StockholmSchoolofEconomics MasterofScienceinEconomicsandBusiness ThesiswithinManagement,15ECTS September2,2013 Managing in an Organized Anarchy An Autoethnography about the Struggle to Deliver in a Young Technology Company RevisedEdition Author AndersSjöqvist(20487) AcademicSupervisor Lars-OlofLychnell Opponent LenaLindvall(20457) Abstract An organized anarchy is governed by vague ideals, and all employees help to define the strategy. Solutions are created without knowledge of the problems and many of the solutions get rejected at an early stage and end up in the “garbage can.” The garbage can is revisited when new problems surface,asthecontentsmightcontainasolutiontooneofthem. Suchastructureoftenworkswellfor organizationslikeuniversities. Butitshouldbenotedthatwhiletheorganizationisgoodatdiscovering newpossibilities,theproblem-solvingspeedislaggingbehind. This thesis describes and analyzes an organized anarchy from the inside, through the eyes of a middle manager. The organization in the study was under pressure to deliver a product, but was better at coming up with creative ideas. As the pressure increased, the lack of accountability made the teams search for people holding the sole responsibility for the setbacks. The dominance of short- term goals limited the focus on sustainability and a healthy work environment. As a root cause, the authoridentifiestheorganizedanarchy—furtherreinforcedbytheresultsofscapegoatingandnegative leadership—and not the individuals themselves. While individuals are certainly responsible for their ownactions,theirrangeofoptionsislimitedbywhattheorganizationpermitsandencourages. The thesis suggests that certain organizations are worse at delivering value than others, despite thecompetenceoftheiremployees, simplybecauseoftheirstructure. Thetopicisimportant, ashuge amounts of resources are wasted while the organization is suffering an internal crisis. The use of autoethnographyasamethodhelpsemployeesandmiddlemanagerstoidentifywiththeauthorand classify the organization they’re working in, and it helps business leaders to understand how their choicesaffecttheworkingenvironmentfortheiremployees. Also,thethesiscanhelpindiagnosingan organizationthathasproblemsdelivering. Keywords: analyticautoethnography, organizedanarchy, garbagecanmodel, leadership, corporatescapegoating Acknowledgments I’d like to thank my academic supervisor Lars-Olof Lychnell for his continued support and patience during the entire research process. He was the one who originally suggested using a self-narrative method, which is rather unconventional within management studies, not expecting me to go along with it. Yet, despite being new to this category of methods, he stoodbymeandhelpedmefindtheanswerstomyquestions. I’m also grateful to Åke Freij for taking time from his research to help me out, not only withunderstandingtheautoethnographicmethodbutalsobydiscussingwhichofmyexpe- rienceswouldbemostinterestingtoareader. Furthermore, I’d like to thank my parents—Arne and Maud Sjöqvist—for their support and their help with proofreading, and finally those of my former colleagues with whom I was discussing the events during my time at the company as well as afterwards, and who later volunteered to read the thesis as it was getting closer to completion. For the sake of anonymity of the organization I made the decision not to mention any names, but I want to tell them this: You helped me to broaden my perspective and to better understand what was going on. Without you, I couldn’t have found the answers that I did. When I shared thedraftwithyouIexpectedsuggestionsforalterationsbutreceivednone,whichI’dliketo think means that I covered all that was important. But your enthusiasm for what you read was even more valuable to me, as it taught me that this is not just a story about me. You toldmethatthethesishelpedyoutoreflectandcomprehendyourownsituations. Eachone of you has unique experiences from your time at the company, but all your experiences fit together. When I became too deeply involved with theories, there were moments when I feared thatmy ideas hadlost touchwith reality, but you helpedme understand thatdespite thesubjectivenatureoftheresearchmethod,theresultsarequiteobjectiveandlikelyrelevant inmanyotherorganizationsaswell. 1 INTRODUCTION Contents 1 Introduction Should management for a start-up really have to 1 Introduction 1 care about building a culture and detailing areas 1.1 Researchaimandresearchquestion . 