16
 The Journal of Systems and Software 83 (2010) 1662–1677 Contents lists available at  ScienceDirect The Journal of Systems and Software  j ourn a l h om e p a g e :  www.elsevier.com/locate/jss Using Scrum to guide the execution of software process improvement in small organizations Francisco J. Pino a,, Oscar Pedreira b , Félix García c , Miguel Rodríguez Luaces b , Mario Piattini c a IDIS Research Group, Electronic and Telecommunications Engineering Faculty, University of Cauca, Calle 5 # 4–70 Popayán, Colombia b Databases Laboratory, Computer Science Department, Facultade de Informática, University of A Coru˜ na, Campus de Elvi˜ na, 15071 A Coru˜ na, Spain c  ALARCOS Research Group, Information Systems and Technologies Department, UCLM–INDRA Research and Development Institute, University of Castilla–La Mancha, Paseo de la Universidad 4, 13071 Ciudad Real, Spain a r t i c l e i n f o  Article history: Received 25 May 2009 Received in revised form 31 March 2010 Accepted 31 March 2010 Available online 15 June 2010 Keywords: Software process improvement Small software companies Model to guide improvements Agile methods Scrum SPI a b s t r a c t For software process improvement – SPI – there are few small organizations using models that guide the management and deployment of their improvement initiatives. This is largely because a lot of these models do not consider the special characteristics of small businesses, nor the appropriate strategies for deploy ingan SPIiniti ati vein thi s type of org aniz ati on.It shou ldalso benotedthat themodels whi chdirect improvement implementation for small settings do not present an explicit process with which to orga- nize and guide the internal work of the employees involved in the implementation of the improvement opportunities. In this paper we propose a lightweight process, which takes into account appropriate strategies for this type of organization. Our proposal, known as a “Lightweight process to incorporate improvements”, uses the philosophy of the Scrum agile method, aiming to give detailed guidelines for suppor ting the management and perfo rmanceof the incorpo rationof impro veme nt opport unitie s within processes and their putting into practice in small companies. We have applied the proposed process in two small companies by means of the case study research method, and from the initial results, we have observed that it is indeed suitable for small businesses. © 2010 Elsevier Inc. All rights reserved. 1. Introd uctio n Software process improvement – SPI – is a planned, managed and control led effort which aims to enhance the capa bilit y of the software development processes of an organization ( Krasner, 2001). SPIaims to und erstan d thesoftware pro ces s wi thi n an org a- nization and, using this knowledge base, it sets out to drive the implementation of change so that specic goals can be achieved (Coleman and O’Connor, 2008) (i.e. to meet its particular purpose). According to the systematic review on SPI in small companies presen ted in  Pino et al. (2008) ,  for some yea rs now the Sof t- wareEngineer ing communit y has beenshowingan ever- incr easin g interest in tackling SPI in small companies. They see this to be a strat egy whichcan be use d to increase sof tware pro duc t qua lit y, as well as to improve the productivity of software development. Even though there are different SPI standards in existence at presen t, the re is eviden ce tha t the maj ori ty of sma ll sof twa re org a- niza tion s are not adop tingthese standards (Coleman andO’Connor, Corresponding author. E-mail addresses: [email protected] (F.J. Pino),  [email protected] (O. Pedreira),  [email protected] (F. García),  [email protected] (M.R. Luaces), [email protected] (M. Piattini). 2008; Laporte et al., 2008 ).  The reasons for this are diverse, but acco rdin g to Lap orte et al.(2008) one imp ort antfact or is the nat ure of the standar ds, which do not provide adequa te guid ance for use in small software organizations. We might also observe that these standards, along with many proposals explicit to small com- panies, focus on offering models related to the best practices of processes (process reference models) and the manner of assess- ing process (process assessment models). They do not provide the explicit guidelines needed for the small companies to drive their SPI activities (improvement models suitable for small companies). An interesting nding of the systematic review presented in  Pino et al. (2008 ) is the perc enta ge of the use of these models in the pr i- mary studies analyzed: 71% used some process reference model, 42% used some assessment model and 24% used some improve- ment model. That is, small organizations have used models that direct the implementation of SPI activities in their initiatives for process improvement to a lesser extent. Regarding this issue, we consider it important to look into improvement models that may be sui tab le forsmallorga niz ati onsbecausesuch a mod el sho uld : (i) help to understand what to do when a small organization is inter- ested in carrying out SPI initiatives within their organization; and (ii ) pro vid e the gui del ine s that areneeded to org anize allthe act ivi - ties rel ate d to pro cess imp rovement as wel l as to con nec t theother models and elements involved. 0164-1212/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jss.2010.03.077

Scrum en Pequeñas empresas

Embed Size (px)

DESCRIPTION

Desarrollo de software

Citation preview

  • The Journal of Systems and Software 83 (2010) 16621677

    Contents lists available at ScienceDirect

    The Journal of Systems and Software

    journa l homepage: www.e lsev ier .c

    Using Scrum to guide the execution of software pin sma

    Francisco oda IDIS Research alle 5b Databases La na, Cac ALARCOS Rese ch andPaseo de la Un

    a r t i c l

    Article history:Received 25 MReceived in reAccepted 31 MAvailable onlin

    Keywords:Software procSmall software companiesModel to guide improvementsAgile methodsScrumSPI

    SPI theiraractof orgall see emose a

    strategies for this type of organization. Our proposal, known as a Lightweight process to incorporateimprovements, uses the philosophy of the Scrum agile method, aiming to give detailed guidelines forsupporting themanagement and performance of the incorporation of improvement opportunitieswithinprocesses and their putting into practice in small companies. We have applied the proposed process intwo small companies by means of the case study research method, and from the initial results, we have

    1. Introdu

    Softwareand controthe softwar2001). SPI anization animplement(Coleman aAccordingpresentedware Engininterest instrategy whwell as to im

    Even thopresent, thenizations ar

    CorresponE-mail add

    (O. Pedreira), FMario.Piattini@

    0164-1212/$ doi:10.1016/j.observed that it is indeed suitable for small businesses. 2010 Elsevier Inc. All rights reserved.

    ction

    process improvement SPI is a planned, managedlled effort which aims to enhance the capability ofe development processes of an organization (Krasner,ims to understand the software process within an orga-d, using this knowledge base, it sets out to drive theation of change so that specic goals can be achievednd OConnor, 2008) (i.e. to meet its particular purpose).to the systematic review on SPI in small companiesin Pino et al. (2008), for some years now the Soft-eering community has been showing an ever-increasingtackling SPI in small companies. They see this to be aich can be used to increase software product quality, asprove the productivity of software development.ugh there are different SPI standards in existence atre is evidence that the majority of small software orga-e not adopting these standards (Coleman andOConnor,

    ding author.resses: [email protected] (F.J. Pino), [email protected]@uclm.es (F. Garca), [email protected] (M.R. Luaces),uclm.es (M. Piattini).

    2008; Laporte et al., 2008). The reasons for this are diverse, butaccording to Laporte et al. (2008) one important factor is the natureof the standards, which do not provide adequate guidance foruse in small software organizations. We might also observe thatthese standards, along with many proposals explicit to small com-panies, focus on offering models related to the best practices ofprocesses (process reference models) and the manner of assess-ing process (process assessment models). They do not provide theexplicit guidelines needed for the small companies to drive theirSPI activities (improvement models suitable for small companies).An interesting nding of the systematic review presented in Pinoet al. (2008) is the percentage of the use of these models in the pri-mary studies analyzed: 71% used some process reference model,42% used some assessment model and 24% used some improve-ment model. That is, small organizations have used models thatdirect the implementation of SPI activities in their initiatives forprocess improvement to a lesser extent. Regarding this issue, weconsider it important to look into improvement models that maybe suitable for small organizations because such amodel should: (i)help to understand what to do when a small organization is inter-ested in carrying out SPI initiatives within their organization; and(ii) provide the guidelines that are needed to organize all the activi-ties related to process improvement as well as to connect the othermodels and elements involved.

    see front matter 2010 Elsevier Inc. All rights reserved.jss.2010.03.077ll organizations

    J. Pinoa,, Oscar Pedreirab, Flix Garcac, Miguel RGroup, Electronic and Telecommunications Engineering Faculty, University of Cauca, C

    boratory, Computer Science Department, Facultade de Informtica, University of A Coruarch Group, Information Systems and Technologies Department, UCLMINDRA Resear

    iversidad 4, 13071 Ciudad Real, Spain

    e i n f o

    ay 2009vised form 31 March 2010arch 2010e 15 June 2010

    ess improvement

    a b s t r a c t

    For software process improvement the management and deployment ofmodels do not consider the special chdeploying an SPI initiative in this typeimprovement implementation for smnize and guide the internal work of thopportunities. In this paper we propom/ locate / j ss

    rocess improvement

    rguez Luacesb, Mario Piattini c

    # 470 Popayn, Colombiampus de Elvina, 15071 A Coruna, SpainDevelopment Institute, University of CastillaLa Mancha,

    there are few small organizations using models that guideimprovement initiatives. This is largely because a lot of theseeristics of small businesses, nor the appropriate strategies foranization. It should also benoted that themodelswhichdirectttings do not present an explicit process with which to orga-ployees involved in the implementation of the improvementlightweight process, which takes into account appropriate

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1663

    With all that inmind,wehavedeveloped an Improvement frame-work in the context of the COMPETISOFT project (Oktaba et al.,2007). COMPETISOFT offers a strategy for supporting software pro-cess improvement in small organizations, and its Improvementframework pguide the exInitially, wetives in sma(Pino et al.work. This p(b) diagnoscuting improffers a setgram, activprocess impprocess in aposal for guwas not yestudies thaconclude th

    For smallcesses thediagnosin

    Small orgout improfalls mainactivitiesments).

