understanding consumer preference and willingness to pay

FINAL REPORT

Understanding Consumer Preference and Willingness to Pay for Improved Cookstoves in Bangladesh

USAID WASHplus Project August 2013

WASHplus, a five-year (2010–2015) cooperative agreement implemented by FHI 360 with CARE and Winrock International as core partners, is funded through USAID’s Bureau for Global Health. WASHplus creates supportive environments for healthy households and communities by delivering interventions that lead to improvements in water, sanitation, hygiene (WASH) and indoor air pollution (IAP). WASHplus uses at-scale as well as integrated programming approaches globally to reduce diarrheal diseases and acute respiratory infections, the two top killers of children under 5 years of age. For information, visit www.washplus.org or email: [email protected].

Contact Information: USAID WASHplus Project FHI Development 360 1825 Connecticut Avenue, NW Washington, DC 20009-5721 Authors: Julia Rosenbaum, Elisa Derby, Karabi Dutta, with inputs from Kirstie Jagoe and David Pennise Communication via: [email protected] and [email protected] Acknowledgments: The WASHplus team would like to acknowledge and thank our colleagues at USAID, both in Washington and Bangladesh: Merri Weinger, Pam Baldinger, Sharon Hsu, Richard Greene, Sher Khan, Ramona El Hamzaoui, and Helen Petach. We are also indebted to those who supported this study both in the field and back in Washington, DC, including Orlando Hernandez, A.K.M.Anowar Hossain Mollah, Kate Kennedy Freeman, Syeda Hosneara Akter, Shajahan Mia, Laila Ishrat Jahan Ruen, Nadia Rahman, Sultana Aziz, A.F.M. Iqbal, Nazmul Haque, Jibun Nessa Begum, Sarah Yagoda, Rachel Lenzi, Dana Charron, Kirstie Jagoe, Todd Wofchuck, David Pennise, Misti MacDonald, and Kathrin Tegenfeld.

TableofContents

EXECUTIVESUMMARY..............................................................................................................................................................1

BACKGROUND................................................................................................................................................................................5

STRATEGICAPPROACH.............................................................................................................................................................6

STUDYOBJECTIVES.....................................................................................................................................................................7

METHODOLOGY............................................................................................................................................................................8

Procedure....................................................................................................................................................................................9

DescriptionofStudyGroup..............................................................................................................................................13

FINDINGS.......................................................................................................................................................................................17

CONCLUSIONSANDNEXTSTEPS......................................................................................................................................41

AnnexA:NGO,Village,andHouseholdSelectionCriteriaAnnexB:WillingnesstoPayWorksheetsAnnexC:StoveProfileSheetsListofFigures

1. CookingFuels2. GatherorBuyWoodFuel3. NumberofPeopleNormallyCookedforinHome4. Husband’sMainOccupation/SourceofIncome5. Women’sOccupation6. PreferredStove7. NumberPreferringICSoverTraditionalStove8. NumberPreferringICS9. NumberPreferringICStoTraditionalStove10. ImprovedCookstoveisGood11. “WordCloud”representingattributesnamedbyallconsumerstryingICS12. CookingProblems13. ChangestoMakeStoveBetter14. EstimatedMonetaryValueasaProportionofAnticipatedSalesPrice(Aggregate)15. EstimatedMonetaryValueasaProportionofAnticipatedSalesPrice(bystovetype)16.StoveUsageDuringandFollowingtheKPTStudy17.Monitoredvs.ReportedStoveUsage18.Boxplotshowingfuelconsumptionbystovegroup19.PercentageofWomenRespondentsReportingIndependentDecisionMakingonHouseholdPurchases

Charts

1.DescriptionofPeopleWhoWouldUseICS2.ChangesinCookingPattern3.PerceptionofPeopleWhoWouldUseTheseNewStoves4.WillingnesstoPayforICS–Method15.WillingnesstoPayPricingScenarios–Methods1&26.ProportionofAllRecordedCookingEventsPerformedbytheInterventionStove7.MeanDailyFuelConsumptionEstimates8.Mean24‐hourAirPollutantConcentrationsintheKitchen

UnderstandingConsumerPreferenceandWillingnesstoPay*USAIDWASHplusProject*August2013 1

EXECUTIVESUMMARYAstheevidencebaselinkingimprovedcookstoves(ICS)withpositivehealthandenergyimpactsgrows,sodoesattentiontohowbesttoinfluencehouseholduptakeandconsistentandcorrectuse.Appropriately,attentionfocusesonboth“hardware”and“software”issues—howtoimprovethefieldperformanceofthestovesthemselvesandmakethemmoreaffordable,accessible,andappealingtotheneediestconsumers.ThisstudyusesqualitativeandquantitativemethodsthatdrawfromsocialmarketingandsocialsciencetoexploreconsumerperceptionsoffiveofthemostpromisingICSpotentiallyavailablefordistributioninBangladesh.Thestudycomplementsothereffortsbyarangeofstakeholderstostrengthenmarket‐basedapproachesandconsumerchoiceforimprovinghouseholdairqualityandreducingtheenvironmentalimpactsassociatedwithdependenceonbiomassfuels.ThroughsupportfromUSAID/Bangladesh,theUSAIDAsiaRegionalBureau/Washington,andanadditionalgrantcontributionfromU.S.StateDepartment’sOfficeoftheSecretaryofState,GlobalPartnershipInitiative,WASHplusislayingafoundationfortheUSAID/BangladeshCatalyzingCleanEnergyinBangladesh(CCEB)programandotherkeyactorsbyconductingacomprehensiveassessmenttobetterunderstandconsumerneedsandpreferencesastheyrelatetoincreasingtheuptakeofICS,includinghouseholdtrialsofimprovedstovesnotcurrentlywidelyavailableinBangladesh.Toassessconsumerpreferences,researchersappliedaninnovativemethodologycalledTrialsofImprovedPractices,orTIPs.TheWASHplusapplicationoftheTIPsmethoduses“elicitationquestions,”whicharesemi‐structuredquestionsthathavebeendevelopedandvalidatedtosystematicallyidentifybarriersandmotivatorstochange,includingwhichfactorsaremostinfluentialinspurringtheperformanceornonperformanceofabehavior.ICSfuelefficiencywasmeasuredusingathree‐daykitchenperformancetest(KPT),widelyacknowledgedasthebestcurrentlyavailablemethodforaccuratelyestimatingdailyhouseholdfuelconsumption.TheKPTwascarriedoutusingacross‐sectionalstudydesignin116studyhouseholdsand24controlhouseholds.Twoapproacheswereusedtomeasuretheextenttowhichhouseholdsadoptedthenewstovesandthemannerinwhichtheyintegratedthemintotheircookingandkitchenmanagementpractices:self‐reporteduseofstovesattheendofeach24‐hourKPTmonitoringperiodandstoveusemonitoringsensors(SUMS).TheSUMSrecordedthestovetemperatureevery10minutesforatotalofapproximately10days;theresultingtemperatureprofileswerethenanalyzedtodeterminethefrequencyof“cookingevents”(i.e.,numberoftimesthestoveswerelit)perday.TheimpactoftheinterventionsonhouseholdairqualitywasexploredduringtheKPTmonitoring;illustrative(notstatisticallysignificant)resultswerecollectedfrommeasuresofminute‐by‐minutekitchenconcentrations(inalocationapproximatingthebreathingzoneofthecook)ofsmallparticles(PM2.5)andcarbonmonoxide(CO).Theimpactoftheinterventionsonwomenandchildren’sexposurewasexploredinthesamesubsetofhomesby


monitoringthe24‐hourexposuretoCOofboththecookandonechildundertheageof5inthehousehold.ProcedureFivedifferentimportedICSmodels1wereplacedinhomes,withthreeofeachstovetypepervillage,thustotaling15householdspervillageineightvillages,or120totalhouseholds.Eachhouseholdinthetrialwasonlyprovidedonetypeofstovetotest.Ineachhouseholddetailedcookstoveoperationandmaintenancetrainingwasprovided,andcookswereaskedtotryoutthestoveundernormalconditions.Eachhouseholdhadtheopportunitytotryanewimprovedstove(notcurrentlywidelyavailableinBangladesh)forthreeweeksandwasaskedtoofferitsfeedbackandopinions.Unlikeothersurveymethods,whereallfactorsareheldconstantandresearchersanalyzethefrequencyandrangeofresponse,thisqualitativemethodologyinviteshouseholdstoidentify,discuss,andresolvebarrierstousingthenewICS.HouseholdswerealsoaskedtocomparecookingontheICSwiththeirtraditionalorpreviousstoveonarangeofcriteria.Throughthesecomparisons,researcherswereabletoelicitcategoriesofattributesvaluedbythetargetconsumer.InterviewswereconductedatDay1,Day3,andDay21.SummaryFindingsThestudyclearlyshowedthatatleasttwostoveswereperceivedaspreferabletotraditionalcookstovesbymanyofthosewhotriedthem.Asiscommonamongmanyimprovedstoveinterventions,2however,noneascurrentlyproducedmetallconsumerneeds,andnonemetsufficientconsumerneedstocompletelyreplacetraditionalstoves.ConsumersmostappreciatedthePraktiandEco‐Chulastoves,withthepreferenceforeachstovevaryingbydistrict. OverallConsumerReactionstoNewStovesonKeyVariablesOverall,consumers“liked”thenewstoves,whichwasadistinctindicatorseparatefromwhetherornotthey“preferred”theICStothetraditionalstove.Thesegeneralreactionswerecommonacrossstovetypes.FemalecooksfeltthatthetasteoftheirfoodwasthesamewhencookedonanICSversusthetraditionalstove.Abouttwo‐thirdsofthestudyparticipantssaidfoodtastesthesame,withtheothersequallysplitbetweensayingitwasbetter(21)orworse(19).Respondentsoverwhelminglyfeltthestoveusedlessfuelthantheiroldstove,withthree‐fourthsofthegroupseeingfuelsavings.Someofthesecooksreportedthatthenewstovessavedupto60percentofthewoodtheywouldhaveusedinatraditionalwoodburningstovepercookingsession.Aboutafifth

1Ofnote,thestovestestedinthisstudywereallimportedfromelsewhereintheregionandwerenotdesignedfortheBangladeshmarket.Thesewerethesinglepot,built‐in‐place,rocketdesignstove(EnvirofitZ3000),asinglepot,portable,rocketdesignstove(EcoZoomDura),a2‐potportablemetalchimneystove(PraktiLeoChimney),asingle‐potportablefangasifierstove(Eco‐Chula),andasingle‐potportablenaturaldraftgasifierstove(Greenway).Onlyoneofthefive(Greenway)wasavailableforpurchaseinBangladeshatthetimeofthestudy. 2Ruiz‐MercadoI.etal.2013.QuantitativeMetricsofStoveAdoptionUsingStoveUseMonitors(SUMs).BiomassandBioenergy.URLhttp://dx.doi.org/10.1016/j.biombioe.2013.07.002PineK.etal.2012.AdoptionandUseofImprovedBiomassStovesinRuralMexico.EnergyforSustainedDevelopment.URLhttp://dx.doi.org/10.1016/j.esd.2011.04.001Ruiz‐Mercado,I.etal.2011.AdoptionandSustainedUseofImprovedCookstoves.EnergyPolicy,DOI:10.1016/j.enpol.2011.03.028SchepersJ.andM.Wetzels.2007.AMeta‐AnalysisoftheTechnologyAcceptanceModel:InvestigatingSubjectiveNormandModerationEffects.Information&Management,44,90‐103.


oftheparticipantsthoughtthestovesusedmorefuel,whichisinterpretedinthediscussionsection.Whenaskedaboutdifferencesinsmokeproduced,avastmajority(85)saidtheICSproducedlesssmokethantheirtraditionalstoves.HusbandspresentatthetimeofthesurveywhoansweredthequestionhadbasicallythesameimpressionsastheirwivesregardingthereductioninsmokefromthenewICS.WhenaskediftheICShadanyimpactoncookingpots,justoverhalftheusers(62)feltthenewstoveskepttheirpotscleaner,afewsawnoimpact(15),andathird(40)feltitmadethepotsdirtierthanthetraditionalstove.Again,thisfindingisdiscussedfurtherinthediscussionsection,butsomeusers“jammed”theICSwithwoodtomakeflamesvisiblymeetthecookingpot,whichwouldclearlyaffectimpressionsandcookstoveperformance.AmajorobstaclereportedisthatthecookingtimewasslowerusingtheICS.Three‐fourthsofrespondents(91)reportedslowercookingtime,afifth(24)reportedfaster,andjustafew(3)respondentssaidcookingtimewasthesame.Whenaskedtheopen‐endedquestion,“Whatdoyouthinkaboutthestove?”afterthreeweeks,manygavetheunpromptedresponsethattheyenjoycookingonthestove(49),andalmostafifth(21)saiditlooksnice.Womennotedthatchangeswererequiredtotheircookingstyle,includingtheneedtoprepareallingredientsbeforeinitiatingcookingandtositinfrontofthestovetendingthefire(asopposedtomulti‐tasking)whilecooking(seechartonpage29).Dislikesandsuggestionsforimprovementfellintotwogeneralcategories,thosethatcanbeaddressedthroughfairlysimplemodificationstothestovedesignandothersmoreappropriatelyaddressedthroughpoint‐of‐purchaseconsumereducationandfollowupfromserviceagentsorhealthoutreachworkers.Themostoverarchingcomplaintaboutallthecookstovesincludedinthetrialwastheirinabilitytocooklargevolumesoffoodinlargepots,especiallythePraktiandGreenwaycookstoves.Studyparticipantscompensatedforthisbyjammingthestovewithmorefuelandwrestlingwithlargepots,whichrenderedsomestoveslessstable.Asiscommonwithotherstovestudies,participantswereunaccustomedand/orunwillingtochopwoodintosmallpieces,thuscomplaintsweremadeaboutthesizeandangleofthewoodopening.Inaddition,traditionalstovesareconstructedsoastoallowa“naturalfeed”oflargewoodpiecesandotheragrofuelsanddungsticks;becausetheopeningintothecombustionchamberanglesdownward,thefuelnaturallyslidesfurtherintothecombustionchamberasitburns.Consumersmissedthisfeatureonthenewstoves;improvedstoveshaveahorizontalfuelentry,sofuelmustbemanuallypushedintothestoveasitburns.Lastly,consumersfoundexcessashcollectedinthestoveandsuggestedatrayforeasyemptying.Whilethislastitemcanbeconsidered,someoftheashbuildupwasduetoexcessiveamountsofwoodbeingburnedinthestoves(Figure12).IncaseofthePraktistovethemajorcomplaintwasthatthesecondpotwasnoteffectiveforcooking.FortheGreenwaystoveamajorcomplaintwasthatthestoveisnotstable.Besidesthesetwospecificconcerns,complaintsweresimilaracrossallstovetypes.Someoftheseproblemsandrelatedsuggestionsforimprovementcanbeappropriatelyaddressedbyimprovedconsumereducation,withoutwhichconsumerswillbelesssatisfiedbytheoverallperformanceoftheirstove,whichwillaffectuseandwordof