2 ofresponsibility,requirementsforcommunication 1.2 Listofterms . . . . . . . . . . . . . . . 3 and which deliverables to use for measuring per- formance? It seems so boring and bureaucratic. 2 Methodology 4 Shouldn’t they be allowed to assume that every- 2.1 Autoethnographyingeneral . . . . . 5 one’s working towards the same goal and for the 2.2 Analyticautoethnography. . . . . . . 6 good of the company? The organization should 2.3 Datacollection . . . . . . . . . . . . . 7 be assumed to work efficiently as long as all the 2.4 Theprocessofwriting . . . . . . . . . 8 employeesaregoodenough,shouldn’tit? 2.5 Ethicalconsiderations . . . . . . . . . 8 This thesis will tell the story of an organiza- tion, with a huge potential because of a skilled 3 Theoreticalframework 9 3.1 Organizationalbehavior . . . . . . . . 10 (albeityoungandinexperienced)staffandpower- 3.2 Groupbehavior . . . . . . . . . . . . . 12 ful investors, that has had to struggle constantly 3.3 Individualbehavior . . . . . . . . . . 13 because of difficulties to deliver on the promises 3.4 Summaryoftheories . . . . . . . . . . 15 to the shareholders. It will then use theory to describe how such an organization can develop 4 Thestoryofamiddlemanager 16 ideas and reject them before they’re even tested, 4.1 Themainplayersinthisstory . . . . 16 andhowitcancauseanatmospherethatwillem- 1 4.2 HowIendedupworkingforthe bracesycophants whiledrivingtalentsaway. Al- company . . . . . . . . . . . . . . . . . 19 though the organization was trying to pinpoint 4.3 Anoutlineofmytimeatthecompany 20 its problems and solve them, it’s important to re- 4.4 BuildingtrustwiththeCEO . . . . . 22 member that solutions which point to any one 4.5 Meanstoanend? . . . . . . . . . . . . 25 particular employee or process are likely wrong. 4.6 Organizationalmythology . . . . . . 27 The organization should instead look for causes 4.7 Inefficientcommunication . . . . . . . 29 among their policies, to find out how they could 4.8 Lackofspecifications . . . . . . . . . 32 haveallowedittohappen. 4.9 Expectationsandprejudices. . . . . . 35 Of course, management is not simple. Some 4.10 Matchfixing . . . . . . . . . . . . . . . 38 companies flourish while others fade away and 4.11 Undervaluingthetalentedemployees 39 die. Sure, this can be due to a business idea that 4.12 Myownshortcomings . . . . . . . . . 41 simply wouldn’t cut it, or economic development that renders some services obsolete. But when 5 Analysis 42 a company has to struggle simply because of its 5.1 Organizedanarchy . . . . . . . . . . . 42 internal culture and structure, then the economy 5.1.1 Problematicpreferences . . . . . 43 and society are worse off as a result of it. When 5.1.2 Uncleartechnology . . . . . . . 44 a company is working for years to develop their 5.1.3 Fluidparticipation . . . . . . . . 45 productwithouteverreachingthefinishline,then 5.2 Garbagecanmodel . . . . . . . . . . . 45 huge amounts of money are lost as development 5.3 Scapegoating . . . . . . . . . . . . . . 46 costs and missed opportunity. If the company ul- 5.4 Narcissisticleadership . . . . . . . . . 48 timately fails it might not be because the idea it- selfwasflawed,asitdidn’tevenreachthemarket 6 Conclusions 48 whereitwouldeithersinkorswimonitsown. The topic about problems to deliver is an im- 7 Discussion 50 portantone. Theorganizationinthestudywasn’t 7.1 Reflection . . . . . . . . . . . . . . . . 50 7.2 Implications . . . . . . . . . . . . . . . 50 1flatterers,bootlickers,yes-men 1 1.1 Researchaimandresearchquestion 1 INTRODUCTION 2 a small bootstrapped venture run from a base- identity, and how this reflects on its ability to de- ment. They had a few dozen employees and of- liver. Theambitionistoprovideacasestudy,but ficesinbothBeijingandShanghai. Theyattracted with focus on my own role in the company and foreign investment, and the most prominent in- how I interpret events that I am an active partic- vestor was a venture capitalist that had invested ipant in. This will then be complemented by a in several renowned IT companies. Why couldn’t theoretical framework, to help entrepreneurs and they spot these problems? They clearly invested other decision makers build solid structures and in the idea that the company had presented, but deal with potential issues within their own orga- shouldn’ttheybeabletoguidethecompanyonto nizations. therighttrackalsointermsofprocesses? Having been a part of a rapidly changing or- Especiallyasastart-up,it’simportanttounder- ganization which was struggling to build a suc- stand