

Technical Products(Governed by Standards)

Running Software

Advert:Features of Release

Release Notes

Documentation, userref. etc.

Training Material

Goal of each :Comprehensiveenough to

ensure that noquestions areasked.

Scope :Governed byTechnologyArchitecture Doc

Baseline

Management Products (Governed by Standards)

Requirements Spec(linked to features)

Software Spec(linked to features)

Project Plan

Audits

Quality Plan

Testing

Config management

Product Delivery

Maintenance ReleasesUpdates

Defect Tracking

Measurement AnalysisMetric Baselines

Figure 5 Working the Baseline

Management products are only used to assist indelivering the products the client sees(technical products are delivered to the client).Maintenance activities start once a product isbaselined and released. All defects must betracked to build a knowledge base about aproduct’s behaviour. Root cause analysis iscritical in eliminating mistakes during the nextbuild. Measurements are defined before theproject starts. Metrics can be analysed andupdated based on product, process and peoplebehaviour. The learning model [4] is used toevaluate the findings.

Management reporting is done at the point ofbaseline. All measurement and metricelements are populated with the latestsuccesses and failures. Time sheets are used toupdate the actual effort where project plansreflect the planning. The results of the projectclosure processes are fed back into the funneland are used by product management to planfuture cycles. A number of elements areconsidered before product concepts areexplored in the funnel including: projectclosure feedback, customer issues, markettrends and technology strategy.

8. Shipped product

Defects, enhancements and requests aremanaged once the product is shipped.

• Defects identified outside the developmentenvironment must be analysed in detail.The cause of the defect should be


investigated after which a corrective planshould be implemented.

• Enhancement requests can either bebecause the product does not have theselected features that the customer requiresor the product is not used as intended bythe author. The feature list of the product isused to manage the delivery ofenhancements.

• Requests can be seen as wishes. What isrequired to make the product more useful?

Internally the team needs to close off the cycleof development before commencing with theproject closure. It is at this time that all themeasurement data are collated and metricsupdated. Some product and product familyrelated metrics include, and are classified intopre-release and post-release:

Pre-release:

• Product size and level of re-usability.

• Product defect types by phase.

• Re-work effort by phase.

• Project development effort.

• Process failures.

• Re-usable components by project and/orproduct.

• Productivity - time by month for activitygroup / item.

Post-release:

• Number of defects / enhancements /requests over time.

• Cycle Time - time to fix defects.

Statistical reports are produced weekly,monthly, quarterly and yearly by productmanagers. The results are evaluated before anyimprovement and/or education programmesstart. Metrics are also updated at this time.

Processes to manage the evolution of standardsand patterns in the various product areas areinitiated after product shipment. This normally

coincides with the evaluation process in thelearning model [4] as discussed in section 7.Processes are normally not clearly defined inthe early stages of using new technologies andinnovations (that is early in the productdevelopment funnel). There is interdependencebetween the kinds of products that are built andthe kinds of processes used. Specialists andproduct managers evolve the processes with theproducts. Stage 4 organisations have thecapability to understand this evolutionaryprocess and manage the associated risks.

9. Learning the lessons

The Learning Model [4] is used as a tool tohelp the process of “learning from one’smistakes and successes”. The basic principleof this model is “Is there a better way of doingthings?” Understanding one’s own learningpatterns will assist in increasing one’sknowledge and in return influence everyone inthe organisation. This is one of the mostimportant aspects of process improvement andmaturity.

The learning model is implemented at variouslevels in an organisation or institution. Thecycle of evolution is slower for direction settersthan for the more focused and detail-orientedenvironments. Each area has its own patternswith which to deal.

Figure 6. Learning Model Processes

The process normally starts with the recordingof project experiences in ways that make futureanalysis possible. These experiences are used


during the evaluation process to determinewhere things can be improved without losingsight of the “bigger picture”. The evaluationprocess produces work-products that will beused to enhance, improve or re-engineereducation or communication programmes.This facilitates re-alignment of processes,people and technology. At this point, newtraining candidates can benefit from theexperiences, and existing staff can go throughcycles of re-training. The re-trainingconcentrates on new ideas and the areas thatchange.

There is a reverse spiral that indicates thecontinuous feedback needed for the process tooperate effectively. Experiences are based onwhat was taught during the education process.Evaluation is based on the experiences over atime frame.

Multilevel learning is when learning happensacross two dimensions in the organisationnamely: vertical (that is, from individual togroup to corporate) and horizontal (that is,across product functional areas) [14]. Thelearning model facilitates both methodswhereby different spirals integrate at differentpoints in time.

Finally, the learning model is a high levelprocess that encapsulates, or co-exists withmany other models, architectures andframeworks.