    It is necesguidelinedo with thnities. Furorganizat

    We havabove. Whaods, for asand Bastarroffered by tOur proposwhich usesguidelines fimprovememulating anobjective oof improvemactivity) wiused the Scbecause it pmanagemen(Abrahamssmethods, annext section

    The purpprocess. It aproposed pies. These owhich usedused PfemCcesses of ththe diagnosFollowing tbackgroundfactors in thout, and the

    the Improvement frameworkofCOMPETISOFTandgivesanoverviewof the PmCOMPETISOFT process. The PfemCOMPETISOFT processand how Scrum supports the activities of formulating and execut-ing improvement is shown in Section 4. Section 5 presents two

    udiets ou

    kgro

    devere orzatiorganLaponheiompe exinclu

    softtrongbetwareturets encally,malicatiodo n

    theyirede comecon

    chars ofapp

    Anotll orthectorsplemses aand son inith t(butand

    l.

    all o

    ordinhat eandese caccoreir are mzatioracticmaladvstarrt sorovides improvement practices, strategies and tools toecution of improvement initiatives in small companies.dened a process to manage and lead the SPI initia-ll organizations step-by-step, called PmCOMPETISOFT

    , 2009a), as the backbone of this Improvement frame-rocess describes ve activities: (a) initiating the cycle,

    ing the process, (c) formulating improvements, (d) exe-ovements and (e) revising the cycle. PmCOMPETISOFTof well-dened process entities (such as activity dia-

    ities, roles, work products and templates) to the wholerovement lifecycle. From an early application of thispilot case study we observed that it was a good pro-iding SPI initiatives in small organizations, but that itt sufcient. This rst application (and other previoust we have performed (Hurtado et al., 2008)) led us toat:

    companies it is not easy to incorporate into their pro-improvement opportunities that are uncovered in the

    g of process activity.anizations have difculties in managing and carryingvement activities inwhich the greatest amount of effortly upon the organization (in PmCOMPETISOFT theseare formulating improvements and executing improve-

    sary to develop a proposal which offers more detaileds to address specically those activities which have toe formulation and execution of improvement opportu-thermore,what is proposed should be suitable for smallions.

    e developed a proposal to tackle the issue describedt we have proposed takes into account agile meth-according to Coleman and OConnor (2008), Hurtadoica (2006) and Abrahamsson et al. (2002) the strategieshese methods are appropriate for small organizations.al thus describes a process called PfemCOMPETISOFT,the Scrum agile method, seeking to provide detailedor supporting the management and performance of thent iteration (which is made up of the activities of for-d executing improvement from PmCOMPETISOFT). Thef PfemCOMPETISOFT is to support the incorporating

    ent opportunities (found in the diagnosing of processthin the processes of the small organizations. We haverum agile method for the denition of this process,rovides support for project management, emphasizest values and practices and focuses on small teamson et al., 2002). A discussion onwhywehave used agiled especially Scrum, in our proposal, is presented in the.ose of this paper is to present the PfemCOMPETISOFTlso describes our experience of the application of therocess in two small organizations, through case stud-rganizations were involved in an improvement project,the Improvement framework of COMPETISOFT. They

    OMPETISOFT to guide the incorporation into their pro-e improvement opportunities that were uncovered inis activity. The paper is organized into six sections.his introduction, Section 2 of the paper presents theon characteristics of small companies and SPI successese organizations; why we used agile methods is setrelated works are also described. Section 3 describes

    case stpresen

    2. Bac

    Tosoftwaorganismall o

    In (Wangesmall cwe havwhich

    Theirare stion

    Theystrucboos

    Typiinformun

    Theyandrequ

    Thesited

    Thecedurerequire2000).in smawhichcess fathe improcestoringreturnsible win SPIipants,contro

    2.1. Sm

    Accargue tativitythat thmore,with thship, aorganiagile pto the sto takeand Basuppors where this process was applied and nally, Section 6r conclusions and future lines of work.

    und

    lop process improvement proposals suitable for smallganizations, the special characteristics of this type ofn and the success factors of process improvement inizations should be taken into account.rte et al., 2008; Horvat et al., 2000; Richardson andm, 2007; Hofer, 2002) the special characteristics ofanies are analyzed and discussed. From these studiestracted some characteristics of the small organizations,de the following:

    ware development is driven by light processes whichly human-oriented and there is constant communica-een the project members and the customer.usually dynamic and exible, with a at organizational(not traditional), in a free-ow management style thatterprising and innovative spirits.they carry out their management process through

    mechanisms, based on face-to-face relationships (com-n, decision-making, problem resolution, etc.).ot have enough staff to develop specialized functions,have little or no room in their budget for buying theexpertise.panies are economically vulnerable and they have lim-

    omic resources.

    acteristics of these organizations, such as specic pro-work and specic relationship between employees,ropriate management of SPI projects (Horvat et al.,her feature is that the success of process improvementganizations is also directly related to the manner inimprovement project is managed. Some of these suc-, which can be found in Pino et al. (2008), are: guidingentation (or deployment) of SPI by means of specicnd the combination of different approaches, the moni-upervision of the SPI project, the attainment of a rapidvestment, the denition of what SPI objectives are fea-he resources available, the participation of employeeston-up), suitable communication between SPI partic-that SPI should be based upon learning, not upon

    rganizations and agile methods

    g to Coleman andOConnor (2008), small organizationsach project and situation is new to them and that cre-exibility are important capabilities; they also afrmapabilities can be supported by agile methods. Further-ding to Coleman and OConnor (2008), agile methods,dvocacy of self-empowered teams and shared owner-ore associated with the style of management of smallns. This is because this type of company considers thates canbeapplied in their softwareprocessbasicallydue

    l initial investment required and because it allows themantage of competitiveness in their personnel (Hurtadoica, 2006). In our opinion, the agile methods also canme success factors involved in process improvement

  • 1664 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    in small organizations, as well as several special characteristics ofthese companies.

    As Abrahamsson et al. (2002) point out, some agilemethods thatfocus on small teams are: Agile modeling (Ambler, 2002), ExtremeprogrammiThomas, 20methods devides a framon describito do withScrum is noit is a frameand techniqmanaging pa specic kn

    Bearingagile methothe introduporation oforganizatioorganizatioopment ofScrum focument value(e.g. PmCOMsuccess factout before.

    2.2. Related

    With rement implehave emerMESOPyME2001), PROet al., 2006medium enprocess moas a processtained. Theis an abstrprojects bylearning anthis experieprojects. Mand effort iaction packapproachesistics of smaeffective esinitiative. Hprocess witinvolved, toof the impr

    When comodel to sstudies pre(2000). Theprise, and itbe adaptedthe improvthroughoutthe businesindicated thupon the oson for this

    the resources needed to carry out the tasks. On the other hand, thesecond study took place in a small enterprise, and discovered thatto improve processes in small organizations a structured modelneeds to be used to organize the process, to adjust the model to

    rticuemeed rodem

    IDEAe sp

    ablerganns aver, nwhic

    to oh regentin Saork pI witesenethorocelantan agrryiness raspeforto trovesedies of thein S

    es suteamsizeses otheremeentaandng imstratployeme

    rove

    Impandup t

    s MeMPEll orgp thethanhat hr tore ind toeaseng (Beck, 1999), Pragmatic programming (Hunt and00) and Scrum (Schwaber, 1995). However, of the fourscribed previously, only the Scrum agile method pro-ework for managing projects, whilst the others centerng practices, activities and work product techniquessoftware development. As stated in Schwaber (2009),t a process or a technique for building products; rather,work within which you can employ various processesues. This feature allows combining the framework forrojects described by Scrum with technical processes ofowledge area, in our context, for instance, the SPI area.in mind (i) the aspects previously described as regardsds and Scrum, and (ii) the particular needs set out inction about developing a proposal to support the incor-improvement opportunities into the processes of smallns, we believe that Scrum is suitable for this kind ofn and that it should be borne in mind in the devel-our proposal. This conviction comes from the fact thatses on small teams, it gives most emphasis to manage-s and practices, it can be combinedwith other processesPETISOFT), and that it is themethodwhich can support

    ors for process improvement such as those we have set

    work

    gard to research on models that direct improve-mentation for small organizations, various proposalsged in recent years. These include, amongst others,(Calvo-Manzano et al., 2002), IMPACT (Scott et al.,

    CESSUS (Horvat et al., 2000), ASPE-MSC (Wangenheim) and the application of the IDEAL model to small andterprises. The PROCESSUS framework is based on thedeling paradigm, in which each procedure is dealt with, which is dened, established, implemented andmain-IMPACT framework is based on the idea that the processaction of the practices carried out in many differentmany different people. It therefore permits continuald the improvement of the process, precisely throughnce that has been gained by a lot of people in a lot ofESOPyME has as its focal point the reduction of timen the implementation of SPI by using the concept ofages as a base. ASPE-MSC integrates and adapts existingon establishment of software process to the character-ll companies. The goal is to support a cost-efcient and

    tablishment process in these companies as part of a SPIowever, none of these proposals presents an explicith which to guide the internal work of the employeesenable them to manage and carry out the deployment

    ovement activities in the small organization.nsidering the application and adjustment of the IDEALmall and medium enterprises, it is worth noting thesented in Casey and Richardson (2004) and Kautz et al.rst study was carried out in a medium-sized enter-identied that the IDEAL models acting phase should

    to the organizations characteristics. In the acting phaseements discovered are created, piloted, and deployedthe organization. The study also highlighted that ifs had carried out this phase as the IDEAL model hadat it should, then it would have had a negative effect

    rganizations process improvement initiative. The rea-is that it did not take into account either the time or

    the paimprovumentstudiesby thefull-timbe suitthese ofunctioMoreodetailapplied

    Witimplemsentedrst wout SPone pragile mupon p(or impuse ofand caof proc

    Theprocessdetailsing imphave uactivitmany ocussedprovidsmallemphapracticwith omanagincremgratedexecutioffer athe emimprov

    3. Imp

    ThemodelmakeFig. 1.