that problems might arise when the com- cessful team but faced persistent problems, I can pany is growing and responsibility and account- describe these processes from my own perspec- ability scatter. In 2013, this company was still tiveratherthanasanexternalobserver. Although presenting itself as a start-up, even after having some of the attempts the company made to im- beenaroundsince2007. Thewilltostayastart-up prove the conditions turned out to be failures, might be a sign that they desired the enthusiasm something good can come of them by learning of the founders in a new company, but then got from the mistakes and helping others in similar thedisorganizedwayofworkingalongwithit. situations. My desire is to pass on these experi- The method of the thesis is unusual, as it ences,withmysubjectiveobservationsinfocus. presentsanarrativefromtheperspectiveofamid- From an external point of view, it would be dle manager within the organization. The narra- possible to examine how the company deals with tive is based on memories, diary entries and dis- communication, which method for software de- cussions with peers, but not on any formal inter- velopmentitusesandsoon. Manysuchattempts viewsordocuments. Thecontributionsofthethe- to find causes for problems had in fact been initi- sisarepartlythedescriptionofaproblem,theset ated by the company, but my ambition is to find oftheoriesusedtoanalyzeitandtheconclusions, a theory that transcends all the others and af- but also the story itself and this method for writ- fects how people think. To do that, I want to use inganacademictext. my own experiences to figure out people’s moti- After the introduction, the methodology of an- vations, since formal interviews might only give alytic autoethnography is presented in section 2, the answers that the interviewee either thinks are followed by a theoretical framework for describ- prudent, or are quotes from the official presenta- ingproblemsonanorganizational,groupandin- tion of how the company operates, regardless of dividual level in section 3. The largest part of whether they’re accurate or not. By conducting thethesiswillbethestoryitself,presentedinsec- research on my own, it might be possible to dig tion 4. The story is analyzed in section 5, which deeperthanthat. is followed by conclusions in section 6 and a dis- When analyzing conflicts, especially those that cussion of the results in a bigger perspective in one is a part of, it’s easy to find oneself in situ- section7. ations of finger-pointing and directing the blame for the conflicts at others. I believe that this is 1.1 Researchaimandresearchquestion inevitable, even in this kind of thesis, not only because parts of one’s personality is visible only Theaimofthisthesisistoinvestigateandanalyze to others and the reasoning of others will always a concrete example of how a small multicultural be more difficult to comprehend than one’s own companydealswithinternalconflictsanditsown thoughts, but also because this thesis aims to 2tomakeitonone’sownwithoutexternalhelp,forexample give a personal account of the events, giving no by reinvesting earlier revenue; from the expression “to pull promise whatsoever about objectivity. However, oneselfupbyone’sbootstraps” 2 1.2 Listofterms 1 INTRODUCTION itisimportanttorememberthatthethoughtsand be designed as a concept without being tested in feelings described are keys to understanding the the process. It assumes that requirements don’t decisionsandoutcomes. Still,ifthetextisallowed changeoverthecourseoftheproject,andthatthe to remain at that level, then the reader will have buyer will have enough money at the end of the littletolearnfromitbeyondtheauthor’spersonal projecttopayforthecosts. feelings towards his colleagues. Hence, although In reality, the Waterfall model often turns out an event is at first described from a profoundly to be impractical, and it is often contrasted by subjective view, the aim is to raise the topic to a agile software development, which is really a group higherlevelofabstractionandgeneralization. of methods. The most famous of these is Scrum, While not an immediate goal in itself but where development takes place during sprints, merelyaconsequenceoftheresearchmethodcho- which typically last for two to four weeks. After sen, the thesis serves as an example of how ana- each sprint, the product is compiled into a work- lytic autoethnography (see Anderson 2006a) can ing version that can be tested by the customer. be used in a business thesis. Given