mouthrecommendationsforthenewstoves.Wesuggestrampingupeffortsinconsumereducationandbehaviorchangebecausesomestovefeaturesunderdiscussion(e.g.,sizeoffuelopeningandlackofvisibleflamesleapingfromthestove)arecriticaltoimprovedcombustionefficiencyandheattransfer;inotherwordstheyarekeyrequirementsoftheimprovedstove.PerceivedValueandWillingnesstoPayStudyparticipantsvaluedstovesforcertainfeatures,butdramaticallyundervaluedthemonetaryworthofthestove.Mostparticipantsestimatedthemonetaryvalueofthestovestobeonehalftoonequarteroftheiractualcalculatedvalue(whichalreadyincludesanassumedcarbonsubsidy)(Figures14and15).However,aswillbefurtherdiscussedinthefindingsanddiscussionsection,reportedvalueswerelikelyinfluencedbyashared(andperhapsdiscussed)perceptionthatparticipantsshouldbegiventhestovesasatokenofappreciationforparticipatinginthestudy.Ofthe120households,105studyparticipantsweregiventheoptiontopurchasethestovesatthemarketvalue.Onlyoneoptedtodoso,andasecondnonparticipantneighborpurchasedastove(seechartonpage33).Usingasecondmethodology,however,theremaining15householdswereofferedthestovesasgifts,andwerethengivenanoptionofacashbuyoutatmarketvalue.Surprisingly,onlythreeoptedforthe(relativelysignificantamountof)cash;theother12preferredtokeeptheirstove.SummaryKPTandSUMSFindingsICSfuelefficiencywasmeasuredusingaKPTin116studyhouseholdsand24controlhouseholds,andtemperature‐loggingsensors(SUMS)affixedtoallstovesinthehousecollecteddataonthefrequencyofcookingperiods.UsagepatternscapturedduringKPTmonitoringsuggesttheinterventionstoveswerecommonlyusedbythestudyhouseholds,butinallcases,didnotfullydisplacetheuseofthetraditionalstoves(seechartonpage35).Homesusingfouroutofthefiveimprovedstoveswerefoundtouseatleast16percentto30percentlessfuelthanthecontrolhomesoverthecourseoftheKPT,3arangethatmaybesomewhatartificiallylowduetounderreportedfuelmixingincontrolhomes(seechartonpage37).

3Itisimportanttonotethatthisdoesnotmeanthattheimprovedstovesused16–30percentlesswoodthanthetraditionalstoves.Rather,homesusingtheimprovedstovesalongsidetheirtraditionalstoves(whichiswhathappenedinmostoftheinterventionhouseholds)used16–30percentlesswoodthanhomesusingonlythetraditionalstoves.


BACKGROUNDConsumersinBangladeshtodatehavenotexperiencedanychoiceintheimprovedcookstovesmarketandhavenothadtheoptiontousehigh‐endimprovedmodels,includingimportedportablemodels.Thecurrentlydisseminated“improved”stovemodel,theBondhuChula,isabasicbuilt‐in‐placestovewithacementcombustionchamberandchimney,surroundedbyclay/mud(seephotobelowright).Thetraditionalstovesconsistofaholeinthegroundwitharaisedclayliponwhichtorestthepot,withaseparatefuelentryhole(seephotobelowleft).

USAID/Bangladesh’sEconomicGrowthOfficeprovidedfieldsupporttoWASHplustoconductanimprovedcookstove(ICS)consumerneeds,preferences,andwillingnesstopayassessmentinBangladesh(“Phase1”).TheUSAIDAsiaRegionalBureauprovidedcomplementaryfundingtoidentifykeybehaviorchangeelementsanddevelopamarketingplanandrelatedtools(“Phase2”)basedonthePhase1researchfindingsandotherregionallessons.UnderPhase1,WASHplusislayingafoundationfortheUSAID/BangladeshCatalyzingCleanEnergyinBangladesh(CCEB)programandotherkeystakeholdersbyconductingacomprehensiveassessmenttobetterunderstandconsumerneedsandpreferencesastheyrelatetoincreasingtheuptakeofICS,includinghouseholdtrialsofimprovedstovesnotcurrentlywidelyavailableinBangladesh.ThisBangladeshICSassessmentrepresentsUSAID’sfirstsignificantinvestmentinbehaviorchangeandimprovedcookstovesandwillformthebasisforitsfirstmajorinvestmentinimprovedcookstovesinBangladesh.Thecookstovesectorhasseenconsistentglobaltrendsofdrop‐offinimprovedstoveuseovertimeandparallelstoveuse(stovestacking)4inpartbecausetheimprovedstovedoesnotmeetalloftheusers’

4 Ibid.

Traditional sunken‐hole stove (two pot version)

Bondhu Chula built‐in‐place chimney stove; the

current model of “improved” stove most widely

disseminated in Bangladesh.


needs;bypayingmoreattentiontoconsumerneedsandpreferences,thebenefitsofimprovedstovescanbemaximized,andattritionandparallelusereduced.Otherreasonsforlowadoptionandsustaineduseofimprovedcookstovesincludedeficienciesindistribution,consumereducation,financing,andafter‐salesservice.STRATEGICAPPROACHWASHplus,afive‐yearcooperativeagreement(2010–2015)managedbytheBureauforGlobalHealth’sMaternalandChildHealthDivision,isimplementedbyFHI360(formerlytheAcademyforEducationalDevelopment),andincludesWinrockInternationalasacorepartner,withprimaryresponsibilityforWASHplus’sindoorairpollution(IAP)activities.WASHplus’soverarchingmissionistoincreasetheavailabilityanduseofwater,sanitation,andhygiene(WASH)andIAPinterventions.WASHplusfocusesonimprovingthepracticeofkeyWASHandIAP‐relatedbehaviors,includingtheconsistentandcorrectuseofimprovedcookstoves.Tothisend,WASHplusincorporatesmethodologiesandapproachesthatfocusonincreasingtheperformanceofimprovedpractices,notmerelyincreasinglatrinecoverageorsalesofimprovedcookstoves.Planningandpromotionareundertakenfromtheconsumerpointofview,incorporatingdesiredbenefitsandconsequencesratherthanfocusingonpromoting“what’sgoodforyou”orwhatmakessensefromapublichealthandorenergyefficiencypointofview.Equalemphasisisplacedonimprovinghealth‐relatedproducts(andservices)thatmeetconsumerneedsandwants—changingtheproductifneededtobettersatisfyconsumersratherthanconvincingconsumerstobuyproductsthattheymaynotvalueorthatmaynotmeettheirexpectationsandneeds.Lastly,WASHplusalsofocusesonincreasinghouseholddemand,inthiscaseforICS,bycraftingpromotionalappealsthatofferdesiredbenefitsthroughcrediblechannelsasdescribedabove;increasingaffordableandaccessiblesupplythroughproductmodification,enhanceddistributionchannels,andfeasiblepaymentoptions;andshapinganenvironmentwithsupportivepolicyandadequatecapacitytoplan,manage,anddeliverproductsandservices.WASHplusoperatesusingtheUSAIDFrameworkforImpact,whichpositsthattoseeimprovedpractices,inthiscaseimprovedcookingpracticesinBangladesh,aprogram(whetherpilotorat‐scale)mustensurethateffectiveandappealingproductsandservicesareavailableandaccessibletoconsumers;thatinstitutionsandpoliciessupporttherelatedproductsorbehaviors;andthattheseproductsarepromotedinawaythatreachesconsumersthroughconvincingappealsandmultiplecrediblechannels.ThismeansthatamarketingplanforICSinBangladeshmusttakeintoaccountstovedesign,paymentoptions,andfuelavailability;assessifgovernmentpoliciesinhibitimport,distribution,orsales;andhighlightwaysforpublicandprivatesectorinstitutionstobuildneededcapacitiesandworkincoordination.Theactualimplementationanduptakeofthemarketingplanwouldoccurthroughabroadsectorsupportprogramorprivatesectorinstitutions;themarketingplanwillpresenttheanalysis,rationale,andoptions.WASHplusformativeresearchwillanswergapsininformationrequiredtodevelopacomprehensivemarketingapproachforincreasingtheuptakeofICSinBangladesh.WASHplusresearchwillcontributenotonlytopromotionalstrategies,butalsotoICSdesign,distribution,andpaymentoptions.Withincreasedunderstandingofwhatboth


womenandmenwantfromastove—theattributes,characteristicsandbenefits—stovescanbemademoreaccessible,affordable,andappealingtolowincomeconsumers.WASHpluswilldrawonlessonslearnedinbehaviorchange,demandcreation,andmarketingofsanitationandwatertreatmentproductstodevelopaneffectivemarketingandbehaviorchangestrategythatwillsuggestalimitednumberofevidence‐basedapproachestoincreasetheuptakeofstoves;concepttestkeyelementsoftheseapproaches;anddeveloppractical“how‐to”toolstocontributetothegoalsandresultsofUSAIDenergyandhealthobjectivesinBangladesh.

STUDYOBJECTIVESFormativeresearchreferstoagroupofresearchmethodologiesspecificallydevelopedtoguideorinforminterventiondesigns.Guidingallformativeresearchisonesimplequestiondevelopedbythe“grandfather”ofsocialmarketing,AlanAndreason,aspartofhisBackwardResearchModel5:Whatinformationisneededtomakedecisions?Inthiscase,thequestionwasframedas:WhatinformationdoweneedtodevelopasolidmarketingplantoincreasetheuptakeofimprovedcookstovesinBangladesh?Withthelargerguidingquestioninmind,theteamdevelopedasetofquestionsthatthisformativeresearchsoughttoanswer.Theseare:ConsumerPreferenceResearchQuestions

1. Whatarethedesiredattributesofacookstove?Thisincludescharacteristicslikesize,portability,stability,color,andfunction(e.g.,timetocook,highandlowpowercapabilities).Thisincludedexploringcurrentstoveattributesandcookingexperienceandinitialexperiencewithnewstoves.

2. WhataretheperceivedbarriersanddislikestothesefivemodelsofICSbasedonathree‐weektrial?Whatmakesastovehardtouse?Whoapprovesordisapprovesofthestove?AnsweringthesequestionsrequiredtheuseofSUMSmonitorstomeasuretheactualnumberoftimesanddurationthatstoveswereusedinhouses,whichcouldthenbecomparedwithself‐reporteduse.

3. Aretherefeasiblesolutionstothesebarriers,eitherbychanginghouseholdbehaviorsorthedesignofthestove?

4. Whatisgoodaboutthenewstove?WhatdocooksandtheirfamiliesperceiveasgoodthingsabouteachparticularICS?

5. Whatcharacteristics,attributes,likes,anddislikesaremostpersuasivetohouseholds?Besidessavingsinfuelcosts,whatotherattributeswillinfluencethepurchaseofanICS?Aretherecross‐cutting“aspirational”attributesorothermoreabstractbenefitspeopleaspireto?Forexample,beingseenasmodern,agoodprovider,agourmetcook—attributesthatresoundandmotivateconsumersfromdeepwithin.

5AndreasenA.1985.BackwardMarketResearch.HarvardBusinessReview.


WillingnesstoPay6. Whatareconsumerswillingtopayforhigh‐endimprovedstoveswithfeatures

theydesire?7. Howdoesofferinginstallmentpaymentoptionsinfluencestovepurchase?

EffectivenessofVariousImprovedCookstoves

8. WhataretheactualfuelsavingsofthetrialstoveswhenusedundernormalhouseholdconditionsinBangladesh?ThefivestovemodelstestedhavealreadybeenshowntosignificantlyreducefueluseandIAPinlaboratorysettings,andinsomecasesfieldsettingselsewhere,andthroughthisactivitytheywillalsobefieldtestedforhouseholdeffectiveness.OnasmallerscaleIAPandsmokeexposurewillalsobemonitored.

METHODOLOGYToassessconsumerpreferences,researchersappliedaninnovativemethodologycalledTrialsofImprovedPractices,orTIPs.TheTIPsmethodologyisaqualitativemethodusedtodevelopandtestbehavioralandproductoptionswithtargetconsumers.IthasbeenappliedsuccessfullytointerventionsrelatedtoHIV,6nutrition,7waterfilters,8dengue,9sanitation,andarangeofothertechnicalareas.TheTIPsqualitativemethodologywasfirstdevelopedfornutritionprojectstorehabilitateundernourishedchildren.10Itdrawsfromassets‐basedmethodologiesthatlookforfeasibleandeffectivebehavioralimprovementsthatuseexistingorreadilyavailableresources.11TheWASHplusapplicationoftheTIPsmethoduses“elicitationquestions,”12whicharesemi‐structuredquestionsthathavebeendevelopedandvalidatedtosystematicallyidentifybarriersandmotivatorstochangeandwhichfactorsaremostinfluentialinspurringtheperformanceornonperformanceofabehavior.ThedatacollectedthroughthesemethodswillfillkeyinformationgapsessentialtodevelopingacomprehensivemarketingapproachforincreasingtheuptakeofICSin 6BeryR.andJ.Rosenbaum.2010.HowtoIntegrateWater,SanitationandHygieneImprovementintoHIV/AIDSProgrammes.WorldHealthOrganization(WHO)/USAID.7GriffithM.1992.ImprovingYoungChildFeedingPractices.USAID/TheWeaningProject.

8RosenbaumJ.2006.BringingtheConsumertotheTableResearchBrief:DevelopingaMarketingStrategyforImprovingHouseholdWaterQualityinNepal.USAID/HygieneImprovementProject.9RosenbaumJ.andE.Leontsini.2002.PlanningSocialMobilizationandCommunicationforDengueFeverPreventionandControl:AStep‐by‐StepGuide.SpecialProgrammeforResearchandTraininginTropicalDiseases,CommunicableDiseases.WHO.10DickinK.,M.Griffiths,andE.Piwoz.1997.TrialsofImprovedPractices(TIPs):GivingParticipantsaVoice,andDesigningbyDialogue:AProgramPlanners’GuidetoConsultativeResearchforImprovingYoungChildFeeding.USAID/SARAProject.11LappingK.,D.Marsh,andJ.Rosenbaum.2001.ComparisonofPositiveDevianceandOtherAsset‐BasedDevelopmentModels.SavetheChildren/AcademyforEducationalDevelopment.12MiddlestadtS.,K.Bhattacharyya,J.Rosenbaumetal.1996.TheUseofTheory‐BasedSemi‐StructuredElicitationQuestionnaires:FormativeResearchforCDC’sPreventionMarketingInitiative,PublicHealthReports.