10. Discussion

The concepts discussed in this paper have beenapplied at Rubico Technologies, the division atRubico responsible for a number of innovativesoftware products. Development ofbreakthrough products started in September1994 and now, in July 1999, development isstill continuing. In the time frame up toSeptember 1994, the innovations that wereused at Rubico were not viable due to certaintechnological and human limitations. Thefollowing diagram illustrates some of the majorreleases [7] of the Rubico Technology ProductFamily.

1994 1996 1997 1998 1999 2001 20022000

Time

1.1.7

1.3.2

1.4.0

2.0

3.0

4.0

Figure 7 Rubico Technology Product FamilyLife-Line

Various innovations were produced up todelivery and many more are anticipated.Breakthrough products include the structuresystem, virtual database that is the evolved andextended structure system, rules engine thatincludes the calculation pad and state transitionengine, generic masterfile, transaction engines,data cube and dynamic user interfacedefinitions. Platform products were introducedat every major release. Today there are manyprojects running that are building derivativeand support products.

With reference to figure 6, releases 2, 3 and 4are very close together. Early innovationsfacilitated extended platform products that willevolve quickly as support products are builtbecause more customers are coming on-line.Refer to Appendix A for a detailed view ofevents. It is anticipated that version 3 willinvolve various breakthroughs where version 4will involve the volume deployment of theversion 3 architecture. The assumptions arebased on deliveries as measured and trackedover the last five years on the Rubico productlines.

Many software engineering principles had to betaught and implemented in the time framebetween 1994 and 1997. Key commercialsoftware development principles were new tothe new recruits and created challenges for theeducation processes.

Delivery capability was (and is) challenged dueto the sale of products to various customers thathad to be implemented in the 1998 to 1999time frame. Focus is firmly on the execution ofearly innovations.

Project Management Institute (PMI)‘Regional African Project Management’South Africa, 3 – 5 November 1999Edited Conference PresentationsDocument Transformation Technologies

Some of the fundamental findings in theevaluations that took place over the last sixyears were that:

• Academia does not prepare individualsproperly for the “real world” of commercialsoftware delivery.

• Decisions are made in circumstances ofresource shortage, too little funding andunanticipated risks.

• Without a major investment in ongoingeducation the deliveries that took placewould not have been possible. Theassumption is that there will never besufficiently skilled and knowledgeablepeople on projects.

• Strong leadership in the field of softwareengineering and product management basedprinciples is essential for the delivery ofcommercial products.

• Strategy and long-term planning for overallorganisational behaviour are essential topredict how the product life cycle willunfold. But can one really predict howtechnology will evolve?

• Technology specialists become tied into theinnovation cycle without getting to deliverworking products effectively. This couldbe because of the focus on working withtools and not developing tools.

Strategic product management concepts shouldalways be implemented incisively witheffective follow-up.

11. References

[1] Glass R.L. (1999). Computing Calamities,Lessons learnt from products, projects, andcompanies that failed. Prentice-Hall. ISBN0-13-082862-9.

[2] Weelwright S.C. and Clark K.B. (1995).Leading Product Development, The SeniorManager’s Guide to Creating and Shapingthe Enterprise. Harvard Business SchoolPress. ISBN 0-02-934465-4.

[3] van Zyl J. (1998). People Maturity Model.Rubico and WITS University. (available onrequest)

[4] van Zyl J. (1998). Software EngineeringApplications Laboratory. WITS University.

[5] ISO/IEC TR 15504. (1998). SoftwareProcess Improvement and CapabilityDetermination, Part 5: An AssessmentModel and Indicator Guidance.

[6] PMI (Project Management Institute).(1996). PMBOK – Project ManagementBody of Knowledge.

[7] van Zyl J. (1994-1999). Rubico feedbackpresentations to the engineering staff.(available on request)

[8] Cary L. Cooper C.L. and Argyris C. (1998).Encyclopedia of Management. BlackwellPublishers. ISBN 0-631-20885-2. P528

[9] ISO/IEC CD 9126. (1997). Qualitycharacteristics and sub-characteristics.

[10] ISO/IEC 12207. (1998). InformationTechnology Software Life Cycle Processes.

[11] SEI. (1998). A Practice for SoftwareProduct Line Practice – Version 1.0,Software Engineering Institute & CarnegieMellon University.

[12] ISO 1006. (1997). Project ManagementStandard.

[13] Cooper, Edgett and Kleinschmidt,.(1998). Portfolio Management for NewProducts. Perseus Books. ISBN 0-201-32814-3.

[14] Weelwright S.C. and Clark K.B.(1995). The Development Challenge. AHarvard Business Review Book: HBSPress. ISBN 0-87584-609-2.


12. Author Contact Details

Jay van Zyl is the director responsible fortechnology, research and education processesat Rubico. He has a background in electronicengineering, information technology andinternal auditing.