that autoeth- The customer could potentially abort the project nographyisuncommonasthechoiceforamaster at that point, if they are satisfied or run out of thesis, the reader is not assumed to be familiar money, or they could reprioritize in the list of re- withit. Consequently,emphasiswillbeputonan mainingtasks,fromwhichthedevelopmentteam explanationofthemethodinsections2.1and2.2. picks a number of tasks from the top, depend- The research question that I’m attempting to ing on how long time they predict that the tasks answeris: will take. The person who leads the Scrum team withinthecompanyiscalledScrumMaster,while How can theories of culture and teamwork the one representing the customer is called Prod- explainirregularitiesinproductivitywithina uctOwner. youngcompanywithalargelyinexperienced AnotheragilemethodisKanban,whichdoesn’t staff? havetheconceptofsprintsandtasksthattheteam is working with together. Instead, tasks are small 1.2 Listofterms enough to be handled by individual developers. This section contains explanations of some of the While Scrum limits work by the amount possible words and phrases used in the thesis. Although within a sprint, Kanban limits it by limiting the theaimhasbeentouseasfewcomplicatedorun- number of tasks in progress at once. Instead of familiar phrases as possible, there are situations focusing on delivering a finished product every where not using the proper terms actually makes few weeks, Kanban thus focuses on getting work thetextmoreconfusing. donebyspottinganddealingwithbottlenecks. Agile Software Development One common Android Android is Google’s operating system way of planning software development projects, forsmartphonesandtabletcomputers. called the Waterfall model, is to write a document describing the desired outcome, down to where Application Programming Interface (API) the buttons should be placed. It’s possible to Most computer software will at some point com- imaginehowsuchamethodcouldbesuitablefor municate with other software, for example when government agencies, which can’t sign a deal ex- getting information from a database, displaying cept after public procurement, where many are text on the screen or uploading pictures to a abletobidonthecontract. Nevertheless,itmakes server on the Internet. In order to achieve this many assumptions about how the process is con- communication, some rules must be set up that ducted, as even the slightest change is normally dictatehowandwhatkindofdatawillbesentin associated with huge costs and a renegotiation the request, and how the response is formatted. of the contract. It assumes that the software can TheserulesarecalledtheApplicationProgramming 3 2 METHODOLOGY InterfaceorAPI.Asimplifiedwayofexpressingit, RoyalInstituteofTechnology(KTH) TheRoyal is that every time two teams of developers need Institute of Technology is the largest institute for to make their code work together, they have to higher education in technology in Scandinavia. come up with an API. The API should preferably I’ve studied Computer Science there, as well as beeasytounderstandandtouse,butatthesame the CEO, the interim CTO and several others in time powerful enough to handle complicated re- thecompany. quests. Also, it should change as little as possi- blebetweenversion,toavoidbreakingthingsthat Stockholm School of Economics (SSE) The werealreadyworking. Stockholm School of Economics is the leading busi- ness school in Sweden, and one of the leading in Back-end Back-end at a software company is Europe. It’s the school where this thesis is pre- usually in charge of developing and maintain- sented, and apart from me also for example the ing the software that’s running on the company’s chairman of the board and the CBO (Chief Busi- servers. For example, many companies have nessOfficer)hadstudiedthere. large databases of users and user-related or user- generated information. These data must be han- 2 Methodology dledsothatthey’researchableandpossibletode- liverinaquick,reliableandsecureway. Theprob- As discussed in subsection 1.1, the thesis will lems in back-end are for example about how to try to find answers that are not easily identi- locateinformationfromterabytesormoreofdata fiable through interviews. Instead, I chose a spread out across many servers, and tell which method that would allow me to unobtrusively informationtheuserhasaccessto,basedoncom- collect data about people’s behavior, to find un- plexsetsofaccessrules. derlying causes that they themselves might have been