Bangladesh.Understandingperceivedbarriersandsolutions;desiredorexecutedmodificationstostovesduringthetrialperiod(e.g.,removaloffuelgrateoradditionofamakeshiftstove‐sideshelf,expressedcolorchange);andperceivedanddesiredbenefitsandattributeswillhelpprogramactivitiesgoingforwardtoidentifyappropriatestovesintargetareasand/ormodifystovesforincreasedeffectiveness,appeal,anduse.Thiswillalsoprovideinformationvitaltodevelopingamarketingandbehaviorchangestrategy.ICSfuelefficiency(reportedintermsofreductionsinfuelusage)wasmeasuredusingathree‐dayKPT(version3.0,www.pciaonline.org/testing),widelyacknowledgedasthebestcurrentlyavailablemethodforaccuratelyestimatingdailyhouseholdfuelconsumption13.TheKPTwascarriedoutusingacross‐sectionalstudydesignin116studyhouseholds(threehouseholdsdeclinedtoparticipateandafourthhadincompletedata)and24controlhouseholds.Twoapproacheswereusedtomeasuretheextenttowhichhouseholdsadoptedthenewstovesandthemannerinwhichtheyintegratedthemintotheircookingandkitchenmanagementpractices:self‐reporteduseofstovesattheendofeach24‐hourKPTmonitoringperiodandtheuseofSUMS.TheSUMStemperature‐loggingsensorswereaffixedtoallstovesinthehouse(includingbothtraditionalandinterventionstoves)tocollectdataonhowoftenthestoveswere“turnedon”(i.e.,lit).TheSUMSrecordedthestovetemperatureevery10minutesforatotalofapproximately10days;theresultingtemperatureprofileswerethenanalyzedtodeterminethefrequencyof“cookingevents”(i.e.,numberoftimesthestoveswerelit)perday.TheimpactoftheinterventionsonhouseholdairqualitywasexploredduringtheKPTmonitoringinasubsetofsevenhomes(twohouseholdsfromthetraditionalstovegroupandonehouseholdfromeachofthefiveinterventionstovegroups)tocollectillustrative(notstatisticallysignificant)results.Minute‐by‐minutekitchenconcentrations(inalocationapproximatingthebreathingzoneofthecook)ofsmallparticles(PM2.5)andcarbonmonoxide(CO)weremeasured.Environmentalandcontextualinformationthatmightimpactindoorairquality,suchaskitchenvolume,wasalsocollectedduringthestudies.Theimpactoftheinterventionsonwomen’sandchildren’sexposurewasexploredinthesamesubsetofhomesthroughthemonitoringof24‐hourexposuretoCOofboththecookandonechildundertheageof5inthehousehold.

ProcedureTheWASHplusteamselectedpartnerNGOsineachofthestudylocations,whothenhelpedthefieldteamidentifyhouseholdstoparticipateinthestovetrial,distributethestovesandcollectthemattheendofthestudy,andensurestoveuserscontinuedtousetheICSduringthestudyevenifsomefunctionalproblemoccurredwiththestove.BasedonthePartnerNGOSelectionCriteria(seeAnnexA),DESHGORIBangladeshinBarisalandInstituteofDevelopmentAffairs(IDEA)inSylhetwereselectedtoconductsitevisitstoeachlocation.

13 Bailisetal.2007;Smithetal.2007;WHO2008.


WASHpluspurchased26modelsofeachstove(twoextraofeachmodelincaseofanyproblems)andhiredBangladeshNGOfieldpartnerVillageEducationResourceCenter(VERC)tocoordinatewithandtrainIDEAandDESHGORI,overseefieldlogistics,andsupportKPTwork.ThepartnerNGOs,alongwithWASHplusstaff,identifiedsixvillagesineachoftheirinterventionareas(foratotalof12villages),basedonasetCommunitySelectionCriteria(seeAnnexA).PartnerNGOsthenidentified20householdsineachvillage(foratotalof240households),usingHouseholdSelectionCriteria(seeAnnexA).Toavoidselectionbiasofanykind,WASHplusstafftogetherwithVERCconductedashortintensivefieldsurveytoensurevillagesandhousesmetallselectioncriteriaandwererepresentativeoftargetconsumers.FourprojectstaffmembersfromeachpartnerNGOweregivenatwo‐daytrainingonstoveinstallation,use,andmaintenanceinaworkshopconductedbyVERC.Praktisentarepresentativetoparticipateinthisworkshopsincemanufacturingandproperinstallationofmetalchimneysinhouseholdsisvitalfortheperformanceofthestove.Theotherstovemanufacturerssentdetailedtrainingmaterialsandstep‐by‐stepguidesforstoveinstallation,use,andmaintenance.WASHplusworkedwithalllocalpartnerstomakefinalhouseholdselections,distributestoves,andprovidetrainingontheirusetohouseholds.Stovesthenwererandomlyassignedto120ofthe240identifiedhouseholds.TheteamplacedoneofthefivedifferentICSmodels(seebelow)ineachofthehouseholdsforcookstouseandprovidefeedbackonthroughsemi‐structuredelicitationquestions.Basedonhighperforming14stovemodelsavailableinotherSouthAsiamarketsandbeyond,theWASHplusteamselectedthefollowingwood‐burningstovesforthisstudy,showninthephotobelow:

- Singlepot,built‐in‐place,rocketdesignstove(Envirofit)

- Singlepot,portable,rocketdesignstove(EcoZoom)

- Two‐potportablemetalchimneystove(Prakti)

- Single‐potportablefangasifierstove(Eco‐Chula)

- Single‐potportablenaturaldraftgasifierstove(Greenway)

Ofnote,allofthesestoveswereimportedfromelsewhereintheregionandwerenotdesignedfortheBangladeshmarket.EcoZoomstovesarenotcurrentlyavailableintheSouthAsianmarket,andonlyGreenwaystovesarecurrentlysoldinBangladesh.

14 The2012ISOInternationalWorkshopAgreementforcookstoveperformanceprovidesasystemforcategorizingstovesbasedonseveralperformancemetrics,includingtwometricsrelatedtoefficiency,fromtier0representingtraditionalstovestotier4representingaspirationalgastechnologies.TheIWAtiersonlyprovidecomparativeclassificationforstovesbasedonlabtests.Allofthestovesselectedforthisstudyhadachievedatier2orhigherratingfortheirefficiencymetricsinthelaboratory.

Trial stoves, clockwise from top left: Eco‐Chula, Prakti,

Envirofit, EcoZoom, and Greenway.


ThestudyhadplannedtoincludetheBioLiteHomeStove,butbecauseBioLiteproductionwashaltedinlate2012toaddressafanissue,itwasnotincluded.ThestudyalsolookedatanimportedricehuskstovefromIndiathatmaybeagoodfitforasegmentofBangladeshiconsumers,butdidnotincludeitinthestudyduetologisticalandgeographicchallengesofdoingso,giventhedistinctusergroupsforwoodversusricehuskfuel.PhotosandmoredetaileddescriptionsofthefiveimprovedstovesincludedinthestudyareattachedhereinAnnexC.Thesefivedifferentstovemodelswereplacedinhomes,withthreeofeachstovetypepervillage,thustotaling15householdspervillageineightvillages,or120totalhouseholds.Eachhouseholdinthetrialwasprovidedwithonlyonetypeofstovetotest.Ineachhouseholddetailedcookstoveoperationandmaintenancetrainingwasprovided,andcookswereaskedtotryoutthestovesundernormalconditions.Eachhouseholdhadtheopportunitytotryanewimprovedstove(notpreviouslyavailableinBangladesh)forthreeweeksandwereaskedtooffertheirideasandopinions.Unlikeothersurveymethods,whereallfactorsareheldconstantandresearchersanalyzethefrequencyandrangeofresponse,thisqualitativemethodologyinviteshouseholdstoidentify,discuss,andresolvebarrierstousingthenewICS.HouseholdswerealsoaskedtocomparecookingontheICSwiththeirtraditionalorpreviousstove.Throughthesecomparisons,researcherswereabletoelicitcategoriesofattributesvaluedbythetargetconsumer.Interviewswereconductedondays1,3,and21. WASHplusrecruitedandselectedaDhaka‐basedteamofinterviewersandprovidedthemwithethicalandtechnicaltrainingtoconductthebaseline,Day3,andDay21questionnaires.OnDay1ofthetrial,thetrainedWASHplusenumeratorsvisitedeachcommunityand:

Explainedthestudytoeachofthe15participatinghouseholdsineachvillage(threehouseholdsforeachoffivestovemodels),usingascriptinBengalipreparedbytheWASHplusteam(Stove‐TrialIntroduction);thisscriptincludedallinstitutionalreviewboard(IRB)‐requiredconsents

Conductedthe“BaselineQuestionnaire”withthesehouseholdsOncethebaselinequestionnairewascompleted,projectstaffofDESHGORIandIDEA(overseenbyVERC)distributedtheimprovedcookstovestothesehouseholdsandtrainedcooksandheadsofhouseholdonthecorrectusageandmaintenanceofthestove.Closeattentionwaspaidtoqualitycontrolintrainingandapplication(bytheNGOs)ofthestandardtrainingprocedureoncorrectusageforhouseholds,giventheimpactthattrainingqualityandquantitycanhaveonimprovedstoveusageandperceptions.

OnDay3ofthetrial,trainedWASHplusenumeratorsreturnedtoeachofthevillagesandconductedthe“Day‐3Questionnaire”witheachoftheparticipanthouseholds.OnDay21(afterthreeweeksofstoveuse),trainedWASHplusenumeratorsreturnedtoeachofthevillagesandconductedthe“Week‐3Questionnaire”witheachoftheparticipanthouseholdsoveraperiodofsevendays.Intheprocessofdatacleaningandanalysis,recordsfromtwohouseholdsneededtobeexcludedfromDay3andDay21comparisonsbecauseofpossibleerror/overlapinthedatacollectionprocess.ThetwohouseholdsremovedfromtheDay3andDay21surveyanalysiswerebothinthe


Kunarchorvillage,resultingin58villagesfromtheSylhetdistrictand60fromBarisal.Thefullcomplementof120householdrecordsisrepresentedintheanalysisandreportingofbaselinefindings.Samplesizesforeachanalysisarenotedincharts.KPTswereundertakenatdifferentpointsbetweentheDay3andDay21questionnaires.Allhouseholdfuelstobeused(wood,cropresidues,charcoal,kerosene,etc.)wereweighedatthebeginningandendofeachofthethree24‐hourmonitoringperiodsusingdigitalhand‐heldscales.Woodmoisturewasmeasureddailyineachhouseholdusingadualpin,electricalresistance‐stylemoisturemeteratthreepointsonthreerandomlyselectedsticksinthewoodpile.Ashortquestionnairewasalsoadministereddailytorecordinformationaboutcookingstoveandfuelusage,thenumberandtypeofmealsprepared,andthenumberofpeoplecookedfor.Thehouseholdswereaskedtomaintaintheirtypicalcookingpatternsforthedurationofthesurvey.AttheonsetoftheKPTs,SUMStemperature‐sensingdataloggerswereplacedonallinterventionstoves,aswellasonthepre‐existingtraditionalstoves(whateverthefamilyhadbeencookingonpriortothetrial),sothatusageofboththenewstoveandtheoldstoveweremeasured.TheSUMStrackedactualcookingperiodsforeachimprovedandtraditionalstoveoverthecourseofthethree‐weektrial.SUMSdatawasdownloadedattheendofthethree‐weektrialandanalyzed,comparingactualtoreporteduse,whichhelpeddetermineanyunreportedproblemswithornonusageoftheimprovedstovestested.Thisactivityappliedthefollowingtools/scripts:Stove‐TrialIntroduction:Thisscriptwasreadtoeachoftheparticipatinghouseholdsbeforethetrialbegan.Itincludedafullexplanationofthetrial,explainedthefollow‐upquestionnairesandwhentheywouldhappen,explainedanypotentialrisks,andaskedpermissionforparticipation,asrequiredbytheIRB.Thisscriptwaspreparedbythecorestudyteam,translatedintoBengali,andreadbytheenumerators.BaselineQuestionnaire:Thisquestionnairewasconductedwiththeparticipatinghouseholdsonthefirstdayofthetrialbeforethehouseholdwasentrustedwithatrialstove.Thisquestionnairedeterminedthebaselinestovemodel,stovecost,stoveusagepatterns,feedbackonexistingstoves,fuelusepatterns,fuelexpenditure,andotherrelevantcharacteristicsofhouseholdsparticipatinginthestovetrials.Day3Questionnaire:Thisquestionnairewasconductedaftereachparticipatinghouseholdhadthechancetousethestoveforthreedays.Thiswasusedtodetermineinitialpreferences,usepatterns,andotherinitialreactionsafteronlythreedays.Examinationofthestoveitselfandquestionsprobedforanymodificationsalreadymadetothestoveand/oritsdesigned/correctuse(userswereNOTencouragedtomakemodificationstothestovedesignoruse,noradvisedaheadoftimethatthiswasallowed).InterviewersnotedanyproblemsandthenrelayedthesetopartnerNGOstaff,whothenvisitedthehouseholdtosolvetheproblems,whetherthroughretrainingtheusers,orservicingthestove.Commonproblemsincludedimproperusage,functionalproblemsofthestove(e.g.,thebatteryofEco‐Chularunningoutduringcooking),andusingpolytheneandplastictostartthefire,whichthencreatedthicksmoke.Boththe


problemsandacceptable/feasiblesolutionswerenotedasdataofinterest.NGOstaffensuredthatthestovewasinproperworkingorderbeforeleaving.Day21Questionnaire:ThisquestionnairewassimilartotheDay3questionnairebutwasconductedafterhouseholdshadtheopportunitytousethestovefor21days.Semi‐structuredinterviewquestionswereagainusedtodocumentpreferences,use/experiencewiththestove,qualitiesattributedtonewandoldstoves,fueluse,cooking,andotheroutcomes.