    Thithe COin smamakeufewerwork tpeculiasoftwaadapteto incrlar conditions of the organization, and to perform thent activities as a project with clearly assigned and doc-les, responsibilities and resources. What is more, bothonstrate that, if the improvement initiatives suggestedLmodel are to be applied, there is a need for at least oneecialist in process improvement. That does not seem tofor the characteristics of small companies, as many ofizations do not have enough staff to develop specializednd their budget forbids buying this required expertise.either of these studies explicitly presented a process in

    h was adapted to the IDEAL model and which could bethers small organizations.ard toworks related tomodels that direct improvementation and agile methods, we should mention those pre-lo and Abrahamsson (2007) and Khknen (2005). Thisroposes an iterative improvement process for carryinghin agile software development teams, and the secondts a lifecycle model for an SPI project which deploys and in a single project. That is to say, these works focusss improvement in projects or teams which work with) agile methods. Our proposed process, however, makesile method such as Scrum to support the managementg out of the improvement process, following any typeeference paradigm.cts set out above justify the denition of a Lightweightimprovement incorporation, which would add morehe activities of formulating improvement and execut-ments of the PmCOMPETISOFT process (see Fig. 1). Wethe Scrum agile method to manage and carry out thef formulation and execution of improvement, in whichorganizations employees are directly involved. As dis-

    ection 2.1, our reason for doing so is that this methodpport for project management and it is suitable fors. Furthermore, apart from those advantages, Scrummanagement values and practices, without including

    n technical issues. This allows Scrum to be integratedmethods and processes, thus offering them more agilent or deployment. Taking into account this aspect, thel and iterativeprocessproposedbyScrumhasbeen inte-tailored to the activities of formulating improvement andprovementsof thePmCOMPETISOFT. Thismeanswecanegy which is useful and practical and through which allees will become involved and take part in the processnt to which they are related within the organization.

    ment framework of COMPETISOFT

    rovement framework, along with the Process referencethe Evaluation model, are the three components thathe Methodological Framework of COMPETISOFT; see

    thodological Framework is the strategy developed byTISOFT project (Oktaba et al., 2007) for supporting SPIanizations, since thevastmajorityof the companies thatLatinAmerican software industry are small-sized,with50 employees (MBI, 2004). In this respect, the frame-as been developed took into account the characteristicsthis type company, thereby offering the Latin Americandustryaprocess improvementapproachwhichhasbeenits particular context. The COMPETISOFT project seeksthe level of competitiveness of small software organi-

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1665

    ologic

    zations throprocesses oand by meaponents ofthese is pre

    The Procegroup forIn this restices, whithey are t

    The Evalucution ofsmall orgnizationaEvaluation15504 Pa

    The Imprstrategies(Pino etment prois the baImprovemcess asseswhich exware procwith ISO/prioritizatduring thsmall comfor improvprovidescuting imsupport thto supporand the imframewor

    straTISO

    Imprich femeTISOove

    ted i

    prov

    COMFig. 1. Overview of the improvement framework and method

    ugh the improvement and certication of the softwaref these organizations, using the proposed componentsns of the Methodological Framework. Regarding com-the Methodological Framework, a brief description ofsented below:

    ss reference model provides a suitable and useful processincreasing theprocess capability in small organizations.pect, each of these processes offers best software prac-ch should be considered by this kind of organization ifo improve their processes.ation model denes the necessary elements in the exe-

    TheCOMPEof thecle, whimprovCOMPEgive anpresen

    3.1. Im

    Pm

    a formal assessment, the objective being to give theanizations a rating of their process capability and orga-l maturity. In this sense, we suggested that the dened

    model must use and be in conformance with ISO/IECrt 2 (ISO, 2004a) and Part 7 (ISO, 2008).ovement framework provides improvement practices,and tools to support SPI initiatives in small companies

    al., 2009b). This framework denes: (i) an improve-cess called PmCOMPETISOFT (Pino et al., 2009a), whichckbone as well as the component integrator of theent framework; (ii) a methodology for software pro-sment called METvalCOMPETISOFT (Pino et al., 2010),tensively describes the activity of diagnosing the soft-esses in small organizations and which is conformanceIEC 15504-2; (iii) a Strategy for process selection andion, which supports the selection of critical processese implementation of a process improvement project inpanies (Pino et al., 2009c); (iv) a Lightweight processement incorporation called PfemCOMPETISOFT, whicha extension of the activities of formulating and exe-provements of PmCOMPETISOFT; and (v) a Tool set toe improvement process, which provides software toolst the dissemination, the knowledge, the managementplementation of the components of the Improvement

    k.

    ISO/IEC 155implantatiomentalman

    Early andof visiblethis imprstage andminimiza

    Ongoing aorganizateffort to vobjectiveorganizat

    Elementalactivitieswhich are

    Effective gment ofthe variouhighlightproject.

    Ongoing lment to thaware thaal framework of COMPETISOFT.

    tegy for process selection and prioritization, METval-FT and Tool set to support the improvement processovement framework are beyond the scope of this arti-ocuses on the description of the Lightweight process fornt incorporation. To help to make the proposed Pfem-FT process easier to understand, we will now go on torview of PmCOMPETISOFT (a complete description isn Pino et al., 2009a).

    ement process PmCOMPETISOFT

    PETISOFT is a light process inuenced by the IDEAL,

    04-4 (ISO, 2004b) and Scrum models, which guides then of an improvement initiative in an iterative and incre-ner, andwhich plans to satisfy the following principles:

    ongoing achievement of improvement: the generationshort-term results implies that the actors involved inovement process see the fruits of their work at an earlythus continue to be motivated. This also permits the

    tion and control of risks in the improvement project.nd rapid process diagnosis: the ongoing diagnosis of theions processes and of the improvement process, in anerify whether the improvement project is fullling itsof raising the level of capability and efciency of theions processes.process measurement: the carrying out ofmeasurementthrough the systematic use of process base measuressuitable for this type of organizations.roup collaboration and communication: the establish-

    communication and collaboration strategies betweens actors involved in the improvement process project,

    ing the importance of each persons role within this

    earning: offering training in software process improve-ose involved in the improvement project,making themt their support, contributions and active participation

  • 1666 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    diagr

    are necesalso encolearnt arehence be

    PmCOMprocesses aalso to assiby focusingimplementoverview oFrom this d

    There areResponsibment grouEvaluator

    Each of timprovemactivities:ing improcycle: theof improv

    Theworkmentplanreport.

    4. Using Scexecuting o

    As a memanagemeand this isods and prevolutive pthat, as thecuting imprprioritizedgeneral impsee Fig. 2),able for thein small buout these a

    lity lhat bolutirporar carrng im

    e lig

    s seceightTISOrolesscripelemof t

    t in

    Purpoofferformroveolveresport inFig. 2. PmCOMPETISOFT activity

    sary for the project to be a success. These actors areuraged to reect upon the work carried out; the lessonsthereby reinforced and the improvement project can

    improved and adjusted.

    PETISOFTs purpose is to improve the organizationss appropriate to its particular business objectives. It isst it in carrying out its software improvement processon small enterprises, dening a guide with which toa step-by-step process improvement. Fig. 2 shows anf PmCOMPETISOFT by means of an activities diagram.iagram we can see that:

    ve roles: Management improvement group (MIG),le for process improvement (RPI), Process improve-p (PIG), Responsible for process or Participant (RP) and(EV).he improvement cycles is formed by one or moreent iterations. An improvement cycle consists of veInitiating the cycle, Diagnosing the process, Formulat-

    vements, Executing of improvements and Revising theactivities of Formulating improvements and Executing

    ements make up the improvement iteration.products are: Improvementproposal, General improve-, Improvement implementationplan and Improvement

    capabition. Tand evto incoline foexecuti

    4.1. Th

    ThilightwCOMPEgram,the deScrumto eachaccoun

    4.1.1.To

    ities ofan impees invto thetake parum to support the activities of formulating andf improvements

    thod, Scrum (Schwaber, 1995) emphasizes values andnt practices and does not include technical practices,just what gives it its capacity to complete other meth-ocesses. This agile method offers an incremental androcess for the development of products. We could add

    PmCOMPETISOFT activities of formulating and exe-ovements are iterative and incremental, initiated byimprovement opportunities which are described in therovement plan (obtained from the diagnosing activity;the strategies proposed by Scrum are in our view suit-management and deployment of both these activitiessinesses. It is important to clarify that after carryingctivities, the desired product is a process with a higher

    opportuniti

    4.1.2. ObjecTo supp

    the processment of a band executinphilosophie

    4.1.3. ActivFig. 3 sh

    process.