unaware of. “Traditional” research methods Front-end Front-end has the responsibility of might otherwise have led me to point out prob- displaying data provided by back-end to the user lems with specific people or in specific processes, through an appealing user interface. In my com- sincethisiswhatmycolleaguesoftendiscussed. pany, different front-end teams were working on I’m discussing naming conventions on page 7 web, iOS and Android apps, although the An- in subsection 2.2, but the truth is that these self- droid project started during my time at the com- narrative methods often vary more between spe- pany and the other ones had been running for a cific authors claiming to use the same method while. Typicalfront-endproblemsarehowtocre- than between each other when comparing their ate graphics and animations that will look good official definitions. Hence, I chose analytic auto- on a variety of devices, and how to constantly ethnography because it describes what I’m try- keepupwiththelatestupdatesontheservers. ing to achieve, not because it was the only self- narrativemethodavailable. iOS iOSisApple’soperatingsystemforthemo- Thissectionstartsoffwithageneraldiscussion biledevicesiPhone,iPodTouchandiPad. about autoethnography, before moving on to the subtype analytic autoethnography. Then, I de- Operations The main responsibility of Opera- scribe how I collected my data and how I went tions is the maintenance of the company’s devel- aboutdescribingmyexperiences. Lastly,Iexplain opmentandproductionservers. what ethical considerations I’ve had during my research. Qualityassurance Qualityassuranceistheactiv- ity of ensuring that the requirements are met by theproductbeforedeployment. 4 2.1 Autoethnographyingeneral 2 METHODOLOGY 2.1 Autoethnographyingeneral either teach us how we would interpret a situa- tion which we haven’t experienced yet, or show “What is autoethnography?” you might ask. us how others who are different from us inter- My brief answer: research, writing, story, pretsomethingthatwehavealreadyexperienced. and method that connect the autobiographi- Moreover, it allows us to use thoughts and emo- cal and personal to the cultural, social, and political. Autoethnographic forms feature tions as tools for understanding concepts that concrete action, emotion, embodiment, self- might for some reason or other be difficult to in- consciousness,andintrospectionportrayedin vestigateorcomprehendobjectively. dialogue, scenes, characterization, and plot. Ellis (2004, pp. 135–137) expands the uses for Thus, autoethnography claims the conven- autoethnographytoalsoincludeself-therapy,and tionsofliterarywriting. (Ellis2004,p.xix) is paraphrasing a conversation she had with her studentsaboutthistopic. Thisis,however,where Ellingson and Ellis (2008, pp. 449–450) say that I personally think that the method is starting to autoethnography presumes that the reality is so- 3 get out of hand. I believe that the strongest need cially constructed , and are thus indirectly say- for therapy is during phases when it’s extremely ing that it’s not possible to be an autoethnogra- difficult to maintain a decent level of objectivity. pher while clinging to the ideas of the Enlight- Theauthorislikelytowritedownmorethanwhat enment of objectively observing, analyzing and is relevant to just convey the story, and the de- searching for the one and only answer. Without scriptions of other people’s actions might be af- getting stuck in the irony of such a statement, we fectedtothepointwhereitturnsunethical,aswill can draw the conclusion that autoethnography is bediscussedfurtherinsubsection2.5. Ellis(1993) aboutfindingonetruth. givesanexamplefromherownlife,bytellingthe In a way similar to autobiography, where the extremelytragicstoryofheryoungerbrotherwho lifeoftheauthorisdescribedthroughtheauthor’s diedontheAirFloridaFlight90crashinJanuary own voice, an autoethnography describes a cul- 1982,onhiswaytovisither. Ellisherselfexplains ture,whichtheauthorbelongsto,throughtheau- howdifficultthisisforher: thor’s own voice. The process of researching the selfiscalledintrospection,andbyallowingoneself Eachwritingandreadingofmytexthasper- tobecomebothresearcherandresearched,theau- mitted me to relive my brother’s death from an aesthetic distance ..., a place that allows thor’s experiences and emotions become part of metoexperiencetheexperiencebutwiththe the story and can be used to better understand awarenessthatIamnotactuallyagaininthis the reasons and purposes of the author’s