DescriptionofStudyGroupThefieldsurveywasconductedinJanuaryandFebruary2013,intwowood‐fuelburningareasofBangladesh—SylhetinthenorthwestandBarisalinthesouth.Bothareasusewoodastheprimaryfuel;thiswasconfirmedin105of120households.Aboutone‐thirdofthestudyparticipantsexclusivelygatheredtheirwoodandabouthalf“mostlypurchasedorexclusivelypurchased”wood.TheremainderusedsomecombinationofpurchasedandcollectedwoodasnotedinFigure2.Duringthestudyitwasfoundthatalthoughwoodwasthepreferredfuelthroughouttheyear,dryleafisusedasasupplementaryfuelinthewintermonthslastingfromDecemberuntiltheendFebruary(Figure1).Manyhouseholdsburnthisfreefuelinspecialleaf‐burningmudstoves,whichtheyconstructoutsideintheopencourtyardtoavoidtheheavysmokethatisemittedbythisfuel.

Figure1


Figure2ThepartnerNGOs,DESHGORIfromBarisalandIDEAfromSylhet,selectedthestudyvillages.InBarisalthevillageswereBillobari,Bihangal,Ichakathi,andGonpara.InSylhetthevillagesselectedforthestudywereJangail,Kewa,Tilargaon,andKunarchor.InBarisalcookingusuallytakesplaceeitherinanopencourtyardinasemi‐permanentstructureorinaseparatekitchenawayfromthemainhouse.InSylhetthecookingtakesplaceinthemainlivingquarters;themajorityofthehouseholdscookontraditionalstovesplacedunderchimneyhoods(whichactaschimneys,pullingsmokeoutofthelivingquarters).Householdswereoriginallyselectedbecausetheyfitthebasiccriteriaofusingprimarilywoodforcooking,havingatleastfourpeopleinthehouseholdwithatleastonechildunder5,andbeingwillingtoparticipateinthestudy.Unfortunately,around20smallerhouseholdsmadeitpastthehouseholdselectionscreeningintothestudy,asthefamiliesincludedintheirreportednumbershouseholdmemberswhodonotlivefulltimeinthehouse.Mosthouseholdshadfourtofivefamilymembers,withtheaveragesize(5.3)fallingjustabovethenationalaverage(averagehouseholdsizeinBangladeshis4.4people).Insomecasestheyreportedthecorrectfamilysizebutfailedtoreportthatextrapeople(farmlaborers)atelunchandsnackswiththefamilysothatcookingwasperformedforalargernumberofpeople(Figure3).

n = 120


Figure3About98outofthe120participanthouseholdswereMuslim,16wereHindu,andsixwereChristian(allsixinBarisal).Themainoccupationofthehusbandwasbusiness(30.83percent),followedbyservice(22.5percent).Othercommonoccupationsweredriverofhiredvehicles,farmers,artisans,andabout5.83percentweredailylaborers(Figure4,representingfrequencies).Amongthewomen,46outof120wereengagedinincome‐generatingactivities.Poultryrearingandsewingwerethemostcommon(Figure5).Abouthalf(61/120)oftheparticipantsbelongtosomesortofwomen’sgroup(suchassavingscooperatives).

Figure4


NOTE ON PRESENTATION OF DATA AND

FINDINGS

This study included both qualitative and

quantitative methods. In depth, qualitative

questions were asked of smaller subsamples

trying particular stoves (where n is 118,

there were 24 each using Eco‐Chula,

Envirofit, and Greenway; 23 using EcoZoom

and Prakti) or small subsamples responding

to particular questions. Following standard

procedures for reporting qualitative data, we

are reporting on these data using words

(most, many, some, few) and numbers, and

only use percentages when reporting on the

entire study group of 120 for the baseline

data, 118 for the 21‐day group, and a few

other rare instances. For the most part,

results of each stove trial group are reported

as follows:

Most = 90% or above (at least 20 of 24) Many = 40% or more (at least 10) Some = 15‐39% (at least 4, less than 10) Few = less than 15% (2‐3)

Figure5Amongeligiblehouseholds,WASHplusdeliberatelyselectedhouseholdsthathadsomeregularincome(thosewhowerenotextremelypoor)andwouldbeabletobuythestovesattheendofthestudyiftheyreallylikedthem.Thisexcludedagriculturalsmallfarmersandfarmhands.Thestudyalsoexcludedrichfarmingfamilies,astheywerelikelytocookwithliquefiedpetroleumgas(LPG)orhavedomestichelpforcooking,ratherthanhavingthewife/motherofthehouseholdperformthattask.Alltheparticipantswerewithintheagegroupof16–65years;about60percentoftheparticipantswerecooksbelow35yearsofage.Some28percentofthewomenwere16–25yearsold,32percentwere26–35yearsold,23percentwere36–45,and13percentwere46–55.Only4percentwereabove55yearsold. ParticipantswererandomlyassignedoneoffiveICSandaskedtotryitoverathree‐weekperiod,providinginformationtointerviewersatbaseline,threedays,andthreeweeksasexplainedabove.


FINDINGSOverallConsumerReactionstoNewStovesonKeyVariablesBasedontheirresponsestotheDay21survey,consumersfeltthatthetasteoftheirfoodwasthesamewhencookedonanICSversusatraditionalstove.Abouttwo‐thirdsofthestudyparticipantssaidfoodtastesthesame,withtheothersequallysplitbetweensayingitwasbetter(21/118)orworse(19/118).Respondentsoverwhelminglyfelttheimprovedstovesusedlessfuelthantheiroldstoves,withalmostthree‐fourthsofthegroupseeingfuelsavings(85/118).Afew(8/118)respondentssaidtheICSusedthesameamountoffuelasthetraditionalstoves.Interestingly,aboutafifthoftheparticipantsthoughtthenewstovesusedmorefuel.ManyPraktiusers(16/118)andmostEco‐Chulausers(21/118)reportedthattheirstovesusedlessfuelthanthetraditionalstove.HoweversomeusersofEcoZoom(6/118),Envirofit(6/118),andGreenway(6/118)reportedneedingmorefueltocookonthesestovesthanontheirtraditionalstoves.Thisisinterpretedinthediscussionsectionbelow,butacoupleofpointsareimportanttonotehere.First,someofthestoveuserstookfreeleaffuelintoaccountintheirmentalcalculations,whichinfluencedtheseimpressions,andsecondly,agroupofparticipantsjammedextrawoodintothefuelentry/combustionchamberstocreatelargerflamesfromthestoves.Bothofthesepointsshouldbetakenintoconsiderationintheresultsrelatedtofueluseforthenewstoves.Thesegeneralreactionswerecommonacrossstovetypes.WhenaskedaboutdifferencesinsmokeproducedbytheICSversusthetraditionalstove,avastmajority(85/118)saidtheICSproducedlesssmokethantheirtraditionalstove.Afewsaidnochange(13/118),andasmallgroup(19/118)reportedmoresmoke.HusbandspresentatthetimeofthesurveywhoansweredthequestionhadbasicallythesameimpressionsastheirwivesregardingthereductioninsmokefromthenewICS.WhenaskediftheICShadanyimpactoncookingpots,justoverhalftheusers(62/118)feltthenewstoveskepttheirpotscleaner,afewsawnoimpact(15/118),andathird(40/118)feltitmadethepotsdirtierthanthetraditionalstove.Again,thisfindingwasinpartduetosomeusers“jamming”theICSwithwoodtomakeflamesvisiblymeetthecookingpot,whichwouldclearlyaffectimpressionsandisdiscussedfurtherinthediscussionsection.AmajorobstaclereportedisthatthecookingtimewasslowerusingtheICS,especiallyforlong‐cookingfooditemslikericeanddaal.Morethanthree‐fourthsofrespondents(91/118)reportedslowercookingtimeusingthenewstovescomparedtotheirtraditionalstoves,afifth(24/118)reportedfastercooking,andjustafew(3/118)respondentssaidcookingtimewasthesame.Inresponsetoanopen‐endedquestion,“Whatdoyouthinkaboutthestove?”afterthreeweeks,aclearmajoritysaidit:wascleaner,releasinglesssootandsmokeintothehouseandkitchen;usedlessfirewood;andemittedlesssmoke.Manyoftheparticipantssaid—unprompted—thattheyenjoycookingonthestove,andalmostafifthsaiditlooksnice.Foreachofthesepositiveattributesnotedhere,asmallminority(lessthan20percentineachinstance)saidtheirstoveemittedmoresmoke(14/118),thatitusedmorewood(12/118),andthattheydidnotenjoycookingonthestove(21/118).


ByDay3,themajorityofhouseholdspreferredtheirnewstovetotheiroldtraditionalcookstove.After21days,however,farfewerhouseholdspreferredthenewcookstovetotheirtraditionalstove,withratesfallingfrom56percentpreferringthenewstoveafter3daystoonly41percentpreferringthenewstoveafterusingitfor21days(Figure6).Thebreakdownbymodelofstove(Figures7&8)andbymodelanddistrict(Figure9)arebelow.Threestoves—theEco‐Chula,EnvirofitandPrakti—wereclearlymoreacceptabletoconsumersattheDay3survey.Morethanhalfofconsumerstryingthosestovespreferredthenewstovetotheirtraditionalstove.BytheDay21survey,however,userpreferencehaddroppedforallfivestovetypes,mostdramaticallyfortheEco‐Chula.Reportedreasonsbehindthisaredescribedinthediscussionsectionandprimarilyhavetodowithchangesrelatedtotheircookingpractices(Figures7,8,and9).

Figure6:HouseholdswereincludedinthiscomparisononlyiftheageoftherespondentatDay3wasthesameasDay21tobecertainpreferencesofthesamepersonwerebeingcompared.


Figure9:ThestovespreferredbytheusersinSylhetindecreasingorderare:Prakti,Eco‐Chula,EcoZoom,Greenway,andEnvirofit.InBarisalthedecreasingorderofpreferenceforstovesis:Eco‐Chula,equalpreferenceforEnvirofit,Greenway,EcoZoom,andthird,Prakti.Thedatapointsthatfollowexplainwhatconsumerslikedanddidnotlikeaboutthedifferentstovesbymodelofstove.Analysisoftheseresultsisincludedinthediscussionsection.Despitethedecreasedpreferenceforthenewstovesversustraditionalstoves,78percentofparticipantsoverallstillsaidtheirnewstovewasa“good”stoveafterthreeweeksofuse.PercentageperceptionbystovetypeisfoundbelowinFigure10.

Figure10:HouseholdswereincludedinthiscomparisononlyiftheageoftherespondentatDay3wasthesameasDay21tobecertainthatpreferencesofthesamepersonwerebeingcompared.


Explorationofthequalitativeandquantitativedatashedslightontheseseeminglycontradictoryfindings(seediscussionsectionbelow).Householdsappearedto“like”and“value”thestovesbutstillfeltthestovesdidn’tsatisfyalltheirkitchenneedsinthewaytheirtraditionalstovesdid.Ourconclusionsonwhatitwouldtake(includingconsiderationofstovedesignmodifications)togetparticipantsinterestedinpurchasingandusinganICSareincludedinthediscussionsection.ProfilesofSpecificStovesEnvirofit–Morethanhalfofuserssaidituseslessfuelandemitslesssmoke,andsomelikedthelooksandsaidthattheirhousewascleaner.Lastly,somementionedthatitwaswellmanufactured.Prakti–Almosteveryonecommentedthatthestoveemitslesssmoke,andamajorityalsomentionedthePraktileavestheirhousecleaner,useslessfuel,andlooksnice.Greenway–Amajoritycommentedthatlessfuelwasneeded,thatitlooksnice,andtheylikedtheportability.Some(butnotamajority)mentioneditemitslesssmoke.Concernscooksnotedincludedthatitappeareddelicateandunstable,andtheyworriedthestovewouldtipover.EcoZoom–Comparedtotheirtraditionalstove,amajoritymentionedituseslessfuelandaroundhalfsaiditemitslesssmoke,itlooksnice,andtheyappreciateitsportability.Cooksnotedthattheylikedthatitlookedbig(indiameterandheight)yetportable,andhasabroadbasethatmakesitstable.Theyalsonotedthattheappearanceandweightofthestoveconvincedthemitwasdurable.Eco‐Chula–Comparedtotheirtraditionalstove,manymentionedituseslesswood,emitslesssmoke,andlooksnice.Abouthalfalsomentionedthehousewascleanerthanwhenusingthetraditionalstoveandmanymentioneditcooksfoodquicklyandisportable.Concernscooksnotedincludedthatitappeareddelicateandthereforemightnotbedurable.Theyreportedlikingthefan,andthegasstove‐likeflamethataidedcooking.Theylikedthatitisportable,andthatthecookingvesselsareplacedonaseparatemetal“quadrapod”frame,sotherewasnofearthatthestovemighttipoverduetotheweightofthepot.PeoplemostlikedtheICSoverallbecausetheyemitlesssmokeanduselessfuel.Rankingalmostashighwasthatthestoveslookednice.Thestoveswereperceivedtobecleanerandproducelesssoot.Lessfrequentbutstillstrongresponsesincludedthestoves’portability,andtoalesserextent,theiroverallqualityandabilitytocookfoodquickly.Otherresponsesincludedbothaspirationalbenefitssuchas:“impressesothersandbringspridetomyhouse,”featureslike“theflameislikethatofanLPGstove,”andthatitretainsheatandproducesmoreflameandheat(seewordcloudbelow).


Figure11:ThewordcloudaboverepresentsattributesnamedbyallconsumerstryingtheICSinresponsetoaskingwhattheparticipantslikedaboutthenewstovescomparedtotheiroldprimarystoveafterthreeweeksofusage.Largertypesizereflectsthefrequencyofmentionoftheattribute.Answersweresimilarinresponsetoageneralquestionaboutwhysomeone(anyone)mightchoosethesestoves(asopposedtowhydoYOUlikethestove,whichcorrelateswiththewordcloudabove),withtheexceptionof“lookingnice”and“cooksfast,”whichwerementionedfarlessoften(seechartbelow).

Chart1:DescriptionofPeopleWhoWouldUseICSWhyWouldSomeone(Else)ChooseTheseStoves?