    4.1.4. ActivOne or

    include theimproved (improvemeam.

    evel and this is put into practice within the organiza-eing the case, we used and adapted the incrementalve process of Scrum to develop the Lightweight processte improvements, which gives a more detailed guide-ying out the activities of formulating improvement andprovements of PmCOMPETISOFT.

    htweight process to incorporate improvements

    tion provides a detailed description of this proposedprocess following the process pattern established byFT, which includes purpose, objective, activity dia-, activities, work products and tools support. Withintion of activities, roles and work products we show theents (described in Schwaber (2009)) that are relatedhese. The goal is clarify how we took this method intothe development of PfemCOMPETISOFT.

    sea guideline for managing and carrying out the activ-ulating improvements and executing improvements (i.e.ment iteration). This guidelinewould allow the employ-d in the SPI initiative of the small organizations (mainlynsible for process improvement) to manage, deploy andthe execution of the incorporation of the improvement

    es found and with which they have some relationship.

    tiveort the incorporating of improvement opportunities ines of small organizations, by means of the establish-reak-down of the activities of formulating improvementsg improvements (improvement iteration), following thes proposed by Scrum.

    ity diagramows the activity diagram of the proposed lightweight

    itiesmore improvement iterations must be carried out toimprovement opportunities within the processes to beimproved processes). To manage the complexity of thent opportunities found in the SPI initiative, these are

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1667

    to inc

    grouped intargets of tand develodently. Inincrementtion of specincrement,tice of the iof ve activcase, Executcess and Prebelow:

    Planningment iteraassociatedestablisheof the gensub-sectioagement i(2009c) csince thisRelease p

    Designingsatisfy thities assoprocess ior days)and assignizationregardingteamresp(this activbacklog).

    tingrumprovlly 1ncorps ofFig. 3. Diagram activities of the lightweight process

    improvement cases according to specic improvementhe organization. An iteration allows the managementpment of an improvement case to advance indepen-this respect, each improvement iteration creates anin the process capability by means of the incorpora-ic improvement opportunities. In order to reach this

    Execuof Scof imusuathe icesseiteration entails thedenition and theputting intoprac-mproved process. Each improvement iteration consistsities: Planning the iteration, Designing the improvementing the improvement Sprint, Presentation of improved pro-sentation of next iteration. These activities are presented

    the iteration: the objectives of the current improve-tion are described. The improvement opportunities setwith each improvement case and the other items

    d in the preliminary improvement plan (which is parteral improvement plan, see Fig. 2 and work productsn) are reviewed, validated and approved by the man-

    mprovement group. The strategy presented in Pino et al.an be used to prioritize the improvement opportunitiesis a key task in this activity (this activity is related to

    lanning meeting from Scrum).the improvement case: the strategies with which to

    e improvement opportunities are identied. The activ-ciated with the strategies for carrying out the currentmprovement are dened and an estimate (in hoursof these tasks is made. These activities are denedned to the members of the work team in the orga-and an agreement is reached with the work teamthe development of the improvement iteration. Theonsible for this activity is theprocess improvement groupity is related to Sprint planning meeting and Sprint

    have tailothe improin this spimprovemimplemencompleteteam respsible for pror participcompliansprint, an Plan exe

    goal is dinitial pment ocase tharespectmodel sbeing imon puttincorpoimprovThe spein practrelated

    Daily wties: theperformorporate improvements.

    the improvement sprint: bearing in mind the concepts sprint, we consider an improvement sprint to be a setement activities conducted over a pre-dened period,4 weeks, which is used to implement and monitororation of the improvement opportunities within pro-the small organization. As can be seen from Fig. 3, we

    red the tasks of the Scrums sprint to the needs ofvement sprint, thereby providing the people involvedrint with a work structure that would place the desiredent opportunities within the processes. Through thetation of several improvement sprints, the process is

    ly dened and put into practice in the organization. Theonsible for these tasks is made up of the person respon-ocess improvementand theperson responsible for processant. The tasks dened in the improvement sprint are

    t and consistent with those described by the Scrumsd they are:cution of improvement sprint: The improvement sprintsened. The rst sprintmust have as its goal to dene anrocess version that incorporates the selected improve-pportunities and which belongs to the improvementt is being addressed with the current iteration. In this

    , the best practices described in a process referencehould be borne in mind to dene the process that isproved. The next improvement sprints should focus

    ing the latest process version into practice and/or onrating in the denition of this process the rest of theement opportunities that have not been addressed yet.cic tasks to dene the improved process or to put itice are identied, estimated and selected (this task isto Sprint planning meeting and Sprint backlog).ork, Manage problems, Update improvement opportuni-specic tasks to reach the improvement sprint goal areed by the work team (related to development work).

  • 1668 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    Improvement sprint meeting: an inspection of the progresstowards the sprint goal is performed in order to regulateimprovement tasks. Process improvement is an activity involv-ing the whole organization, permeating all of the ongoing dailywork ofan activrequiresprintming on(this tas

    Improvework cais assesit is imis to incincreasenition osatisfy tfor thegestionreceivedment spmust beations (

    Improvethe woin theobtainesprint aimprovimprovment opimprovmake itsprint aidentibeing imment opsprint).

    Presentatiorganizattionalize

    Presentatiannounce

    The worDesigning thment implemother activi

    4.1.5. RolesBoth pro

    process desof each rolimprove sofimprovemedenitionaSOFT, it isto Scrum fsible for prresponsibilagement imOn the othework produlightweight

    General improvement plan: this is the input work product to thelightweight process and it is a document is made up of two parts:(i) the assessment report which describes the state and analysisof the processes, and (ii) the preliminary improvement plan which

    es: tmproft of

    manat oppessesovem-leves arelog).nt itovemdesigh incovemon otionto m

    ecorovemovedts attmmelusioof th

    Toolsfac

    a procal suableroceemenedi

    te dichin C

    portprovwhicctivitecutited kcatiotande Imp

    rresp

    ed onew os entScru

    ng mScrumthe Srodubackve ethesedowimprthe software development in the company. It is alsoity that complements those day-to-day tasks and it

    s an additional effort. That being so, we propose that theeeting should take place once or twice aweek, depend-

    the progress and needs of the improvement iterationk is related to daily Scrum).ment sprint review: thegoal is to showthe improvementrried out during the current sprint. Theworkperformedsed against the sprint goal. To carry out this review,portant to consider that the objective of each sprintrease the improved process capability. This capabilityis given by incorporating best practices in the de-

    f the process and its implementation. These practiceshe improvement opportunities that have been selectedcurrent sprint. On the other hand, in this meeting, sug-s about the lessons learnt and problems found are also, along with the most relevant results of the improve-rints. This allows for feedback fromthese sprints,whichtaken into account for the following improvement iter-this task is related to Sprint review).ment sprint retrospective: a retrospective analysis ofrk carried out by team members that were involvedlast improvement sprint takes place. The feedbackd supports ongoing learning and improvement of thend iteration. In our context the ongoing learning andement involve: (i) to improve the process used foring processes, and (ii) to identify new process improve-portunities. This task entails revising and adjusting the

    ement process used, in this case PfemCOMPETISOFT, tomore effective and enjoyable for the next improvementnd iteration. Furthermore, actionable improvement

    ed from the daily work related to the process that isproved or other processes is added to the improve-portunities group (this task is related to retrospective

    on of improved process: the improved process of theion is presented, and the manner in which to institu-it within the entire organization is determined.on of next iteration: the new improvement iteration isd.

    k carried out in the activities Planning the iteration ande improvement case should be registered in the improve-entation plan. In addition, the work carried out in the

    ties is registered in the iteration improvement report.

    and work productscess entities are derived from PmCOMPETISOFT. Thiscribes a competence set needed to perform the worke. These competences involve the skills necessary totware process and to increment their capability, such asnt project execution andmanagement, and assessment,ndexecutionof softwareprocess. TousePfemCOMPETI-necessary to consider additional competences relatedor some roles. In this respect, the individual respon-ocess improvement must have knowledge about theities of the ScrumMaster, and one person from the man-provement group must play the role of Product owner.r hand, in the following bullet points, we describe thosects which have a direct relationship with the proposedprocess:

    denthe ia drariskmenproc

    ImprhighcasebackcurreimprThiswhic

    Imprmatiiteratiedare rimprimprmenrecoconcview

    4.1.6.One

    directnologiwill enment pimprov2.0 andgeneraand whviewedto supthe im2008),of all aand exgeneraan eduundersand th

    4.2. Co

    Basoverviproces

    TheplanniDailyroles (work pSprintwe hasincebreak-latinghe improvement opportunities and their prioritization,vement cases, the number of improvement iterations,the general planning (it includes measures, training,gement and chronogram sketch). The list of improve-ortunities is the requirements tobe incorporatedwithinthat are being improved (related to Product backlog).ent implementation plan: a document which denes thel activities which must take place if the improvementto be created, designed and executed (related to SprintThis document contains: the probable planning of theeration and a draft design of the improvement case (orent opportunities) that is addressed by this iteration.n involves a denition of the process to be improved,orporates the selected improvement opportunities.ent iteration reports: this report includes relevant infor-n the performance and assessment of the currentand its sprints. In this respect, the specic tasks iden-eet the objectives of the different performed sprints

    ded (related to Sprint backlog). The analysis of theents brought into the organizations processes, theprocess documentation, the effort involved, achieve-

    ained, lessons learnt, an iteration post mortem review,ndations for improvement iteration adjustment, andns about the improvement iteration from the point ofe organization are recorded too.

    supporttor that may assist small companies to successfullycess improvement initiative is that of providing tech-pport to such companies through software tools whichthem to guide, implement and manage their improve-ss (Pino et al., 2008). Lightweight process to incorporatetshas thereforebeendescribedwith the standardSPEM

    ted with the EPF Composer (Eclipse, 2007), in order toocumentation in a standard format which is updatedis available to organizations through the Web. It can beYTED (2008). We have also developed a couple of toolsthe person responsible for process improvement duringement cycle, these are: (i) GENESIS (Hernndez et al.,h can be used for themanagement and implementationies of an improvement cycle (including the formulationon of improvements), aswell as in the administration ofnowledge; and (ii) HEPALE! (Cruz et al., 2009), which isnal tool to support dissemination-getting to know andthe processes described by the Process reference modelrovement framework.

    ondence between PfemCOMPETISOFT and Scrum

    the description presented in the previous section, anf the correspondence between the PfemCOMPETISOFTities and the Scrum elements is shown in Table 1.m framework consists of a set of activities (the Releaseeeting, the Sprint planning meeting, the Sprint, the, the Sprint review, and the Sprint retrospective),crumMaster, the Product owner and the Team) andcts (the Product backlog, the Release burndown, thelog, and the Sprint burndown). From these elements,mphasized activities for dening PfemCOMPETISOFT,

    directly support the goal of this process, giving an of the improvement iteration (activities of formu-ovements and executing improvements) in managing

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1669

    Table 1Relationship between PfemCOMPETISOFT process entities and Scrum elements.