actions. situation, and thus I muster the courage to (Ellingson and Ellis 2008, pp. 450–452) Further- continue grieving. This process may not be more,EllingsonandEllis(2008,pp.452–453)even attractive to everyone. Often it was painful refute the classical view of objective science, by forme,sopainfulthatevennowin1992,ten pointingoutthatscienceitselfisasocialconstruct. years later, I experience intense emotions of Subjectiveaccountsshouldn’tbeseenaslessratio- losswhenIreadthisstory. Indeed,myemo- nal than objective ones, since “reading emotions tional reactions over the years have been so ofselfandothersoftenformsanecessaryprecur- pronounced that at times I have noted them sorforrationalaction.” asIedited. (Ellis1993,p.727) Autoethnography thus allows us to see the Ellis’ article was published in the end of 1993, world through someone else’s eyes. This could almost 12 years after her brother’s death. At that 3Socialconstructionismandsocialconstructivismtracebackto point in time, it was probably possible for her to the work of Berger and Luckmann (1966), who claimed that thinkrationallyenoughaboutittoknowwhether the terms “reality” and “knowledge” are products of the so- she actually wanted the article to be published. ciety we live in, and that even the simplest common sense But her own description of her feelings makes it knowledgeistheresultofsocialinteractions. Thisthesiswill possibletodoubtit. Ibelievethatthemoresevere notattempttoexplainthistheoryinanyfurtherdetail. the experience was, the more thought has to be 5 2.2 Analyticautoethnography 2 METHODOLOGY put into whether to actually publish the text. It’s beliefs,thoughtsandactionstoemphasizeon likely to work for some of the people, some of the awareness of the effects of the researcher the time. But I’m skeptical towards therapeutic onthegroupandviceversa. autoethnographyasageneralpieceofadvice. 3. Narrative visibility of the researcher’s self: The researcher should not only reflect on her 2.2 Analyticautoethnography thoughts and actions, but also describe her- Autoethnographiescomeintwoflavors—evocative self as an active participant, to avoid seem- and analytic. This distinction was originally sug- inglikeasilentobserver. Thepotentialpitfall gestedbyAnderson(2006a): hereistowrite“authorsaturatedtexts.” Theauthorproposesthetermanalyticautoeth- 4. Dialoguewithinformantsbeyondtheself: While nographytorefertoresearchinwhichthere- evocative autoethnography focuses on intro- searcher is (1) a full member in the research spection, analytic autoethnography needs to group or setting, (2) visible as such a mem- ensure that the results are generic enough ber in published texts, and (3) committed by involving others within the group. If too to developing theoretical understandings of much effort is spent on the self, then the broadersocialphenomena. genre might have closer ties to autobiogra- Anderson had noticed that autoethnography phy. was underrepresented within the genre of ana- 5. Commitment to theoretical analysis: An evoca- lyticethnography,andcameupwithfivekeyfea- tive autoethnography will often leave the turesofananalyticautoethnography: readerafterhavingdeliveredthe“story,”but 1. Complete member researcher status (CMR): The an analytic autoethnography needs to keep researchermustbeafull-timememberofthe going, to pierce the surface and at least at- group, but different from the others in that tempt to analyze or draw conclusions in a she has dual responsibilities. Apart from de- broader perspective. Anderson explains it voting full attention to the activities within best: thegroup,shemustcontinuouslyspendtime Thepurposeofanalyticethnographyis and effort recording the events. This forced not simply to document personal ex- reflectionwillalsomakehermoreattentiveto perience, to provide an “insider’s per- theunderlyingprocesses,whichofcourseaf- spective,” or to evoke emotional res- fectsherunderstandingoftheevents. Ander- onance with the reader. Rather, the sonalsoexplainsthedifferencebetween“op- definingcharacteristicofanalyticsocial portunistic” and “convert” CMRs. In short, scienceistouseempiricaldatatogain an opportunistic CMR is part of the group insight into some broader set of social before making the decision to write about it, phenomenathanthoseprovidedbythe datathemselves. and a convert CMR initially has a research interestandimmersesherselfasamemberof Oneofthemostprominentsupportersofevoca- thegroupduringthecourseoftheresearch. tiveautoethnographyisCarolynEllis. Inthesame 2. Analytic