FrequencyN=118 %

Lesssmoke 80 68Savesfuel 80 68Portable 61 52Kitchen/potsstaycleaner 58 49Lookssmart/modern 4 3Looksnice 3 2.5Cooksfast 27 23

Noonewillliketouseit/nogoodreasontouse 4 3Other 6 5

Whethertheypreferredtheimprovedstoveovertheirtraditionalstoveornot,allusersencounteredsomeproblemsorbarrierstousingthenewstove.Therewaslittlevariationacrossstoves,withsomenotableexceptions,oftendirectlyattributabletothedesignofthatparticularstove.Someofthemajorproblemswerethatinallthestovemodelsittookalongertimetocooklargequantitiesoffoodin


largevessels.Usersfeltthatbigpotsdidnotworkwellwiththesesmallportablestovemodelsbecausetheflamedoesnotspreadtocoverenoughofthepotandthecookingpotmighttipover.InBangladeshthestaplefoodisrice,anditisconsumedinallthreemealsoftheday.Dependingonthefamilysize,thestoveusersfounditverydifficulttocooklargequantitiesofriceinthesestovemodels(Figure12).Thiswasanespeciallylargeobstacleduringthemonthinwhichthestovetrialstookplace,asfamiliestendedtocooklargerquantitiesofriceallinthemorningduringthecoolerDecember‐Februaryseason,ratherthanduringmultiplecookingperiodsspreadthroughoutthedayasismoretypicaltherestoftheyear.Userswhobelongedtosmallfamiliesofuptothreetofourmemberslikedthestovemodels.Stoveuserswereaskedaboutwhetheranythingabouttheimprovedstovewasn’tfunctioningproperlyduetothedesignofthestove.InthecaseofthePraktistove,themajorcomplaintwasthatthesecondpotwasnoteffectiveforcooking(13/23).Forallstovemodels,especiallyPraktiandGreenway,arecurringcomplaintwasthatthestovesizewastoosmallformosttasks(15/23and14/24).FortheGreenwaystoveanothermajorcomplaintwasthatthestovewasnotstable(13/24).WhenuserswerethenaskedaboutproblemscookingontheICS,usersfounditdifficulttochopwoodintosmallpiecesforthesestovetypes;thiswasnotedespeciallyfrequentlybyEco‐Chula(16/24)andEnvirofit(10/24)usersandtoalesserextentGreenway(7/24)andPrakti(8/23).SomeusersofthePraktistove(5/23)complainedtheycouldnotusethesecondpothole(thiswastheonlystovethathadthisissuebecauseitwastheonlystovewithtwopotholes).SomePrakti(3/23)andEcoZoom(5/23)usersreportedthatthefuelchamberwassmall.AfewusersofEco‐ChulaandGreenwayfounditdifficulttoignitethestoveevenafter21daysofregularuse.OtherproblemsforsomeGreenwaystoveuserswerethatashwouldbuildupquickly(10/24)andfuelwoodkeptfallingoffthetraywhilecooking(4/24).Beyondthesecomplaints,othercomplaintswerecommonacrossallstovetypes.Differencesbystovetypearehighlightedinthechartthatfollows(Figure12).


Figure12:*Cookingproblemsdenotedwithanasteriskareissuesthatareopportunitiesforconsumerawarenessandeducation,ratherthanchangesthatshouldactuallybemadetostoves,sinceenlargingthecombustionchamberandincreasingflameheightaredetrimentaltostoveperformanceintermsoffuelsavingsandemissionreductions.Rather,consumerscanbeeducatedonhowtogetthebestperformanceoutoftheirstovesusingmethodsthatoptimizeitsdesign.Afterusingthestoveforthreeweekstheusersprovidedsomesolutionsthattheyperceivewillmakethesestovemodelsbetterandmoreacceptable.Morethan90ofthe118usersforallstovemodelssaidthatthestovesshouldbelargerinsize.Theyreportedthatthecombustionchambershouldbelargerforallstovemodelssothatmorewoodcanbefedintothestove.SomeGreenwaystoveusers(4/24)suggestedthatthestovecouldbemademorestablebymakingthetopplatethickerandsturdiersotobetterbeartheweightofthepotsandvesselsplacedonthem.SomePraktistoveuserswantedmoreheatinthesecondpotmouthandsuggestedplacingthecombustionchamberbetweenfirstandsecondpotsothatbothpotscanbeused


forcooking(8/24).Asmallpercentageofuserswantedthestovestohavevisibleflameswhichwouldreachthepots,particularlyforEnvirofitandGreenway.Althoughsomeofthesuggestionswereconstructiveandwouldrequiresomesimpledesignchanges,othersuggestionsarebetteraddressedthroughconsumereducationatthepointofsaleandduringaftersalesserviceforthesestovemodels(Figure13).

Figures13:*Suggestedchangesdenotedwithanasteriskareissuesthatareopportunitiesforconsumerawarenessandeducation,ratherthanchangesthatshouldactuallybemadetostoves,sinceenlargingthecombustionchamberandincreasingflameheightaredetrimentaltostoveperformanceintermsoffuelsavingsandemissionreductions.Rather,consumerscanbeeducatedonhowtogetthebestperformanceoutoftheirstovesusingmethodsthatoptimizeitsdesign.Inthisstovetrialamajorityoftheuserswereforcedtomodifytheircookinghabitstoaccommodatethedesignsofthenewstoves.Insteadofmulti‐taskingduringcooking,59percentofusershadtositinfrontofthestovefortheentirecookingsession,addingwoodpiecesatregularshortintervals.Some31percentofstoveuserssaidtheyhadtoplantheircookingandprepareeverythinglikechoppingthevegetablesandcleaningthedaalandriceinadvancebeforestartingtheactualcooking,sincethenewstovesaffordedlesstimeformulti‐tasking.Only29percentofuserssaidtheydidnotneedtomakeanychangeincookingstyletousethesestovemodels(seechartbelow).


Chart2:ChangesinCookingPattern

ChangesinCookingPattern,ifAny,asAdaptedbytheUsersfortheNewStoves(AllStoves)

Frequency(n=118) %

Sitinfrontofthestovetocookeverything 70 59Prepareeverythingfirstandthencook 37 31

Nochange 34 29Perceptionsof“WhoistheStoveGoodfor?”and“WhatisitWorth?”Inadditiontoconductingarigorouswillingnesstopayexercise,stoveuserswerealsoaskedtodescribethekindofpersonthesestovesweregoodfor,aswellastoestimatehowmuchthestovewasworth.Thiswasdistinctfromwhethertheywantedtobuythestove,rathertheirestimateofitsvalue.Theusersreportedthatthestovemodelsweresmallinsize,andoverwhelminglysuggestedthatthesestovesaregoodforsmallfamilies.Tothedescriptionsofsmallfamilies,fewtosomeaddeddifferentotherdescriptions:“smallfamilieswhobuytheirwood,”“wholiveinurbanorperi‐urbanareas,”“whorentorlackspacetoinstallstovesoutdoors,”andafewsuggestedtheyarebestforbachelors!Anotherfewuserssaidthatsincethestovesareexpensive,thepeoplewhocanaffordthestoveswouldhavetobesalariedprofessionalpeopleorpeoplewhohaveagoodincome(seechartbelow).

Uponcompletionofthestudy,generalcommentsaboutthestovesincludedthattheyweregoodbuttoosmallfordailycookingsincetheaveragefamilysizefortheseruralhouseholdswasmorethanfive.Participantsreportedthattheywouldliketousethepresentsmallermodelsinthesummerandrainyseasonwhentheycannotcookoutside.Theynotedthattheyuseonlywoodfuel(gatheredandsavedduringwinter)duringthesemonthsandneedtosaveonfuelwood.Participantsrecommendedthatthesestoveswouldbeindemandwithsmallfamilies(57/118)and,toamuchlesserextent,mentionedtheyareappropriateinurbanandsemi‐urbanareas(9/118).Becauseofthelackofspaceinurbanareas,users(5/118)suggestedthosefamilieswouldwelcomeportablestovemodelsthatcanbeusedinsidetheapartment.Someparticipants(9/118)alsonotedthatwoodfuelisalmostexclusivelypurchasednotcollectedinurbanareas,andsincethesestovessavefuelthereshouldbeagooddemandforthesestoves(seechartbelow).


Chart3:PerceptionofPeopleWhoWouldUseTheseNewStovesWhatKindofPeopleWouldUseThis(These)NewStove(s)?

Frequencyn=118

%

Smallfamilies 57 48Modernpeople 52 44Thriftypeople 25 21Poorerpeople 11 9Simple,ordinaryfamily 23 19Someonepeoplerespect 11 9People/familieslivinginurbanorperi‐urbanareas 9 8Smallfamilieswhobuywood 9 8Smartpeople 8 7Peoplelivinginrentedorlackspaceoutdoors 5 4Professionalpeopleorpeoplewithgoodincomes 4 3Middleclassfamilies 3 3Bachelors 3 3

Studyparticipants“valued”stovesforcertainfeatures,butdramatically“undervalued”the(anticipated)priceofthestove.Many(49/111,or44percent)estimatedthestoveat0to25percentofanticipatedsalesprice(whichalreadyincludesanassumedcarbonsubsidy);another35/111(32percent)estimatedthevaluetobebetween26and50percentoftheanticipatedsalesprice.Only13/111(12percent)estimatedbetween51and75percent,and6/111(5percent)between76percentandthesalesprice.Fewoverestimatedpricingacrossallmodels(Figure14).However,aswillbefurtherdiscussedinthefindingsanddiscussionsection,reportedvalueswerelikelyinfluencedbyashared(andperhapsdiscussed)perceptionthatparticipantsshouldbegiventhestovesasatokenofappreciationforparticipatinginthestudy.


Figure14:Theseproportionsaregeneratedbydividingtheestimatedvaluebytheanticipatedsalesprice,thuscreatingavaluetocompareacrossstoves,whichrangedinprice.

Figure15:Studyparticipants’estimatesofstovepricesaredisplayedasaproportionofanticipatedsalesprice,bystovetype.Stoveswereforthemostpartundervaluedacrossallstovetypes,withtheexceptionofEcoZoom,whichconsumersthoughtwasworthmorethanotherstoves(thusitfellintothehigherratios),apparentlybecauseofitslargersizeandheavierweight;someconsumersmistakenlythoughttheentirestovewasmadeoutofcastiron,whichwoulddemandahigherprice.


WILLINGNESSTOPAYTwodifferentinnovativewillingnesstopay(WTP)assessments(basedonareviewofregionalandotherIAPWTPmethodsusedtodateanddiscussionswithexpertsincludingTRActiongrantrecipients)markedthecompletionofthestovetrial.15FinalformatsarefoundasAppendixB. Insevenvillages,all105householdsweregiventheopportunitytopurchasethestudystovesinabargainingexercisethatincludedinstallmentpaymentoptions,andinonevillageall15householdsweregiventhestoveasagift,butofferedcashto“sell”itback.Ofthe105householdsofferedthechancetobuytheirstove,12enteredintonegotiations,butonlyonestudyhome(andonenon‐studyhome)eventuallypurchasedthestove.Thewillingnesstopayformswereusedonlyinthese12households;therestdidnotwishtoevenbargaingiventhehighstatedvalueofthestove.Theywantedthestovemodelstobegiventothemforfreeasatokenofappreciationforhavingparticipatedinthestudyforthreetofourweeks.SomeuserssaidtheyhadparticipatedinthestudytohelptheconcernedorganizationstobringinnewstovemodelsintoBangladesh.Tohelpthestudy,theyusedwoodfuel,whichtheyeitherbuyorgatherandkeepforthesummerandrainyseason.Duringthewintermonthstheygenerallyusedryleafybiomassasfuel,whichisavailableforfree,inspeciallydesignedmudstovesthattheythemselvesconstructoutsideintheopencourtyardtosavemoney.Theuserswereveryvocalinstatingthatalthoughtheyknewthecostofthestovevariedfrom1600–4000Taka,theywantedtobuyitatasubsidizedpricerangingfrom200–500Taka,dependingonstovemodel.Therewereclearsignsof“collusion”anddiscussionamongstudyparticipantswithinvillagesandperhapsacrossstudyvillages.Theydidnotthinktheywouldbemakingagooddecisioniftheyboughtthesesmallerstovesatthequotedpricebecausethesestovesalthoughgoodwouldnotreplacethetraditionalstovefortheirdailycookingneeds.Itwouldremainanadditional/supplementarystoveforthefamily.Themajoritywerenotwillingtobargainornegotiate.Belowisadescriptionofthe12householdswhowereinterestedinpurchaseandengagedinthewillingnesstopay“bargain”butdidnotpurchaseastove,alongwiththeonestudyparticipantandonenon‐studyparticipantwhodidpurchasestoves.Inallhouseholdsbothhusbandandwifeparticipatedinthenegotiations,withhusbandendingupbeingthemainrespondentforbargaining. 15AfterreviewingallavailableIAPstudiesandamid‐depthreviewofthewillingnesstopayliterature,aswellasafewdiscussionswithresearchers,noneoftheresearcherswereparticularlysatisfiedwithavailableWTPassessmentmethodologies.Essentially,mostmethodologiesconsistofASKINGpeopleiftheywouldWANTtobuyaconsumeritem,andthenaskingwhatpricetheywouldpay.Themostpopular/bestpracticeofengagingparticipantsinan“auction”turnsoutnottobeatrueauction,buta“stepDOWNoffering”betweenparticipantandinterviewer(wouldyoupayX?okaythen,wouldyoupayX–$10?).IntheendtheitemissoldfortheSECONDhighestbid,tothehighestbidder.Onlyonepersonisallowedtobuyastoveintheend.Theconcernwiththisapproachisthat1)itdoesn’ttakeintoaccounttheacceptedcustomofbargainingforANYpurchase,wherethepurchaserwouldbeina“losingposition”torevealhowmuchtheyarewillingtopayforanitem;and2)itdoesn’ttakeintoaccounttheimportanceoffinancingtotriggerandenableapurchase.


Chart4:WillingnesstoPayforICS–Method1 WillingnesstoPay—BargainandFinancingOffered

DescriptionofThose12InterestedinBuying(2BoughtinEnd)

Village Stove LowestAcceptableOfferBasedonStoveValue(BDT)

InitialPriceOfferedbyParticipant(BDT)

FinalPrice/Resolution

SylhetDistrictKewa Prakti 3000 200 250

Finalpricedeclaredbythehusband[notpurchased,asfinalofferbelowthreshold]

Kewa EcoZoom 1600 300 Refusedtobargainbeyondthatpriceevenafterrepeatedpersuasion[notpurchased]

Kewa Greenway 2400 300 400afterlotofpersuasion[notpurchased,asfinalofferbelowthreshold]

Kunarchor EcoZoomalthoughusedGreenwayinstudy

1600 500 Husbandwasnotreadytonegotiate[notpurchased]

Kunarchor EcoZoom 1600 Wantedtoparticipatebutbackedoffafterhearingthestatedvalueofthestove[notpurchased]

Tilargaon Prakti 3000 5 installments 5 x 750

PURCHASED

Tilargaon Prakti 3000 3000 3000PURCHASEDLandladyofotherpurchaser

BarisalDistrictBillobari Prakti 3000 300 500afternegotiation[not

purchased,asfinalofferbelowthreshold]

Bihangal EcoZoom 1600 300 [notpurchased]Gonpara Eco‐Chula 4300 600 [notpurchased]Gonpara Greenway 2400 1200 Saidunabletopaymoreashe

waspoor[notpurchased]Gonpara Envirofit 2000 500 Refusedtonegotiatebeyond

thatprice[notpurchased]InonevillageinSylhethouseholdswereofferedtheirstudystovesasgifts,thengiventheoptionofsellingbackthestovesatthepricesdetailedbelow.In12ofthe15householdseveryfamilymemberoptedtoretainthestoves,irrespectiveofthestovemodel.Inonehouseholdthehusbandsaidthatalthoughhe,hiswife,andsonliketheEco‐Chulaandwouldliketokeepittheywerebeingforcedtosellsincehiswifehasaheartproblemandherequiresreadycashforhertreatment.Assuch,threeofthe15householdstradedtheirstovesforcash,andtheothersturneddownthemoneyinfavor


ofthestove.ThethreestovesthatwereexchangedforcashweretheGreenway(for2400Taka),theEco‐Chula(for4300Taka),andtheEnvirofit(for2000Taka).