    PfemCOMPETISOFT Related Scrums element

    Process entity Key description related to Scrum

    Improved pr ut into

    Improvemen tices d

    Planning the ning,of the

    ities aDesigning th igh-le

    rovemo thefor t

    Improvemen r a prrationzation

    Plan executi re decess oks des

    blisheImprovemen l is pe

    ss imtwice

    Improvemen mentessed

    Improvemen ers, prd itera

    ResponsibleManagemen role oPreliminary t opp

    casesImprovemen ed to

    and carryinin the procand work pthese elemements, in oas in PmCOorder to guprocesses.

    5. Applicat

    The Actiused for thponents deAccording tA-R involvethe knowleactivities wleast the fotion, and rethe componincluding Pfoped by mthese compproblem-solsupport theproblem-solallowedus tthe case stucycle, to disocess Process with higher capability level that is porganization.

    t opportunities set Improvement opportunities are all best pracimproved.

    iteration Goals and a draft plan (this is mid-term planprobable delivery date, work team and cost)established.The highest priority improvement opportun

    e improvement case Strategies are determined, along with their himprovement sprints so as to satisfy the impThese activities are estimated and assigned tAn agreement is reached with the work teamimprovement iteration.

    t sprint Set of improvement activities conducted oveused to implement and monitor the incorpoopportunities within processes of the organi

    on of improvement sprint Goals and scope of the improvement sprint aThe specic tasks to dene the improved proidentied, estimated and selected. These tasactivities perform.A short-term planning for these tasks is esta

    t sprint meeting A progress inspection towards the sprint goaimprovement tasks. In the context of a procepropose that this meeting take place once or

    t sprint review What was accomplished during the improveThe success of the improvement sprint is assperformed against the sprint goal.

    t sprint retrospective How the sprint went as regards team membThis feedback on the improvement sprint anongoing learning and improvement.

    for process improvement This person has the role of ScrumMaster.t improvement group One single person from this group plays theImprovement Plan This work product contains the improvemen

    prioritization and grouping in improvementt Iteration Reports This work product contains the tasks identig out the incorporation of improvement opportunitiesesses of the small organizations. Regarding the rolesroducts, Table 1 establishes a relationship betweennts from PfemCOMPETISOFT and Scrum. For these ele-ur proposed process we have kept the same namesMPETISOFT (without using the names from Scrum), inarantee the coherence of these elements between both

    ion of the proposed lightweight process

    on Research (A-R) and Case Studies methods have beene denition, renement and application of the com-veloped in the context of the COMPETISOFT project.oMcKay andMarshall (2001) and Chiasson et al. (2009),s a research cycle and a problem-solving cycle, in whichdge is applied and discovered interactively betweenith different goals and outcomes. Each cycle includes atllowing activities: problem diagnosis, action interven-ective learning (Avison et al., 1999). In this respect,ents of the COMPETISOFTs improvement framework,emCOMPETISOFT (theoretical knowledge),were devel-eans of the execution of various research cycles, andonents were applied by means of the execution of theving cycles. We have used the case study method toexecution of the action intervention activity from the

    ving cycle. The execution of the problem-solving cyclesodiscoverpractical knowledge,whichwas registered indy reports. This knowledge inuenced thenext researchcover new theoretical knowledge, which was included

    in the comnew, reneframework

    On the othe projectresearcherscation, anincludes cotions. As mand it has aand the comof the Imprthe action iA-R, the res1999). Thatthe small cothe case stuPfemCOMPby followinBrereton et

    It is impstudies of PPmCOMPETa set of proopportunitifrom PmCOing the proimprovemethe processSOFT. Thuspractice within the Product

    esired in the processes to be Product backlog

    by including the major risks,improvement iteration are

    nd cases are determined.

    Release planningmeeting

    vel activities to attempt theent opportunities.

    members of the work team.he carrying out of the

    Sprint planning meetingSprint backlog

    e-dened period, which isof the improvement.

    Sprint

    ned.r to put it into practice arecribe how the high-level

    d.

    Sprint planning meetingSprint backlog

    rformed to regulateprovement project wea week.

    Daily Scrum

    sprint is shown.by contrasting the work

    Sprint review

    ocess and tools is analyzed.tion should support their

    Sprint retrospective

    ScrumMasterf Product owner. Product ownerortunities, their.

    Product backlog

    meet the sprints objectives. Sprint backlogponents of the Improvement framework, thus creatingd and improved versions of the components of this(including PfemCOMPETISOFT).ther hand, by using the A-R research method we divideparticipants into two groups: the rst is made up offrom various universities and an national body of certi-d the second is called the critical reference group andmputer professionals from small software organiza-entioned previously, the application of A-R is iterativellowed us a continual feedback between the researcherspanies involved, so we can polish up the components

    ovement framework. With regards to the execution ofntervention activity of the problem-solving cycle fromearchers applied the empirical variant (French and Bell,is, PfemCOMPETISOFTwas applied by the researchers inmpanies (critical reference group) bymeans of the use ofdy research method. In order to validate the proposedETISOFT process we have conducted two case studiesg the protocol template for case studies presented inal. (2008) (see Fig. 4).ortant to highlight that in order to carry out the casefemCOMPETISOFT, it has also been necessary to useISOFT, due to the fact that, to apply PfemCOMPETISOFT,cess improvement opportunities is needed and thesees are discovered by means of the diagnosis activityMPETISOFT. That is to say, by means of the diagnos-cess activity of PmCOMPETISOFT, we determined thent opportunities, and these have been incorporated ines of the small organizations by using PfemCOMPETI-, that proposed process has been used to conduct the

  • 1670 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    e COM

    activities ofIn this respeapplied alothe improvprises involall the activtion 5.3 weOur next steformed in tdata collect

    5.1. Design

    The maicess to incordriving thesoftware enthese caseenable smaof the imprprocess enaopportunitithe effort ofsmall comp(2003), thecases holiof two smacess to incorthe researcthe activitiebility levelbe improvedescribed bare the implating and eimproved b

    For the awith two cof the COMIn order toto these twsome featurproducts, b

    lComs of eis mlopmmainganizmatiAlthmerarebee

    thatentedasinstablcomprtanowthlComty) wbineslopmtraderchists oFig. 4. Application of A-R and case studies to th

    formulating improvements and executing improvements.ct, this section showshowPfemCOMPETISOFThasbeenngside PmCOMPETISOFT to give a complete vision ofement initiatives performed by the two small enter-ved in the case studies. Although this section describesities performed in the improvement initiatives, in Sec-emphasize the results of applying PfemCOMPETISOFT.p in this paper will be to describe the case studies per-

    erms of design, analysis unit, subjects, eld procedure,ion, and analysis.

    , analysis unit and subjects

    n research question addressed is: Is the Lightweight pro-porate improvements suitable (useful and practical) forformulation and execution of improvements in smallterprises? Additional research questions addressed by

    studies are: (i) Does the proposed lightweight processll companies to involve their employees in theexecutionovement activities? (ii) Does the proposed lightweightble small companies to incorporate the improvementes in their processes and put them in practice? (iii) Isapplying the proposed lightweight process suitable foranies? Taking into account the focus presented by Yin

    SmalyearstaffdeveTheof orinforters.custosoftwhavewasprevincrefor ethisimpoof gr

    Smalversicomdeveandreseaconsdesign type of the case study in this work is multiple

    stic, since the strategy has been applied in the contextll business. The object of study is the Lightweight pro-porate improvements. The measures used to investigateh questions are: (i) the effort involved in carrying outs associated with the SPI initiative, and (ii) the capa-of the processes under analysis (those which need tod). Furthermore, we also took into account the benetsy the small organizations. In this vein, the analysis unitsrovement iteration composed of the activities of formu-xecuting improvement, along with the processes to bey the organizations.pplication and validation of the study object, weworkedompanies that form part of the critical reference groupPETISOFT project, which were our research subjects.protect their anonymity, in this paper we shall refero companies as SmallComp1 and SmallComp2. Next,es about these companies are described, including theirusiness area and motivation to improve:

    ation andpreviousdevelopmstrategy ping the deare criticasiders thaif we kno

    It is impclosely consoftware prtime, eithera software

    The topted to procewell as forpanies. NeiPETISOFT project.

    p1 is a small software development company with 5xperience at a nationwide level. The companys presentade up of seven people, six of whom are devoted to theent, operation, and maintenance of software products.work areas of the company include the developmentational web portals, e-commerce solutions, geographicon systems, and support software for education cen-ough this company offers off-the-shelf products to itss, most of its workload is given by projects for customdevelopment. Its average sales during the last 3 yearsn $153,000. One of the main problems at SmallComp1the absence of visible and explicitly dened processes

    the organization from growing accordingly to theg demands of its customers. That was the main reasonishing a process improvement program. Furthermore,any considers that software process improvement ist because without it, the company has few possibilities.p2 is a small academic organization (in a Spanish uni-ith 13 years of experience at a nationwide level. Itresearch and development activities with softwareent projects for other organizations (through contractsagreements), which serve as a way of transferring

    results to the industry. The groups staff currentlyf 21people, 15ofwhomwork in thedevelopment, oper-

    maintenance of software products. As in the case of

    company, there were no dened processes for softwareent or project management. The organization took theroposed by COMPETISOFT as the reference for improv-velopment and project management processes, whichl due to the nature of its work. This organization con-t software process improvement is important becausew what we are doing, we can improve.