reflexivity: In traditional ethnogra- 2006 issue of Journal of Contemporary Ethnography whereAndersonpublishedhisproposal,Ellisand phy,theresearcherlimitsthefocusonherself, Bouchner (2006, p. 432) questioned even the na- and instead directs it outwards. In the cases tureofwhatAndersonsuggested: wherethereisareflexivepartinthetext,itis clearly separated from the analysis of the ex- “I’m having doubts. The more I looked ternal culture. In analytic autoethnography, closelyatwhatLeon[Anderson]waspropos- ontheotherhand,it’svitalfortheresearcher ing, the more I thought to myself, ‘but this to incorporate a reasoning around her own isn’tautoethnography. Whydoeshewantto 6 2.3 Datacollection 2 METHODOLOGY call this autoethnography? It’s just another 2.3 Datacollection genreofrealistethnography.’” Whileworkingatthecompany,Istartedwritinga Anderson(2006b,p.455)respondedbycriticizing moretraditionalthesisandcollecteddatathrough thecircularargumentthatsinceautoethnography interviews, but I had to abort this project as the had been redefined to mean evocative autoeth- workload increased and since I couldn’t collect nography, no such thing as analytic autoethno- the weekly feedback that I needed. It should be graphy can exist. He doesn’t recognize the right noted, that any information collected during my of Ellis and Bouchner to redefine the word, and firstattempttowriteathesisforthecompanywas claimstohavedefineditwithameaningcloserto discarded as the work started with the new one. theoriginalintent. Although I had collected information that the Two years later, Ellingson and Ellis (2008, companyvolunteeredthroughinterviews,thatin- p.445)acknowledgetheexistenceofthetwosep- formationwasmeanttobeusedinanotherthesis arate orientations within autoethnography in the and I found it unethical to make use of it once I beginning of their chapter, but towards the end haddecidedtochangetopics. Ellis once again gives voice to her concerns that Having realized that I needed a new topic for Anderson’s attempts at widening the concept of mythesis, Istartedkeepingadiaryofthingsthat autoethnography will make her and others like happened around me. I took notes on an almost herlosecontroloverthegenretorealistethnogra- daily basis from the end of November 2012 until phers, while the name itself gets watered down the beginning of June 2013, when I left the com- until it no longer challenges mainstream ethno- pany. After that, I occasionally continued taking graphies. (EllingsonandEllis2008,p.460) notes, when I was in contact with the company. For my part, I find it sad that the debate must Allinall,Ihave203diaryentries,averagingabout focus so much on naming conventions. I don’t 100wordsperentry. mind the variety of names for similar genres, in- Apart from the diary, I also maintained a file cludingbutnotlimitedtopostmodernanthropology withrandomthoughtsthathadcrossedmymind (see e.g. Reed-Danahay 1997, p. 17), postmodern about the situation of myself and the company, ethnography, ethnic autobiography, autobiographical and I saved some of the more important e-mail ethnography(seee.g.Reed-Danahay1997,p.2),au- messagesandchatlogsforreference. tobiographical sociology, auto-anthropology and self- The process of collecting data in this way first narrativeresearch(seee.g.Anderson2006a,p.373). became a habit, and then almost an obsession. I In many cases, I believe that these overlap, but it feltlikethedayhadbeenmoreproductiveifIhad should also be possible to pick a name that not summarizedandreflectedontheeventsthattran- only suits the actual content of the text but also spired. Thesummariesalsoturnedouttobeabit encapsulates its purpose. I have nothing against therapeutic. If I was upset or annoyed, I could Ellisorherwork—infact,Ithinkmanyhavealot just write down my thoughts and then be done to learn from her way of writing—but I would at with it. The thoughts was already there, written the same time love it if I could gain something down and saved for future, so why would I have morethananemotionfromreadinghertexts. to spend more energy on them? Furthermore, I I like it that autoethnography allows me to tell found it rewarding to be forced to write down a the story about who I was, how I felt and what summary, where I needed to reflect not only on I did. I also like it that traditional research lets others’ behavior, but also on my own. Actually, me analyze, get an overall picture and convey most of the material for this thesis is the result of it to others. Fortunately, it seems