Chart5:WillingnesstoPayPricingScenarios–Methods1&2Stovemodel Stove

value(US$)

Buy‐backoffer,lowestsaleprice(US$)

Stovevalue(BDtaka)

Buy‐backoffer,lowestsaleprice(BDtaka)

Prakti 70 38 5000 3000Greenway 45 29 3300 2400Envirofit 40 24 3000 2000Eco‐Chula 70 54 5000 4300EcoZoom 35 19 2600 1600

The“stovevalue”isbasedontheactualcostWASHpluspaidforthestove(notincludingshippingandhandling),plus$5/stoveforshipping(assumingbulkshippinginafuturemarketscenario),plusa10percentmark‐upforanationaldistributer,$4fortransportoutofDhaka,and10percentmark‐upforrural/localdistributer(upto$5).The“buy‐backoffer,lowestsalesprice”takesthathighendandsubtractspossiblecarbonrevenuefromit(assumingafour‐yearlifespanforthePraktistoveandtwoyearsforallothers,and$8/ton/stove/yearforcarbonpricing)toreacharealisticvaluethatthesestovescouldsellforintheBangladeshimarketoncemorewidelypromoted.Ofnote,theselifespansareconservativeestimates;accordingtomanufacturerspecifications,expectedlifespansforthesestovemodelscancommonlyreachfiveyears.SUMSSemi‐structuredsurveyinstrumentswerecomplementedbySUMS,temperature‐sensingdataloggersplacedonallinterventionstoves,alltraditionalstovesinthecontrolgroup,andonthetraditionalstovesin51percentoftheinterventionhomestotrackactualstoveuse.WithadditionalfundingfromtheU.S.StateDepartment’sOfficeoftheGlobalPartnershipInitiative,BerkeleyAirMonitoringGrouptrainedagroupof10fieldworkersinkitchenperformancetestprotocolandprocedures.TogetherwithaBerkeleyAirsupervisor,thisteamundertookkitchenperformancetestsinallbutfourstudyhouseholds,aswellas24controlhouseholds,totrackchangesinfueluse.BerkeleyAiralsooversawlimitedindoorairpollutionmonitoringandpersonalexposuremonitoring.SUMS‐measuredusageratesforallinterventionstoveswerebetween2.1(Envirofit)and3.3(Eco‐Chula)usesperdayduringtheKPTmonitoring,withallgroupsincludingthetraditionalstoveintheircookingsystemsbetween1.3and1.9timesperday.TheseusagepatternsduringKPTmonitoringsuggesttheinterventionstoveswerecommonlyusedbythestudyhouseholds,butinallcases,didnotfullydisplacetheuseofthetraditionalstoves.ThefollowingchartshowsthepercentageofcookingtasksperformedintheinterventionhomesthathadSUMSonbothinterventionandtraditionalstoves,bothduringandaftertheKPT.


Chart6:ProportionofAllRecordedCookingEventsPerformedbytheInterventionStove(bystovegroup)

%CookingperformedonICS:DuringKPT

%CookingperformedonICS:PostKPT

EcoZoom(n=9) 65% 34%

Prakti(n=11) 72% 43%

Eco‐Chula(n=11) 73% 46%

Envirofit(n=16) 60% 29%

Greenway(n=9) 69% 30%

Interestingly,oncethefieldteamsstoppedvisitingthetesthomesdailytotakefuelmeasurements,allstovegroups,includingthetraditionalstovecontrolhomes,showedamarkedreductionintheuseofanystoves,bothinterventionandtraditional.Thelargestdeclinewasseenintheuseoftheinterventionstoves(Figure16).

Figure16ThestoveuseratesreportedattheendofeachdayofKPTmonitoringwerecomparedtotheSUMSdatafromthesameperiod.Cookswithaninterventionstovewerelikelytounder‐reportuseofthetraditionalstovebutreporteduseoftheinterventionstovewithrelativeaccuracy(Figure17).

Stove Usage During and Following the KPT Study


Figure17 KitchenPerformanceTestingTheKPTwascarriedoutin140households.Afterremovalofinaccurateormissingdata,thefinalsamplesizewas134(Barisal:65andSylhet:69households).Allhouseholdsineachstovegroupusedwoodastheirmaincookingfuelduringthemonitoringperiod,withasmallnumberofhomesinBarisalreportingusingcropresidue(intheformofdriedleaves)asasecondaryfuel(9percent,n=12).TheKPTdatasuggestthatallinterventionstovegroupsexceptoneused16to30percentlessfuelperhouseholdperdaycomparedtothetraditionalstove‐usinghomes.16

16Itisimportanttonotethatthisdoesnotmeanthattheimprovedstovesused16–30percentlesswoodthanthetraditionalstoves.Rather,homesusingtheimprovedstovesalongsidetheirtraditionalstoves(whichiswhathappenedinmostoftheinterventionhouseholds)used16–30percentlesswoodthanhomesusingonlythetraditionalstoves.

Monitored vs. Reported Stove Usage


Chart7:MeanDailyFuelConsumptionEstimates(reportedaskgperstandardadult(SA)perdayandbyhousehold(HH)perday.±represents1standard

deviation)

Wood(kg/HH/day)

%savingscomparedtotradstove

Wood(kg/SA/day)

%savingscomparedtotradstove

Pvalue*

Traditionalstove(n=23)

3.09±1.69 ‐ 0.73±0.30 ‐ ‐

EcoZoom(n=22) 2.39±0.77 22.7 0.60±0.19 17.8 0.106

Prakti(n=22) 2.58±1.16 16.5 0.69±0.41 5.5 0.746

Eco‐Chula(n=22) 2.19±0.79 29.1 0.63±0.23 13.7 0.223

Envirofit(n=24) 3.63±1.24 ‐17.4 0.87±0.47 ‐19.2 0.214

Greenway(n=21) 2.32±0.94 24.9 0.62±0.22 15.1 0.217

* Comparing intervention stove with traditional stove for (kg/SA/day value). Equal variances assumed in all cases.

Aboxplotofthekgwood/standardadult/daybystovegroupwasexaminedforthepresenceofoutliersthatmighthaveanimpactonthesamplemean.Figure18belowidentifiesoneoutlier17(denotedascircles)inthePraktistovegroupandtwointheEnvirofitgroup.

Figure18

17Outlierisdefinedas1.5timestheinter‐quartilerangefromthethird(75th)quartile.

Box Plot Showing Fuel Consumption by Stove Group (Kg/SA/Day)


Althoughthesedatapointswerefoundtobevalid,theirremovalwasexplored,whichchangesthemeanwoodfuelconsumptionforthePraktistoveto0.63kg/SA/day(SD0.28n=21).Thisestimateis13.7percentlowerthanthewoodfuelconsumptioninthetraditionalstovehouseholds(versusa5.5percentreductionwhenthedatapointisincluded).RemovalofthetwooutliersintheEnvirofitstovegroupwouldreducethefuelconsumptionto0.77kg/SA/day(SD0.34n=22),anincreaseof5.5percentcomparedtothetraditionalstoveestimates(versusa19.2percentincreasewhentheoutliersareincluded).Ofnote,itwasexpectedthatallofthesestoveswouldachieveatleasta35percentreductioninfueluse,basedontheirlaboratoryperformance.18Ourresultsdonotshowthepercentreductionthateachstoveachieved,butratherthepercentreductioninfueluseinthehousehold.Giventhatweknowthatthehouseholdswereusingtheinterventionstovesandtraditionalstovesinparallel,wecannotsaywhethertheinterventionstoveswereperformingasexpected,intermsofefficiency,inthefield.Wedoknowthattheyweren’tmeetingcooks’needs,basedonthisparalleluse.TheoneexceptionistheEnvirofitstovegroup,whichwasfoundtouseMOREfuelthanthetraditionalstovegroup.Basedonthisstove’sperformanceinthelaboratory,itislikelythatthestudystoveswereinstalledorusedincorrectly,whichhighlightsthegraveimportanceofpropertrainingforbothstoveinstallersandstoveusers.Itisalsopossiblethatwoodsavingsfortheimprovedstoveswasactuallyhigherthanthedatasuggestduetounderreportedleaflitteruse,asfurtherexploredinthediscussionsection.Inaverylimitedsnapshotofindoorairpollution,alloftheinterventionstoveswereseentoreducekitchenconcentrationsofcarbonmonoxideandparticulates,althoughnottothehealth‐protectivelevelofWHOorEnvironmentalProtectionAgencyguidelines.Thepilotmeasurementsof24‐hourexposuretocarbonmonoxiderevealedlowexposurelevelsthatwerenothealththreatening,eveninhomeswithtraditionalstoves.ThekitchenconcentrationsofPM2.5andCOarereportedinthechartbelow.Thisexploratorydatashouldbeseenonlyasanindicativepilot,however,sincethereisonlyonehouseholdperstovetype(twointhetraditionalstovegroup).Withoutalargersamplesize,thecomparisonofhouseholdairpollutionlevelscanbemisleading,sincemanyofthefactorsthataffectpollutionlevelsvaryfromhometohome.Factorsincludeventilationrates,thesizeandtypeofkitchen,themixofstovesandfuelsused,thenumberofpeoplecookedfor,lighting,andotherindoorsourcesofpollution,suchasincenseandcigarettes.

18The2012ISOInternationalWorkshopAgreementforcookstoveperformanceprovidesasystemforcategorizingstovesbasedonseveralperformancemetrics,includingtwometricsrelatedtoefficiency,fromtier0representingtraditionalstovestotier4representingaspirationalgastechnologies.TheIWAtiersonlyprovidecomparativeclassificationforstovesbasedonlabtests.Allofthestovesselectedforthisstudyhadachievedatier2orhigherratingfortheirefficiencymetricsinthelaboratory.


Chart8:Mean24‐hourAirPollutantConcentrationsintheKitchen

PM2.5

(g/m3)

CO

(ppm)

TraditionalstoveHH1

11,017 31.5

TraditionalstoveHH2

2,737 14.1

EcoZoom 1,744 2.8

Prakti 626 9.1

Eco‐Chula 2,587 7.8

Envirofit 1,343 0.9

Greenway 1,472 3.2


DISCUSSIONLimitationsoftheStudyIdeally,astovetrialwouldbelongenoughforuserstotryoutastoveforseveralmonths,ratherthanseveralweeks,sinceitcantakethatlongtogetusedtoanewcookingapparatusandanewstyleofcooking,andusersmaynotsettleintonewusepatternsforanumberofmonths.Givenbudgetandtimingconstraints,wewerenotabletoextendthetrialbeyondthreeweeks.Forthecurrentstudy,WASHplusselectedfivepromisingimprovedstovemodelsbasedontheirlaboratoryperformancetestingresultsandtheiracceptanceelsewhereintheregionandbeyond.WewerenotabletoincludeinthestudythetrulyaspirationalBioLiteHomeStove,whichmayhavebeenverypopularinBangladesh,givencellphonepenetrationratesandthestove’sabilitytorechargecellphones.TheBioLiteisnowbackinproduction,andwerecommenditsinclusioninafuturestovetrial.Inaddition,giventhemixedreceptionoftheimprovedstovestrialedinthisstudy,andclearunwillingnesstopayfortheimprovedstoves,WASHplusrecommendsthattheseimprovedstovesbecomparedwithBondhuChulamodelstoassessrelativepreferencesandperformance,sinceourpolicy‐makingandprogramimplementingaudiencemayinterpretfromthefindingsofthisreportthattheyshouldcontinuetopromoteBonduChulastovesgoingforward,despitetheirmediocrefieldperformance.Furthermore,WASHplusrecommendsthatlargerandhigherfirepowertwo‐potstovesbetrialedinBangladesh.Althoughwedidincludeatwo‐potstoveinthetrial,userscomplainedthatthesecondburnerdidnotburnhotenoughtoboilwaterorcookrice.Becauseofdependenceonfreeagrofuels,trialingofaricehuskand/ormixedfuelstoveissuggested.Lastly,whilethestudywasdesignedtogenerallyapplytoallwood‐burningstoveusers,theresultsofthisstudyintwosmalldistrictsofBangladeshcannotnecessarilybeextrapolatedtotheentirecountry;furthertrialsshouldbeundertakeninotherpartsofthecountrytoaddmoredatapoints.FuelUse/SeasonalityOverall,studyparticipantsfeltthattheimprovedstovestrialedwerenotbigenough(intermsofbothphysicalsizeandfirepower)tocookthemealsneeded.Thiswasinpartbecauseduringtheseasoninwhichthestudytookplace,whentheweatherisrelativelycoolerandfoodthereforekeepslonger,householdsprefertocookriceforthewholedayallatonceinthemorning,ratherthanateachmealtime,asiscustomaryduringtherestoftheyear.Noneofthestudystovesweredesignedtocook1.5kgofriceatonetimeanddidnotmeetusersatisfactionforthattask.Assuch,mosthouseholdsendedupusingthestudystovesforsidedishes,whilecontinuingtocookriceintheirtraditionalstoves.