    ortant to highlight that these two small companies arenected and have a strategic agreement to tackle certainojects together. Depending on their workload at a givenof the two companies can request the development of

    product from its partner through outsourcing.management groups of both companies have commit-ss improvement as the support for the organization asthe systematic consolidation and growth of the com-ther of the two companies has previous experience in

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1671

    software process improvement. However, at the present time it isimportant for them to increase the capability level of their pro-cesses, in the quest to gain clarity, organization and tracking of theprocesses involved in the development of their software products.There is alstheir customcommencemlast year, wfrom the grrst improvshould incocess referencdevelopmeand the Spe

    5.2. Field pr

    The Pfemsented in t(see Fig. 2),From theseing out thecycles. An ithe cycle, dcuting imprcycle, one othe formulaencountereimproved).the initiatinSOFT to guidthat the proies and datroles and wCOMPETISOproducts, wposal, Geneand the pretation planiteration reSOFT and Peachworkpmore, the dself-containinterviewedfrom the im

    A descricollection is

    5.2.1. InitiaThe initi

    The rst onPETISOFT adcompanies.PETISOFT adfor process iparties in th

    After thawere perfonized thatGiven thatestablish onadministratmentcycle.themselvesstrategy. A

    capability of the software development and specic project admin-istration processes by one level. People were assigned to theroles established by the improvement process. The managementimprovement group of the organizations and the COMPETISOFT

    r sel, eacoratelly,

    zatioementh allwithemeation

    Diagnpro

    or thor thre devd tot al.,is c

    r plaby aprveyrposen ath anmenactiadmools.cicsed tto tsoftwweres perroduPET

    ses hprocetherels oftheor oimp

    epored inoth crs

    . Ford, inwcyc

    theyraft ons,r relmporere

    red.impresp

    emenemeno the desire to improve their processes, thus offeringers higher quality products and services. This led to theent of a process improvement cycle in both companies

    ith the support of an advisor in process improvementoup of researchers in the COMPETISOFT project. For thisement cyclewe suggested to these companies that theyrporate some process related to the Prole 1 of the Pro-e model of COMPETISOFT,which includes: the SoftwarentprocessSD, theSoftwaremaintenanceprocessSM,cic project administration process SPA.

    ocedure and data collection

    COMPETISOFT process (see Fig. 3) that we have pre-his paper, along with the PmCOMPETISOFT processwere the guidelines for both improvement initiatives.diagrams we can see that the procedure for carry-eld work is formed by one or more improvement

    mprovement cycle consists of ve activities: initiatingiagnosing the process, formulating improvements, exe-ovements and revising the cycle. In each improvementr more improvement iterations must be carried out fortion and execution of the improvement opportunitiesd for the processes under intervention (processes to beIn this respect, we used PmCOMPETISOFT to conductg, diagnosingand revising activities andPfemCOMPETI-e the formulating and executing activities. This impliescedures governing the eld procedure of the case stud-a collection plan are directly related to the activities,ork products described in PmCOMPETISOFT and Pfem-FT. The data collected are those included in the workhich are, as shown in Figs. 2 and 3: Improvement pro-ral improvement plan (made up the assessment reportliminary improvement plan), Improvement implemen-and Improvement report (made up the Improvementports). It is important to highlight that PfemCOMPETI-mCOMPETISOFT provide a self-contained template forroduct, tomake its constructionanduseeasier. Further-ata collected were stored by means of the use of theseed templates. At the end of the improvement cycle wethe companies on their opinion of the results obtainedprovement initiative.ption of the execution of the eld procedure and datapresented in the following sub-sections.

    ting the cycleation of the improvement cycle involved several steps.e was the start-up of the improvement cycle. The COM-visor compiled all the necessary information about theThe top management of both companies and the COM-visor signed a document called collaboration agreement

    mprovement, which guaranteed the agreement of all thee execution of the improvement cycle.t, the tasks for the creation of the improvement proposalrmed. In their initial meeting, both companies recog-their processes for developing software were chaotic.fact, the management improvement group decided toly the software development SD and specic project

    ion SPA processes within the scope of the improve-Thegoalwas to improve thepracticesof theseprocesseswithin the company by following the COMPETISOFT

    nother quantitative goal consisted of increasing the

    advisoFinallyincorp

    Finaorganiimproving wisharedimprovexpect

    5.2.2.The

    tasks fscope fsoftwaoppose(Pino e(whichadvisocessesand suthis pu

    Whing wideveloppanysprojectment tthe speexpresrelated

    ThecessesProcesWork pvalCOMprocesis, thegoals,the goaventedresults

    Thewere rcollectnies. Bprocessprocessdeneup theorderThis diteratioanothemost iment wregistein thepersonimprovimprovected a person responsible for process improvement.h company selected a pilot project through which tothe process improvements.the improvement proposal was socialized into then. The advisor and the person responsible for processt presented the project through a cycle start-up meet-the people involved in the pilot projects. The advisorthem the work which was to take place during the

    nt cycle and obtained feedback about the employeess of the improvement project.

    osing the processcess diagnosis involved different tasks. First of all, thee creation of the assessment report were performed. Theisrst improvement cycle includes anassessmentof theelopment and specic project administration processes asthe second capability level of the assessment method2010) of METvalCOMPETISOFT assessment methodology

    onformance with ISO/IEC 15504-2). The COMPETISOFTyed the role of evaluator. The advisor assessed the pro-plying the technique of evidence gathering: interviews

    s, using the information-gathering tools developed fore.ll the above had been done, the advisor had a meet-other member, the person responsible for the softwaret process of each company, in order to obtain each com-vity diagram of the software development and specicinistration processes (see Fig. 5) and to apply the assess-There was no meeting with the person responsible forproject administration process, because both companieshat they did not carry out any visible and formal activityhis process.are development and specic project administration pro-assessed in the following process attributes: PA 1.1

    formance, PA 2.1 Performance management, and PA 2.2ct management from the assessment method of MET-ISOFT (see Table 2). The assessment concluded that thead a zero level of capability (incomplete process). Thatsses were not well implemented or did not attain theirwas little evidence of any systematic achievement ofthe process, and there were failures that limited or pre-fulllment of the goals of the process. In addition, fewutputs were identied in the process.rovement opportunities of the processes evaluatedted and prioritized by using as a base the informationthis activity, together with the needs of the compa-

    ompanies decided to improve the software developmentt and then improve the specic project administrationboth companies the preliminary improvement plan washich the number of improvement iterations thatmake

    le of improvement was presented, together with theneed to be carried out in and the overall schedule.f the plan for the improvement cycle included three

    two related to the software development process, andated to the specic project administration process. Thetant risks for the performing of the cycle of improve-determined and their corresponding management wasMoreover, the training on SPI of the employees involvedrovement cycle was established. The advisor and theonsible for process improvement created the generalt plan with the assessment report and the preliminaryt plan.

  • 1672 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    llCom

    5.2.3. FormFor the

    developmenagement imup of theof each comimprovemenliminary improcesses, taccount, thspecic proimplant iniiteration thopportunitidesign theopment andregistered i

    During timprovingthe improvimprovemenimprovemeadministratof an effectthe employnies; (ii) coan assessmactivities, tactivities esof the impperformedlearning, w

    proorkemenlantig inolvempart

    ecicent tses. Tlannces, aprevhe caplets initstratprovinte

    spring ouasksny. Ties a

    Table 2Capability of t

    Organization

    SmallComp1

    SmallComp2

    Values of procFig. 5. Activity diagram for Sma

    ulation improvementsformulation of the improvements of the software

    t and specic project administration processes, the man-provement group and process improvement group (madeadvisor and the responsible for process improvement)pany carried out various activities. The managementt group reviewed and validated the draft of the pre-provement plan. However, after taking the state of thehe companys needs, and the available resources intoe management improvement group decided that for theject administration process it would be sufcient totial basic practices in the process. Therefore, for eache process improvement group used the improvementes that had been found to make a mid-term plan andimprovements incorporation for the software devel-specic project administration processes, which were

    n the improvements implementation plan.he rst improvement iteration, which was related tothe software development process, the strategies forement of this process were dened by the processt group. These strategies were also used in the nextnt iterations related to improving the specic projection process. These strategies are: (i) the establishingive communication channel between the advisor and

    cess imjoint wimprovof impworkin(v) invcess orand spdocumprocesterm presourposed

    As t(incomprocesadminirst im(underinitialcarryinment tcompaactivitees responsible for process improvement in the compa-ntinuous and rapid evaluation of the process followingent process; (iii) basic measurement of the processhrough the measurement of the effort involved in thetablished in the improvement project, both at the levelrovement process and at the level of the activitiesby the participants in the pilot projects; (iv) ongoinghich trains the people involved in the project in pro-

    COMPETISOlished by thdenition.improveme

    5.2.4. ExecuThe pro

    practice wi

    he organizations processes.

    Process Process attributes

    PA 1.1 P

    Software development 0.18 (P) 0Specic project administration 0 (N) 0Software development 0.21 (P) 0Specic project administration 0 (N) 0

    ess attributes: F, fully achieved; L, largely achieved; P, partially achieved; N, not achievedp2.

    vement. In addition to this training, there needs to bebetween the advisor and the responsible for processt so that the latter can be better equipped for the taskng an improvement process which will then carry onthe company on a continuous and permanent basis; and

    ent of the companys employees (responsible for pro-icipants) in the denition of the software developmentproject administration processes and the techniques andemplates to be used in the activities specied for theseheprocess improvement group thenestablished themid-ing of each improvement iteration, dening activities,nd responsibilities which support the strategies pro-

    iously.pability level of the software development process is zeroe process), it was necessary to create a denition of thisially (the same thing happened with the specic projection process). To support the achievement of this goal, aement sprint, focusing on the denition of the processrvention by each iteration) took place. By means of thist, the process improvement group (team involved in thet of this sprint) indentied and assigned the improve-so as to create a draft design of the processes for eachhis version of these processes was dened using thend work products which satisfy the rst level of the

    FT process reference model. The process pattern estab-is process reference model was the basis for the processEach company had to use this new process in its pilotnt project.

    ting improvementscesses dened in the previous activity were put intothin the company by means of the execution of new

    Capability level

    A 2.1 PA 2.2

    .04 (N) 0 (N) 0 (incomplete)(N) 0 (N) 0 (incomplete).10 (N) 0.10 (N) 0 (incomplete)(N) 0 (N) 0 (incomplete)

    .