like analytic thethoughtprocessitself,associatedwithwriting autoethnography can combine the best of both notes, and not the diary. By writing things down worlds. It would be a shame if naming conven- inanorganizedway,Icouldalsokeepthemorga- tionswoulddiscourageresearchersfromchoosing nizedinmymind. Althoughthediarybecamemy thismethod. mostimportanttooltodevelopanunderstanding 7 2.5 Ethicalconsiderations 2 METHODOLOGY of the culture and events, I never really felt that I my short thesis might accurately compare a case neededtoreaditintheend. study with academic theory, it would do very lit- The information I collected has never been tletofindthecausesfortheunderlyingproblems shown to anyone else—I consider it too personal in this particular organization. Instead, I started to share with others—but I have often discussed to view my experiences not as a sequence of sep- things with my colleagues, usually not with the arate events, but as one continuous story which intentionofcollectinginformationbutsimplybe- containsasmallsetofconsistent“truths.” cause many of us found ourselves in frustrating After I had found a few theories that seemed situations and needed to talk about it. My per- toexplainwhatwasgoingonatamoreprofound ception of things is of course to a large extent the level, both my academic supervisor and I gradu- outcomeoftheseconversations. ally began to understand that the strength in my thesis is my unique position to see every part of 2.4 Theprocessofwriting the daily activities in the company, something re- searchers normally can’t as they are not part of AsIfirststartedwritingdownmythoughtsabout theorganizationinthesameway. Tryingto“cover my experiences in the first draft of this thesis, I everything”didn’tmakesensetousfromatradi- wanted to write down everything that came to tionalacademicperspective, butwhenwritingan mind, to let my ideas and memories start to flow. autoethnographyitsuddenlymadeperfectsense. I knew that later on I’d have to filter these stories My story isn’t one about communications prob- with respect to the theories that I would eventu- lems or organizational learning. It’s about what allyselect. I saw, heard, and felt! It’s about my relationship Once I had started writing, I kept writing un- tomycolleagues,theproblemswefacedandhow til I had 17 pages of stories off the top of my we (for better or for worse) decided to deal with head, only occasionally referring to my notes to them! verify some details or order of events. With a During the process of writing, I spent a lot of few stories still on my mind, but with consider- time going through events in my mind and try- ablymuchmorefreedomtofocusononethought ing to see how different theories would match. I at a time, I started considering which theories also got help focusing my thoughts from some of to use. Among the interesting ones were inter- my former colleagues who supported me, and of nalandcorporatecommunicationtheory,employeeen- course from my academic supervisor. In a way gagement, human resources, social, cultural and in- similartowhenIwaskeepingmydiary,Ineeded dustrial/organizational psychology, escalation of com- towritetofindoutwhatsoundedreasonableand mitment,single/doublelooplearning,decisionmaking, whatIbelievedin. cross cultural management (especially Hall’s high Adifficultpartoftheworkwastodecidewhat andlowcontextcultures and Hofstede’s culturaldi- to tell and what to leave out. I had to choose to 4 mensionstheory)andafewothers. Butsittingand tellarepresentativepartofthewholestory,while attempting to compare these theories, I realized leaving out parts as they could sound incriminat- thatallofthemwouldworkaboutequallywellto ingorsimplydidn’tfitinanyofthesections. explaincertainaspectsofmyexperiences. Onthe other hand, that also meant that all of them were 2.5 Ethicalconsiderations equally bad at explaining the full story. Know- ing that the company had spent considerable re- The thesis primarily focuses on the company. sources on finding answers to how to increase Sincethetopicwaschosenbymewithoutexplicit efficiency, I came to the conclusion that while permission from the company, I will present the companyinananonymizedformattoavoiddraw- 4 It’soutofthescopeofthisthesistogointoanyofthese ing unnecessary attention to it. Furthermore, I’m theoriesindetail,orevengiveadviceonliteratureforfurther reading.AsearchontheInternetshouldbeenoughtogivethe bound by my non-disclosure agreement and may readeraquickintroductiontoanyoftheseconcepts,however. 8