Duringthestudyperiod,householdswerealsoaccustomedtosupplementingtheirwoodusewithfreegatheredleaflitterforfuel,enablingthemtosaveupfuelwoodforthesummerrainyseasonduringApril–August/September,whendrywoodoragrowasteishardertocomeby.Soalthoughmoststudyparticipantsreportedthattheimprovedstovesusedlesswoodthentheirtraditionalstoves,theywouldhave


preferredtobeusingleaflitter,andinfactdidsointheirtraditionalstoves.Theabilitytoburnleaflitterintraditionalstovesmayhavebeenasecondarycontributingfactortohighertraditionalstoveusageandlowerimprovedstoveacceptance,preference,andusageduringthestudyperiod.Weonlydiscoveredduringthecourseofthestudythatmanyhouseholdsbuiltspecialleaf‐burningstovesforuseduringDecember–February.Thesestoveswerebuiltawayfromthehousesinanopencourtyardand/orsemi‐enclosedspacebecauseofthethicksmokethatburningleavescancreate.Inaddition,resultsfromtheDay3andDay21qualitativesurveyssuggestthatleaveswereusedextensivelyforcookingfuelduringthetimetheKPTswereconducted.Thiscouldhavepotentiallyinfluencedtheextentandnatureofwoodfuelconsumptioninboththecontrolandinterventionstovehouseholds.Thatsaid,theKPTteamdidnotseemanyhousesusingleaves.Theteamweighedtheleaveswheneverthehouseholdreportedusingthem,whetherusedintheiroutsidestoveorinthetraditionalstoveintheirhouse.Some12ofthe134totalhouseholdshadleavesweighedduringtheKPT,andall12ofthesehouseholdswereinBarisal.ThisfactisinlinewiththefieldteamreportsthatleavesareusedasacookingfuelinadditiontowoodinBarisalduetotheabundanceoftreesthere,whileinSylhet,wheretreesarelessabundant,leafuseislesscommon.Onlyoneofthese12householdswasacontrolhousehold.Itispossiblethatsomeunreported,andthereforeunmeasured,leafusetookplaceduringtheKPT.TheKPTteamdidnothaveanystrongevidencetoindicatethatthisoccurred,orthatitoccurredmoreforleavesthanforwood(e.g.,smallsticksandtwigs).Alsothis“leakage”inthemeasurementofleaflitterislikelytohavehadasimilarimpactacrossallstovegroups,includingthecontrolhouseholds,sothatanybiasisspreadoutacrossthestudypopulation.Thereisalsothepossibilitythattheprocessofmonitoringinfluencedthewaythehouseholdusedtheirstovesandfuels.IfthehouseholdsperceivedthemainfocusoftheKPTtobewoodfuel(eventhoughtheyhadbeenaskedtoshowallfuelstobeweighed),theycouldhavealteredtheirhabitstousemorewoodandlessleavesfortheperiodoftheKPT.TheSUMSdatasupportthishypothesisinthattherewasareductioninstoveuse(includingbothtraditionalandimprovedstoves)inallhouseholdsaftertheendoftheKPTmonitoring.Thissuggeststhattheymayhavemovedfromtheimprovedandtraditionalwoodburningstoves,whichhadSUMSunitsplacedonthem,totheleafburningstove,whichwasnotmonitored(duetofearstheymightbestolenasthestovewaslocatedoutside).Theextenttowhichleafusereduceswooduseisnotclear.AsdemonstratedinFigure8,thetwodistrictshaddifferentstovepreferences.InSylhethouseholdsoftenhavechimneyhoodsunderwhichtheyusetheirtraditionalstovesinkitchensattachedtothemainhousehold.Becausetheywereaccustomedtocookingindoorswithoutheavysmokeemissions,thePraktistovewasthebestmatchforthem(asithasachimney,socanbeusedinsidewithverylittleindoorsmoke).Thesehouseholdsthoughttheotherstovestoosmoky,especiallyduringthelightingprocess.InBarisalhouseholdsusuallycookoutsideinsemi‐enclosedspaces,orseparaterooms,ratherthaninthemainhouse/livingspace.SinceitrainsalotinBarisal,andthey’reaccustomedtocookingoutofthelivingarea,theyliketheportabilityoftheEco‐Chula.


OnesurprisingfindingfromthestudywasthedramaticdecreaseinacceptanceofALLtheimprovedstovesbetweentheDay3andDay21surveys.ThiswasespeciallypronouncedfortheEco‐Chulastove.OurreportingindicatedthatpeopleinitiallylikedtheEco‐Chulabecauseitwasportableandcleanburningwithlittlesmoke(especiallyvaluableinBarisalwherepeoplecookinsemi‐enclosedareas),butthatovertimetheygrewtoresenthavingtochopwoodintosmallpieces,asrequiredbythestove,andhavingtositbythestovecontinuouslyaddingwoodpieces,ratherthanbeingabletomulti‐taskastheywereaccustomedtodoingwiththeirtraditionalstove.WillingnesstoPayAninitialinterpretationofthewillingnesstopayfindingssuggeststhatwhenacquisitionbarriersareremoved(asmodeledinthesecondWTP“buy‐back”scenario,wherehouseholdswere“given”thestoveandthenofferedasumofmoneyto“buyitback”),householdsvaluedthestoveshighly.Priceswereidenticalinbothcases;thelowestpricetheteamwouldacceptforsellingthestoveinscenariooneequaledtheofferedpriceforwhichtheteamwouldbuybackthestoveinscenariotwo,asperthetablefoundintheWTPfindingssection.Peoplesawbenefitsandpositiveattributestothestoves.Some,butnotamajority,preferredtheICStotraditionalstoves.Buteveryoneunderestimatedthemonetaryvalueofthestove,andfewwerewillingtopayanythingclosetomarketvalueforthestove,evenwhenofferedinstallmentoptions(thatincludedinterestof20percentoverfiveinstallmentpayments).Thiswassurprising,astheauthorshypothesizedthatfinanceoptionsnotoftenavailableforstoveswouldincreasetheirappealandacquisition.Accordingtothestoveusers,theylikethestovesandwouldhavelikedtokeepthemifgivenforfreeoratanominalcost.Improvedstoveswouldnotreplacetraditionalstoves,rathercomplementtheiruseundervariousconditions.Householdersrealizedthatthesemetalstovesareexpensive,buttheywerenotreadytobuythematmarketprice.Thevariousreasonsputforwardbythemwere,asfollows:

a. Thestovemodelissmallandcannotcompletelyreplacetheprimarystove.Itwillbeasupplementarystovethatwillbeveryusefulinthesummerandrainyseasonwhenwoodistheonlyfuelandcookingneedstobedoneindoors.

b. Theyhadparticipatedinthestudy,sothestoveshouldbegiventothematanominalpriceorfree.

c. Theycouldnotriskpayingsomuchmoneyforanexperimentalmodelsinceafterthestudytherewillbenoaftersalesservice.

d. Theydonotwanttobuythestovesoninstallments(oratleastontheinstallmentplanoffered)sincetheydidpaymentcalculationsintheirheadsandrealizedtheywouldenduppayingmuchmoreforthestoveonceinterestand/orservicefeefortheloanwasincorporated.

Fundamentally,studyparticipantsviewedtheWASHplusstudyfieldteamasNGOstaff,andinruralBangladeshthereisastrongculture/backgroundofNGOsgivingawayoratleastsubsidizinggoodsandservices.Assuch,studyparticipantsstronglyfeltthattheyshouldbegiventheimprovedstovesfreeorataheavilydiscountedrate.Thiswasespeciallytrueincaseswheretheyfelttheimprovedstovewouldonlybeusedforspecifictasks,andwasnotatotalcookingsolution.


Oursecondwillingnesstopayassessment(thebuy‐backexercise),however,demonstratedthatconsumersDIDtremendouslyvaluetheimprovedstoves,oncetheyownedthem.Giventhelowpurchaserateofthestoves,theteamwassurprisedtofindthatsomanyfamiliesoptedtokeeptheirgiftedstoves,ratherthanexchangingthemforcash.Thisleadtheteamtoobservethattheparticipantsvaluedthestoveandpreferreditoveritscashequivalentwhentheydidnothavetomakesacrificestotheirhouseholdeconomytokeepitandwhentheydidnothavetocomeupwithfundsfromwhatwasanalreadytighthouseholdbudgetinmostcases.Ofnote,participationinawomen’sgrouphadaslightbutnotsignificantpositivecorrelationwithexpressedinterestinpurchasingastove(independentofwhethertheyultimatelydidpurchaseastove).Almosthalf(55/120)oftheparticipantsbelongtosomesortofwomen’sgroup(suchassavingscooperatives).Ofthiswomen’sgroupsubset,60percentexpressedinterestinpurchasingastove,comparedto55.4percentofthe65womenwhodidnotbelongtoawomen’sgroup.Withsuchasmallsamplesize,it’sverydifficulttosaywhetherwomen’sgroupparticipationhadanyrealinfluenceonstovepurchasedecisions.Thesewomen’sgroupsarestillbeingconsideredasapotential“vehicle”forpromotionanddistributionofstoves.Thebaselinesurveyalsoaskedwhetherrespondentswereabletoindependentlymakethedecisionwhetherornottopurchaseastove.Thegraphbelowshowstheresultsbyagebrackets.Womenparticipantswithinthe16–25agebracketwereprobablynewlymarriedorunmarrieddaughterswhodidnotyethaveauthoritytomakedecisionsforthefamily.Thegraphshowsagradualriseindecisionmakingpowersasageincreasesandthenaslightdipwhentheolderwomenlikelybegintohandoverhouseholdresponsibilitiestothedaughter‐in‐law(Figure19).Outof120studyparticipants,113answeredthequestion“Areyouthepersonthatwouldmakethedecisiontopurchase[theimprovedstove]?”Outof31womenrespondentswithinthe16–25agebracket,onlyaboutaquartersaidyes.Outof38respondentswithinthe26–35agebracket,themajoritysaidyes.Outof26respondentswithinthe36–45agebracket,almostallsaidyes.Outof14respondentswithinthe46–55agebracket,mostsaidyes.Outoffourrespondentswithinthe56–65agebracket,mostallconfirmedtheywereabletoindependentlymakethedecisionwhetherornottopurchaseastove(Figure19).


Percentage of Women Respondents Reporting Independent Decision Making on Household Purchases, by Age Group

Figure19CONCLUSIONSANDNEXTSTEPSAstheevidencebaselinkingimprovedcookstoveswithimprovedhealthandenergyimpactsgrows,sodoesattentiononhowbesttoinfluencehouseholduptakeandconsistentandcorrectuseofstoves.Appropriately,attentionfocusesonhowtoimprovestovesandmakethemmoreaffordableandappealingtotheneediestconsumers.Importantworkhasstarted,includinginSouthAsia,toidentifythe“drivers”ofcookstoveadoption.UnderPhase2ofWASHplusactivities,WASHpluswilldevelopagenericmarketingandbehaviorchangestrategy;suggestalimitednumberofevidence‐basedapproachestoincreasetheuptakeofstoves;concepttestkeyelementsoftheseapproaches;anddeveloppractical“how‐to”toolstocontributetothegoalsandresultsofUSAIDenergyandhealthobjectivesinBangladesh.ThiswilldrawonlessonslearnedinBangladeshandothercountriesintheSouthAsiaregioninbehaviorchange,demandcreation,andmarketingofsanitation,watertreatmentproducts,andcookstoves.Amarketingstrategyaddresseswhatisclassicallyreferredtoasthe4Psofmarketing—product,place,price,andpromotion—tosuggestavibrant“marketingmix”ofelementsthatwillmakeimprovedcookstovesappealingandaffordabletothemostvulnerableBangladeshimarket.Tohighlightthewayforwardandhowthesefindingswillbeapplied,we“preview”someapplicationsinthefollowparagraphs.Thestudyhasshedlightonsomeessentialchangestoallfivestovesbeforetheyareappealingenoughtoconsumersforthemtoopentheirpursestopurchasethem,andbeforetheyareabletousethemconsistentlyandcorrectly.

n=31 n=38 n=26 n=15 n=4


Thestovemanyconsumers“want”islargeandstable,yetportable,withamodern,well‐madedesign.Itcookslargevolumesoffoodandhastwoactiveburners.Flamesarehotandleaptotouchthepots,butdonotmakethemblack.Thestovecanusemultiplefuels,whichcanbefedinrelativelyunprocessed.Woodinparticularcanbestuffedinlargepieces,andlefttofeedalmostautomatically.

Thestovedescribedabovewouldbeconsideredthe“ideal”formanyinthestudy,butnotwhatiscurrentlyavailablenornecessarilywhatweareaimingfor.Someofthecriteriaarecontradictory(leapingflamesandcleanpots),thermo‐dynamicallyimpossible,undesirablefromafuel‐efficiencystandpoint,andfarfromwhatiscurrentlyavailableinthemarket.Thefivestovestestedbyconsumersoverthethree‐weektrialperiodmetsomeofthedesiredattributesdescribed,butusersexpressedmanyproblemswiththecurrentICSmodels.Someoftheproblemsandsuggestedchangescanbeaddressedbymanufacturerswithoutmucheffortor“R&D”(researchanddevelopment),suchasstabilizingthestoveand/orenlargingtheburnertoaccommodatelargerpots,oraddinganashtrytocatchburntashes.However,otherproblemsandsuggestedchangesrevealconsumerpreferencebutarenotrecommendedchanges,becausetheywillclearlyaffecttheefficiencyandemissionsofthestoves.Thisdoesnotmeantheyshouldbedismissed;however,theyclearlyindicatearangeofeducationandinformationthatshouldbedeliveredtoconsumersthroughpointofpurchasesalesmaterialsandinteractionwithdistributorsandsalespeople,inpromotionalmaterial,andthroughhealthorotheroutreachactivities.AddressingsuchissueswillbeessentialforconsistentandcorrectuseofICS,forconsumersatisfaction,andrelatedword‐of‐mouthrecommendations.Understandingthepreferenceandobstaclestopurchaseandconsistentandcorrectuseofstovesfeedsdirectlyintopromotionstrategies,aswell.Althoughthestudydidnotrevealmuchaboutthedriversofstoveuptake(becausethestoveswerenotparticularlywellreceived),whentakenincontextwithwhatisalreadyknownaboutdriversofadoption19wehaveidentifiedkeyissuesaroundseasonalityofcookingpatternsandoffreefuelavailabilitysuchasleaves;ofthevisualappealofthemetalstovesas“wellmade,nicelooking,andmodern”;andanumberofotherfeaturesthatwerelikedordislikedandcanserveasthefoundationofpromotionalappeals.Thisstudyhasalsoidentifiedconsumergroupsmostlikelytobeinterestedandempoweredtopurchaseimprovedcookstoves,andpromotionalstrategiesshouldtargetthesepotentialearlyadopters:smallfamilies,especiallypoorbutnotdestituteperi‐urbanfamilies,headedby30‐to55‐year‐oldwomen.Thestudyhasalsorevealedthelowwillingnesstopayforstoves(price),atleastforcurrentmodels.Financingoptionswereexploredbutinconclusivebecauseconsumerswerenotparticularlyinterestedinbuyingsucha(relatively)highpriceditemthatdidn’tdeliverdesiredbenefits.Thefewinterestedinbuyingvaluedthestoveshigherthantheywerewillingtopay,expectingsomesortofsubsidyfromtheNGOsbringingthemtotestformarket.Overalltheyrejectedinstallmentswithanysignificantinterestorloanservicecostsattachedtotheloans.