He is currently involved with the managementand research of education programmes forbusiness and technology professionals. Theseeducation programmes cover the entiresoftware engineering spectrum and specificsoftware development methodologies. He isalso involved with the research anddevelopment of implementation, developmentand change management methodologies atRubico. Being at the head of research: productdirection, software requirements engineeringand strategic planning for the Rubico softwareproduct suite, is also managed by Jay.

His research interests include: organisationalprocess, design and behaviour in the context ofsoftware engineering and product delivery,process innovation and strategic productmanagement, process assessment, processimprovement, people measurement andbusiness modelling methodologies.

Jay is a post-graduate, part-time student at theSoftware Engineering Applications Laboratory,Department of Electrical Engineering,University of the Witwatersrand,Johannesburg, South Africa.

He is a member of the ISO/IEC workgroupsand is a participant in the SPICE (ISO 15504)Trials. He also serves on the board of advisersfor the International Quality AssuranceInstitute (QAI) and is a Certified QualityAnalyst (CQA).

Contact details: Jay van Zyl, Rubico, P.O. Box5016, Rivonia, 2128, Republic of South Africa.Office Phone: +27-11-808-1000, Fax: +27-11-808-1101, Internet E-mail: [email protected]

Internet URL: www.rubico.com/JayVanZyl

Alastair Walker is presently an associateprofessor in the Department of ElectricalEngineering, University of the Witwatersrand.He was responsible for establishing theSoftware Engineering Applications Laboratoryin 1988. Since that time, the SEAL has becomeknown as the leading provider of advancedprofessional courses on topics in SoftwareEngineering, particularly in the areas ofapplication of object-oriented designmethodology in industrial applications and insoftware quality systems management. SEALwas the first academic institution in the worldto receive an ISO 9001 certification forsoftware development.

He is a Certified Quality Analyst, a CertifiedSoftware Quality Systems Auditor and is thecountry’s only Certified Auditor in Educationand Training.

He is presently a member of the Advisorycommittee and the Information TechnologyCommittee of the South African Bureau ofStandards. As a member of the SouthernAfrican Council for the Certification of QualitySystems Auditors, he advises on issues relatedto training of software quality system auditors.

He has recently been appointed convenor of theSABS work group involved in the developmentof the new international standard for softwareprocess assessment and for the upgrade of the1991 version of ISO 9000-3 and is a SouthAfrican representative on theISO\IEC\JTC1\SC7 work group. He is thesecretary of the Software Quality SpecialistDivision of the South African Society forQuality.

Contact details: Professor AJ Walker, SoftwareEngineering Applications Laboratory,Department of Electrical Engineering,University of the Witwatersrand, Private Bag 3,PO Wits, 2050 Johannesburg. Phone: +27 11716-5469, Fax: +27 11 403-1929, Internet E-mail: [email protected]

mailto:[email protected]

www.rubico.com/JayVanZyl

mailto:[email protected]


Appendix A. – Summary view of Rubico Technology Products life line

Time frame General happenings1993 Pilot projects, anticipated breakthroughs based on past research.

Education focus, preparing environment.Funnel started and concept development explored.

1994 Product architecture and initial infrastructure, breakthrough products.Education process, technology and architecture.Standards, patterns and process implementation focus.Funnel explored further and initial projects put in place to evaluatebreakthroughs.

1995 Product innovation and concept development, breakthrough andplatform products.Process and measurement focus, commercial product developmentprinciples introduced to team.Learning process evaluations took place to accelerate developmentand new product funnels kicked off.

1996Firstcommercialready release1.1.7

Product delivery and stabilization.Deployment and testing focus.Product commercialized at end of first development funnel. Sold onemajor account with three sites.

1997External salesRelease 1.3.2

Product innovation and new markets research, mostly in the areas ofperformance.Total integration focus, 3rd party reporting tools, etc.Evaluations took place that resulted in major re-education of all staff.Sold two accounts and twenty sites with implementations of over a yeareach.

1998RubicoReleases 1.4and 2.0 (late1998)

Product portfolio enhancement from breakthroughs – reporting tools ver1 through 3, replication.Internationalization and benchmarking.Internal re-design phase 1.Sold seven major accounts with implementation times ranging twelve totwenty four months. One hundred and thirty three sites to beimplemented.

1999RubicoRelease 3.0 –ramp-up fromver 2, Release2 updates 2.1etc.

Independent and synchronized product delivery.Logistics and deployment.Product usability and stabilization.Application partitioning.Sold two major accounts with thirteen sites. Ready for next wave ofmajor account sales.

2000RubicoRelease 3.n –commercialrelease

International deployment and issue handling.Product architecture re-implementation moving towards ver 4.Dynamic application partitioning.Non-functional attribute focus i.e. performance, scalability, robustness,etc.

2001 – 2005 International volume deployment and logistics management –breakthroughs anticipated.Product stabilization and additional platform and derivative products.

Documents

Strategic Product Development...Project Management, defined the seven deadly sins of project management: • Mistaking half-baked ideas for viable projects. • Dictating unrealistic