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1673

    Table 3Intermediate development process assessment.

    Organization Process Increase of capability of PAs Capability level

    PA 1.1 PA 2.1 PA 2.2

    SmallComp1 (from N to P) 0.48 (from N to P) 1 (performed)SmallComp2 (from N to P) 0.38 (from N to P) 1 (performed)

    N, not achieve ely achieved (>50% to 85% achievement); F, fully achieved (>85% to 100%achievement).

    improvemeresponsibleindividualresponsiblethe improvpractice wination of thbefore, theselected bytion. Thus, tthe complexa small partbetween thmentwas thtraining inqwas in consweekly virtimproveme

    An imprprocess stardirected tow(responsibleble for proceview of thdetailed vieware develoresponsiblereviewed thpleted up uetc. From thad at leastimprovemeof improve(or latest vethe processtions of thobtaining thinvolved incess in themanage proand clariewere resolvtemplates focesses (for ethe projectuser manuaticipants upthat remainimprovemenopportuniti

    The joinand the mefact that thewere key posprints andsprint, a me

    er their opinions of, and suggestions about, the good and bads of the improvement sprint and iteration current and howiewed the implementation and use of the new processes inot project. Moreover, the institutionalization strategy for therocesses was that these should be used in all of the organi-s new projects from that moment onwards.

    Cycle revisionlve weeks after the commencement of the improvement

    the advisor carried out a new assessment of the develop-rocess (only in the pilot projects) in both organizations. Assee in the results shown in Table 3, this process increased

    ability to level one. Furthermore, the improvement reportsted for each organization show that base practices of theprojns. Ts attrecit, totes istask

    gisteevottabl

    r to S. Thisthe ee resh orgthe ove, bponswhie canple

    r the i

    ingllComllCom

    singllComllCom

    latingllComllCom

    tingSoftware development 0.69 (from P to F) 0.37Software development 0.66 (from P to F) 0.31

    d (015% achievement); P, partially achieved (>15% to 50% achievement); L, larg

    nt sprints which pursued this goal. The work teamfor carrying out these sprints was composed of the

    responsible for process improvement and the personsfor processes under intervention. This team identiedement tasks needed to put the dened process intothin the company. The team self-organized the assig-ese tasks between members. As we have pointed outimprovement tasks were only applied in pilot projects,the management improvement group of each organiza-hese taskswere carriedout in a controlledenvironment,ity of their applicationwas reduced and it affected onlyof the company.Moreover, due to the close relationshipe two organizations, the responsible for process improve-e same for both of them. Although this person had basicualitymanagementandprocess-basedorganization, hetant contact with the COMPETISOFT advisor (through aual meeting) in order to report on the evolution of thesent tasks and to solve any possible doubts.ovement sprint focusing on implementing the denedted in both organizations with an informative meetingards the employees participating in the pilot projects

    for process or participants). In thismeeting, the responsi-ss improvement presented the assistants with a generale COMPETISOFT process reference model and a morew of the process under intervention in the sprint (soft-pment or specic project administration process). Thefor process improvement, along with the assistants, alsoe goals of the improvement sprint, the activities com-ntil that point, project milestones, way of working,his time on, the responsible for process improvementone weekly meeting with each team, to work on the

    nt tasks. In the rst meeting associated with this typement sprint, the denition of the corresponding draftrsion) process was presented to the participants, andwas enriched and rened by following the sugges-

    e employees involved in the improvements, therebye approved nal version of the process. The employeesthe pilot improvement project used this new pro-

    ir daily activities. During the following meetings (toblems and conduct sprint) the participants discussedd how to carry out each process activity and doubtsed. Furthermore, they worked together to create ther thedocuments produced in each activity of thesepro-xample for the software development process, alongwithplan, requirements specication, system specication,l, operation manual, etc.). As well as all that, the par-dated the plan for the work of the improvement sprinted unresolved. In addition the responsible for processt updated the state and priority of the improvementes in the general improvement plan.

    discovaspectthey vthe pilnew pzation

    5.2.5.Twe

    cycle,ment pwe canits capgeneraspecicnizatioproces

    A spproductemplaout thewas reeffort d

    Theadvisocutionduringwith thfor bot

    Oninitiatithe resprocessnies, wthe com

    Table 4Effort fo

    InitiatSmaSma

    DiagnoSmaSma

    FormuSmaSma

    Execu

    t work of the person responsible for process improvementmbers of each team, their active participation, and they could make decisions with the advice of this personints in the execution and success of the improvementiterations. Upon the completion of each improvementeting took place with the participating employees, to

    SmallComSmallCom

    RevisingSmallComSmallCom

    Ps. = person; hect administration processwere implanted in both orga-he implantation of these practices allowed the value ofibutes of this process to be increased.c self-contained template was developed for each workmake its construction easier. An important item in theseto register theeffort. Bydoing this, theeffort of carryings associated with each activity related to say productsred in each of the work products. Table 4 shows theed to each activity of the improvement cycle.e shows that the time devoted by the COMPETISOFTmallComp1 is zero in the activity of improvement exe-is because the time the COMPETISOFT advisor devotedxecution phase corresponds with the weekly meetingsponsible for process improvement, which were the sameanizations.ther hand,we asked companies about the improvementy means of an unstructured interview carried out withible for process improvement and with one responsible forch was improved. From the point of view of the compa-highlight several important conclusions reached after

    tion of the rst improvement cycle:

    mprovement cycle.

    COMPETISOFT Company

    p1 1 Ps.4h 1 Ps.4hp2 1 Ps.9h 1 Ps.4h

    p1 1 Ps.6h 1 Ps.4.5hp2 1 Ps.19h 1 Ps.4h

    p1 1 Ps.1h 1 Ps.0.5hp2 1 Ps.3h 1 Ps.1hp1 1 Ps.0h 1 Ps.24hp2 1 Ps.5h 1 Ps.30h

    p1 1 Ps.4h 1 Ps.6hp2 1 Ps.5h 1 Ps.8h=hours.

  • 1674 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    The application of COMPETISOFT and specically the approachoffered by PmCOMPETISOFT and the Lightweight process to incor-porate improvements in driving theprocess improvement allowedboth organizations to obtain an important and rapid improve-ment in twand speci

    Obtainingtantaspecin a projeparticipatprocess imlar, despicaused.

    Applyingreduced timplemenpanies.

    At the enhad moveto a tangiprojects. Bnies havethey haveimprovem

    5.3. Analys

    In this srst applicatwo case stof this analevidence gi

    In both cimprovemeand to do bstrategy ofwaredevelothe commeopment prowhich the cpanies is zeprocesses odevelopmeupon proceble productthere is nothe problem

    To implatheir capabresistance tconsidered.project, butway inwhicis the formemphasizinwork withis to fosterin such a wimplementimproveme

    We havelish improvof the Lightstudies carand implanof the propo

    process that had been developed, the improvement iterations andsprints were used to strengthen the work and collaboration of theemployees taking part in these iterations, in the quest to generateearly andongoing achievement of improvements in bothprocesses.

    rmorde awors witave shtwee of pOMPlity len by) on tthe oatinghe aps hasmprohis een ay thne prrintr toan wary boreande. Thof t

    impred thne tof anthe eto premecou

    tivitis is tc impPfemestutionortandoug

    ) purof Scles d

    ed inlatiois oftionle, eveturn ofe direseenty oflatedrninationted.o of the key processes for them: software developmentc project administration.good results in relatively short periods was an impor-t for themotivationand involvementof theparticipantsct like this. Seeing such rapid results and their direction caused the employees to realize the possibilities ofprovement in general, and COMPETISOFT in particu-

    te the initial reticence that these projects might have

    the improvements in pilot projects signicantlyhe resources needed and the risk associated with thetation of improvements in key processes of the com-

    d of the rst cycle of improvement, both organizationsd from a chaotic and unpredictable software processble one, which is currently being used on developmentoth themanagement and the employees of the compa-seen the benets of this result and, most importantly,realized the need to maintain continuous and ongoingent, following the same approach as in this rst cycle.

    is and discussion

    ection, we highlight the most relevant aspects of thistion of the proposed PfemCOMPETISOFT process in theudies by means of a qualitative data analysis. By meansysis we try to derive results and conclusions from theven by the data collected.ompanies the management group is committed to thentproject, since theyhave realized that, inorder togrowetter, the company needs to be managed through theprocess improvement, from the execution of the soft-pmentprocess to the softwareprocessmanagement.Atncement of the improvement cycle, the software devel-cess was chaotic, as can be deduced from Table 2 inapability level of the processes assessed in both com-ro. This makes it clear that there is no visibility in thef these companies, which signies that the softwarent has a far greater dependence uponpeople rather thansses. For example, as we can see in Fig. 5, the only visi-of the SmallComp2 is the software product. In addition,possibility of traceability of this software product, withs that this implies.nt the improved processes, and in an effort to increaseility, strategies with which to tackle the skepticism ando changeon thepart of thedevelopment grouphad tobeThose attitudes are innate to this type of improvementit is plain that developersmust feel condent about theh they areworking.One strategy for addressing this riskation of group development in process improvement,g the advantages that can lead to conducting the dailyprocesses which incorporate better practices. Anotherthe groups implication in the improvement projectay as to support and contribute to the denition andation of the processes to be improved, thus causing thent to be bottom-up.used the Scrum agile method satisfactorily to estab-

    ement strategies through the denition and applicationweight process to incorporate improvements. In the caseried out, the Scrum agile method was used to denet processes in the two small organizations, by meanssed lightweight process. Through this new lightweight

    Furtheto incluin thetunitiethey hour ligpracticPfemCcapabiis givemodel

    Onformulfrom tprocesin the iities. Thave bcially bto derst spin ordeteam cboundit is mrenepracticplexityin theobserv(to decutiongoal ofcess inimprov

    Wethe acprocesspecition ofthecasmentato suppitoring

    AlthmentssomeprincipinvolvThe reanalysconnecexampment rcreatiois mor

    Asmajoriples reand leamunicsuppore, in both organizations PfemCOMPETISOFT allowed usll thosewhowere involved in an improvement iterationk carried out to incorporate the improvement oppor-hin the processes under intervention and with whichome relationship. To be specic, we have used Scrum inight process to support the denition and putting intorocesses within small organizations. After carrying out

    ETISOFT, the desired product is a process with a highervel and this is put into practice. This capability increaseincorporating best practices (from a process referencehe denition of the process and its implantation.ther hand, PmCOMPETISOFT describes the activities ofand executing improvement prescriptively. However,plication of PfemCOMPETISOFT we observed that thisallowed us to give more exibility to the team involvedvement iteration during the carrying out of these activ-exibility is supported by all elements of Scrum thatdapted and included in lightweight process, but espe-e improvement sprint, since some sprints should seekocesses and others to implement these processes. Themust have as its goal to dene an initial process versioncreate a work product upon which the improvementork in the following sprints. In PfemCOMPETISOFT theetween formulating and executing is not clearly traced;that an improvement sprint can identify activities toextend the denition of the process and also put it ine scope of an improvement sprint depends on the com-he improvement opportunities and cases to be tackledovement iteration. In this sense, from case studies weat the goal of the activity for formulating improvementhe improvement) was satised by means of the exe-early improvement sprint with this aim. Similarly, thexecuting improvement activity (to put the dened pro-actice within the company) was achieved through thent sprints that followed, which focused on this goal.ld add that, according to Komo-Sirvi (2004) one ofes which is critical to the success of an improvementhat of carrying out (implementing and monitoring) arovement plan. Basing our opinion upon the applica-COMPETISOFT, as well as from results obtained from

    dies,weconsider thatusingScrumtomanage the imple-of improvements in small businesses is a suitable waythis success factor. It was also appropriate for the mon-supervision of the SPI project.h the whole idea of Scrum (and not of just a few ele-sues self-management, we have tried to analyze howrums elements are related both to some of the earlyened to PmCOMPETISOFT and to some success factorsprocess improvement in small businesses (see Table 5).nship shown in this table is based on our subjectivethe description of Scrums elements and their strongerwith some of these principles and success factors. Foren though the whole of Scrum supports a rapid invest-

    n, we believe that the sprint concept, which allows theincrements of the improved process in the short-term,ctly related to this success factor.from the table, Scrum has a relationship with the vastprinciples and factors considered. Factors and princi-to: (i) early and ongoing achievement of improvement

    g, and (ii) effectiveparticipation, collaborationandcom-of employees in the improvement project are stronglyWe have not related any element of Scrum with the

  • F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677 1675

    Table 5Relationship between practices of Scrum and SPI success factors.

    Scrums elements

    Product Release planning Sprint Sprint planningg

    Sprintbacklog

    Scrummeeting

    Sprintreview

    Sprintretrospective

    PmCOMPETIEarly and o

    of improX X X X

    Continuoudiagnos

    Elementalmeasure

    Effective gand com

    X X X

    Continuou X X X

    Success factoAttainmen

    investmDenition

    objectivresource

    X

    ParticipatiSPI (but

    Suitable cobetween

    SPI shouldlearning

    principle ofrying out afocus on reImprovemenvalCOMPET

    In additiments is thand the advScrumthrouities of formdetermininits managemcess improvethem an acimprovemeshort amoubetween thpating empproblems wthus allowicases. On toexibility intion) of theexibility wments proviemployeesaspects: rswhat they c

    On theimprovemeprocess to ineffort for Smto the adviSmallCompsponds to thNote that thperformanc(39 and 47SOFT advisoThis is beca

    activrs

    efforisorryingouldto a

    rryinort rrovemCOemennvolf theowsnalis effportbacklog meeting meetin

    SOFTs principlesngoing achievementvements

    X X X X

    s and rapid processisprocessment

    X X

    roup collaborationmunication

    X X X

    s learning X X X

    rs of SPIt of a rapident return

    X

    of feasible SPIes with the availables

    X X X

    on of employees inton-up)

    X X X

    mmunicationSPI participants

    X X X

    be based upon, not upon control

    X X X

    process diagnosis, because this principle involves car-rapid software process assessment by following a

    viewing the process capability. The component of ourt framework that gives support to this principle is MET-ISOFT assessment methodology.on, as can be seen in Table 4, the execution of improve-e direct responsibility of the organizations employees,isors effort in this is minimal. The fact of having usedgh theproposed lightweight process to guide the activ-ulation and execution of the improvements has been ag factor in the success of the improvement cycle, since

    ent practices allowed the person responsible for pro-ment to involve theemployeeswhoparticipated, givingtive part in the process improvement, thus generatingnts (dene the process and its implementation) in antof time. Communication (usually informal) tookplacee responsible for process improvement and the partici-loyees in the improvement iterations, through which

    to theformedis, thisthe advthe car

    It cspondswas cathe effan impwith Pimproveffort ieach oure shin thethat ththe imere quickly resolved and decisions were made rapidly,ng to continue the implementation of improvementp of all this, it offered the employees who took partthe incorporation into the processes (under interven-

    improvement opportunities through the iterations. Thehich the Lightweight process to incorporate improve-ded for the responsible for process improvement and theparticipating in the improvement project is seen in twot of all in who built the improvements and secondly inhose to work on and in what order.other hand, Table 4 shows the effort involved in thent cycle followingPmCOMPETISOFTand the Lightweightcorporate improvementsbybothorganizations. The totalallComp1was54h (h), ofwhich28% (15h) correspondssor and 72% (39h) corresponds to the company. For2 the total effort was 88h, 46% (41h) of which corre-e advisor and 54% (47h) corresponds to the company.e effort involved on the part of the companies in thee of the activities of the improvement cycle is similarh, respectively). However, the effort of the COMPETI-r is greater (almost triple) in the case of SmallComp2.use the gathering and analysis of information relating

    and realizaparticipatiodiagnosing

    If we beweeks, theapproximatand the comassigned to

    FigX X X X

    X X X

    X X X

    ities of initiating, diagnosing and formulating were per-t for this organization, and then for SmallComp1. Thatt is related to the learning and experience acquired byin the tasks and products that had to be carried out forout of the improvement process activities.be said that the effort related to SmallComp2 corre-company which was inexperienced in improvement (itg out an improvement cycle for the rst time), and thatelated to SmallComp1 corresponds to the execution ofment cycle in a company which was already familiarMPETISOFT and the Lightweight process to incorporatets. Fig. 6 shows the distribution and tendency of the

    ved on the part of the advisor and of the companies inactivities carried out in the two case studies. This g-that the greatest effort in the improvement cycle wasactivities of the process (executing and reviewing), andort falls back upon the organizations. This demonstratesance of offering agile strategies for the management

    tion of this work. We can also note that the advisorsn in the initial activities was of a high level (initiating,and formulating).ar in mind that the rst improvement cycle lasted 12n the average effort for each of the companies wasely 6h per week (taking into account the advisorspanys time). We propose that one person should be

    the responsibility for process improvement for a quarter

    . 6. Distribution and tendency of effort in the case studies.

  • 1676 F.J. Pino et al. / The Journal of Systems and Software 83 (2010) 16621677

    of their time, which is a load that we consider can be borne suf-ciently by small businesses. If we bear this aspect in mind thenthe processes proposed and described in this paper are better t-ted to the characteristics of small organizations. This is because itis not necesimprovemeand Richardmodel is apthese studiegroup in thment. In oumanagemen

    The resuevaluated ters. These ithe result oexplicitly dmanagememore custoappreciatedsteps with athe base foenhancingSmallComp2to have a band projectthe organizwith small

    It is impthe COMPEsible for proable to acqThe rst imlished the bfor incorpomanagemeongoing ansense is imstarted in o

    The casesome limitsbased on twtion, and (iiof daily actbecause the

    6. Conclus

    In thisment suchsets out tonies. The rimprovemenments needimprovemePfemCOMPmary PmCOmethod. Ththe goal bein charge oorganizatioagile methothe proposOur proposcess improv

    resources and in a short time. We have also presented how thisprocess has been set up in two small organizations by means of animprovement project.

    The description of the case studies undertaken and the analysisoutemenion ois relroveonee perprovnies.ainedrovemrporamploorating aapp

    on anprovn theho wapprprovnies.is thcom

    s. Fonatuagerouge acPET

    ere ipro

    rmoron ws impen themecounonitore prnAmin fed to remenork

    wled

    wouentsnce oof SpGAS. Ackshipspme

    nces

    ssonhods:sary for them to employ a person who carries out thent initiatives on a full-time basis, as is the case in Caseyson (2004) and Kautz et al. (2000) in which the IDEALplied to small and medium organizations. Furthermore,s highlight the importance of theManagement steeringe IDEAL model as a success factor in process improve-r proposed processes this aspect corresponds to thet improvement group.lts of the improvement cycles in both companies werehrough unstructured interviews to the main stakehold-nterviews revealed that SmallComp1 was satised withf the improvement cycle, since having a visible andened process for software development and projectnt permitted the company to attend the demands ofmers, and to extend the