19 LewisandPattanayak.2012.EnvironHealthPerspectives.


Placereferstotheplaceofsalesanddistribution.AgainthestudyshowsthatNGOsmayhaveanunintendedeffectonpoorhouseholds’willingnesstopay.AlthoughconsumersinthisstudytrustedNGOs,theyputtheminacategoryof“do‐goodersforthepeople”andnotsalespersons.Inthemindoftheparticipantsofthisstudy,NGOsgivethingsaway;theydonotfinanceorsellthem.Itispredictablethatallresearchpresentationsendwiththecallformoreresearch,andthisstudyisnodifferent,sinceitgeneratesanothersetofquestionsandareastoinvestigate.WASHplusoffersthefindingsofthisfinalreporttoarangeofstakeholdersinBangladesh,includingUSAID’sCatalyzingCleanEnergyforBangladeshprojectandplanstodiscussthefindingsinvariousforumswiththehopesofapplyinglessonsanddeepeninglearning.Andwhilesomestakeholderstakestockofthesefindingsandotherinputstodate,andmoveforwardwithevidence‐basedinterventions,otherscanfurthertheappliedresearchagendainacoordinated,paralleltracktoadvanceoursystematicunderstandingofthemarketdriversandconsumercontextthatwillopentheimprovedcookstovemarketinBangladeshandsupportuptakeanduse.


AnnexA

SelectionCriteriaPartnerNGOSelectionCriteria- TheNGOmustworkinoneofthethreedivisionswherewoodismostprominently

usedasaprimaryfuelsourceforcooking(Sylhetdivision,Chittagongdivision,Barisaldivision,datafrom2011DHSsurvey)

- TheNGOmusthaveanofficeoron‐the‐groundstaffinoneofthethreedivisionswherewoodismostprominentlyusedasaprimaryfuelsourceforcooking

- NGOmusthavepreviousexperiencewithworkingatthehouseholdlevelandgoodcommunityrelationships.

CommunitySelectionCriteria- Thevillagesmustbelocatedinoneofthethreedivisionswherewoodismost

prominentlyusedasaprimaryfuelsource- Mustbewithingeographicrangeofpartnerorganization- MustbeaccessiblebyroadinOct/Nov- Atleastoneofthevillageschosenmustnothavehadanyformerstoveinterventions.

AtleastoneofthevillageschosenmusthavehadaBCSIR/bondhuchulhainterventioninthepast

- Villageswillbeselectedreflectingthereligiousmake‐upofBangladesh(primarilyMuslimwithsomeHindu)

HouseholdSelectionCriteria- Primarycookingfuelmustbewood- Musthaveatleastfourpeopleinthehousehold(averageHHsizeinBangladesh=

4.4people)w/childunder5- Themajorityofthehouseholdsselectedshouldhavenopriorexperiencewithor

ownershipofICS;approximately10percentofthetotalsamplewillbepurposivelyselectedforhavingusedanICSpreviously,forpurposesofcomparison.Note:WecouldnotfindbondhuchulaorotherICSprevioususerstorecruit.

- IswillingtoparticipateintrialsCookstoveSelectionCriteria- FollowthecriteriaoftheInternationalStandardsOrganization(ISO)International

WorkshoponCleanandEfficientCookstovesInternationalWorkshopAgreement(ISO–IWA)—theinternationallyagreeduponcookstovestandardsandprotocols:http://pciaonline.org/files/ISO‐IWA‐Cookstoves.pdf

- ThestovesmustmeeteithertheTIER2orTIER3requirementsinISO‐IWA- Maybeeitherportablestovesthatmaybemetalandcapableofcookingmealsforat

least4‐6personsorfixedstoveswithchimneysthatareeasytoinstall(maybein2/3piecesforeasycleaning.)

- Woodstovesbutmayalsobemulti‐fuel


AnnexB

Allsurveyinstrumentsandquestionnairesavailableonrequestandinfullreport.

COST/WILLINGNESS TO PAY SECTION House # Village #

Interviewer: Unique ID# ______________

Select the ROW of the stove being used by your respondent. Begin by repeating that they have the opportunity to buy the stove, but are under no

obligation at all.

Then say:

1. This stove is worth [insert the value in column A] 2. But because you’ve participated in the study, and because there is only limited

servicing available on the stove at this point, we can offer it to you for ________ [insert the value in column labeled B]

3. Would you like to buy the stove? Record in column C [check X if yes, make – if no and proceed if no]

4. We can offer installment payments if easier. Are you interested in the stove if you could buy it for [say the amount in column D] __________ RECORD response in column E [check X if yes, make – if no and proceed if no]

5. You know, here in Barisol/Sylhet, we never buy for the asking price, of course we bargain. So please, I invite you to bargain with me and tell me what you are wanting to pay.

6. Note amount in column F, then accept if above the minimum, or bargain using columns G and H.

7. Note if they accept that price in column I. 8. Give last chance to counter-bargain. Note in J. 9. If still no, go back to the thank you and close the questionnaire. 10. If yes, make the financing/payment arrangement.

A B C D E F G H I J K L

STOVE

TYPE

Value

(BDT)

.

Offering

price

(BDT)

2.

1.

A

c

c

e

p

t?

Offer

payments

(calculate

at 20%

interest)

2.

2.

A

c

c

e

p

t?

Invite

bargaining.

Note below

the price

they offer

IF they

offer …

(circle

which)

(BDT)

Your

counter

offer

(BDT)

2.3.

They

agree

Offered

price

(BDT)

2.4.

They

make final

offer?

(note)

IF NO,

PUT ‘X’

Agree to

anything

this

amount

or above

(BDT)

2.7. They

request

installment

payments

for your

counter

offer? Note

and accept

if above

min

Stove 1:

Prakti

5600 5000 5x1200 4000

3500

4500

4000

3000

Or


3000 okay 5x750

Stove 2:

Greenway

3700 3300 5x800 2500

2200

2000

3000

2600

2500

2400

Or

5x600

Stove 3:

Envirofit

3300 3000 5x700 2500

2200

2000

2800

2500

okay

2000

Or

5x500

Stove 4:

Eco‐Chula

5600 5000 5x 1200 4500

4000

3500

5000

4500

okay

4300

Or

5x1050

Stove 5:

EcoZoom

2900 2600 5x 600 2000

1800

1500

2500

2100

1600

1600

Or

5x400

A B C D E F G H I J K L

Interviewer: House # Village #

Unique ID# ______________

Before beginning, please find the row that corresponds to the stove given to the respondent. Then

say:

1. We thank you for your participation in this survey, and as part of our thank you, we’re giving you this stove. It’s actually valued at [pick from column B] _________

A B C

Value Buy

Back*

Prakti 5000 3000

Greenway 3300 2400

Envirofit 3000 2000

Eco‐Chula 5000 4300

EcoZoom 2600 1600 Buy back value considers potential but realistic carbon credit subsidy in pricing.

2. Note any reaction. Wait a little bit, like one minute, before offering the buy back. __________________________________________________________________


__________________________________________________________________ __________________________________________________________________ __________________________________________________________________

3. Now say: As an alternative, if you don’t want to keep the stove, we can buy it back from you and give you cash. The amount is a bit lower because the stove is now used of course, so it’s not worth as much. It’s completely your choice. Would you prefer the stove, or XX [select the corresponding amount from column C]. Note choice, and any reaction. [ ] Chooses stove [ ] Chooses cash

Reactions:

[ ] Asks for a different stove [ ] Tries to demand full pricing [ ] Other __________________________________________________________________ __________________________________________________________________ __________________________________________________________________ __________________________________________________________________


AnnexC

ATTRIBUTES NAMED BY ECOZOOM

USERS WHO PREFERRED ICS OVER

TRADITIONAL STOVE: When comparing

EcoZoom to their traditional stove, a

majority mentioned it uses less fuel,

over half said it emits less smoke and

they appreciate its portability, and

some said it looks nice.

PREFERENCE RATES: Just under half

(9/19) preferred the EcoZoom to their

traditional stove after three days; this

slipped to 8/19 after three weeks. Of

note, the EcoZoom and the Prakti had

the smallest decrease in preference

rates out of the five stoves.

PROBLEMS: Whether they preferred the EcoZoom to their traditional stove or not, many

experienced some problems with the stove. Users were most bothered by the difficulty of cooking

large quantities of food in bigger pots on the stove. Specifically, 15/23 said it takes longer to cook

(on ICS than traditional stoves) with large vessels and large amounts of food. Ash buildup was a big

problem for EcoZoom users (9/23). Less often than with other ICSs, some users (8/23) said that big

pots cannot be used as the flame does not spread. Related to this same issue, a few (4/23) also said

they had problems cooking rice in large quantities. Lastly, some (5/23) mentioned problems of wood

slipping out due to the slant of the opening. Just a few users found chopping wood into small pieces

(3/23) and the small fuel chamber (4/23) to be a problem.

SUGGESTED IMPROVEMENTS: The majority (17/23) said the stove should be larger to accommodate

bigger pots and bigger families. They suggested addressing wood falling out and not self‐feeding;

specifically, they suggested the place for entering the wood in the stove should be slanting inwards

to prevent the wood pieces from falling out of the chamber (10/23), and also an ash tray to make it

easier to remove the ash (4/23).

As with other stoves, users wanted to have a larger combustion chamber to add wider and bigger

wood (16/23). A few said the flame should reach the vessel bottom and spread (4/23). These final

suggestions fall into the category of suggestions that will be taken into account and addressed

through sales and education efforts, but not implemented because it would diminish the

effectiveness of the stove.


ATTRIBUTES NAMED BY PRAKTI


TRADITIONAL STOVE: Everyone

commented that the Prakti stove

emits less smoke, and a majority

also mentioned that it leaves the

house cleaner. Half said it looks nice

and just under half said it uses less

fuel.

PREFERENCE RATES: At three days,

just over half of those trying the

Prakti said they preferred the stove

compared to their traditional stove.

After three weeks, this slipped

slightly to just under half (10/21) of the users.

PROBLEMS: Whether they preferred the Prakti to their traditional stove or not, many experienced

some problems with the stove. Users were most bothered by the difficulties in cooking large

quantities of food in bigger pots on the stove. Specifically 17/23 said it takes longer to cook (on ICS

than traditional stoves) with large vessels and large amounts of food and that big pots cannot be

used as the flame does not spread (15/23). Related to this same issue, some (7/23) specifically

addressed problems cooking rice in large quantities. Users of the Prakti stove complained that they

could not use the second pot for cooking (5/23), and also that they needed to chop the wood (5/23)

and could not use large wood pieces. Some mentioned problems of ash buildup (6/23) and wood

pieces slipping out (6/23) due to the slant of the opening. Lastly, some (3/23) complained that the

fuel chamber was too small and that they wanted to add more wood than the existing chamber

allowed.

SUGGESTED IMPROVEMENTS: Most users (18/23) said the stove should be larger to accommodate

the cooking needs of big families. As mentioned above as a problem, many (15/23) wanted the

chamber to be bigger/wider to allow for larger pieces of wood (this suggestion falls into the category

of suggestions that will be taken into account but not implemented because it would diminish the

efficiency of the stove). Many suggested addressing the problem of wood “falling out” of the entry;

specifically, they wanted a slanted entry to hold the wood and have it “self‐feed” (10/23). Many had

suggestions about changing the stove to make the second pot more functional. Related to this, some

suggested that the combustion chamber should be between the first and second pot so that both

pots can be used for cooking, and an equal number said that the second pot should have more heat

for cooking (8/23). A few suggested adding an ash tray to make it easier to remove the ash (2/23).


ATTRIBUTES NAMED BY GREENWAY


TRADITIONAL STOVE: Regarding the

Greenway, a majority commented

that less fuel was needed, that it

looks nice, and half said they liked

the portability. Some, but not a

majority, mentioned that it emits less

smoke.

PREFERENCE RATES: Just under half

(10/21) preferred the Greenway to

their traditional stove after three

days, and this fell slightly after three

weeks to 7/21.

PROBLEMS: Whether they preferred the Greenway to their traditional stove or not, many

experienced some problems with the stove. Users were most bothered by the difficulties in cooking

large quantities of food in bigger pots on the stove, more than with all other stoves. Specifically,

19/24 said it takes longer to cook (on ICS than traditional stoves) with large vessels and large

amounts of food, and almost everyone complained that big pots cannot be used as the flame does

not spread (22/24). Almost half (10/24) mentioned problems of ash buildup (most probably due to

overfeeding), more than any other stove. The most critical complaint, not as much for user

satisfaction as for safety concerns, was the large number (9/24) (far more than any other stove)

saying the stove was not stable when stirring pots, requiring pots to be held when stirring to avoid

the pot falling from the burner. Some (3/24) complained that wood or embers fall off the tray, and

6/24 users found it difficult to chop the wood into small pieces and complained that they could not

use large pieces of wood. A small group of users (2/24) said that pots become black.

SUGGESTED IMPROVEMENTS: More than any other stove, the vast majority (21/24) said the stove

should be larger to accommodate bigger pots and bigger families. They suggested addressing the

problem of wood pieces falling out of the chamber by slanting the place for introducing wood into

the stove (13/24). As with the other stoves, many (16/24) wanted the chamber to be bigger/wider to

allow for larger pieces of wood (this suggestion falls into the category of suggestions that will be

taken into account and addressed through sales and/or education efforts, but not implemented

because it would diminish the efficiency of the stove). A few (4/24) suggested adding an ash tray to

make it easier to remove the ash, and a few also said the flame should reach the vessel bottom and

spread (5/24). A few strongly suggested dealing with the stability issue by making the plate with

“stands” on top thicker to prevent vessels from sliding and tipping over (4/24).


ATTRIBUTES NAMED BY ECO‐CHULA


TRADITIONAL STOVE: The popular

Eco‐Chula was preferred over the

traditional stove because it uses less

wood, emits less smoke, and looks

nice. Over half also mentioned the

house was cleaner than when using

the traditional stove, and that it

cooks food quickly.

PREFERENCE RATES: Many (16/20)

preferred the Eco‐Chula to their

traditional stove after three days,

and although it was the most

popular of all the stoves, it fell

dramatically in preference after

three weeks, to 10/20.

PROBLEMS: Whether they preferred the Eco‐Chula to their traditional stove or not, many

experienced some problems with the stove. Many Eco‐Chula users (16/24) found chopping the wood

into small pieces very difficult, and 3/24 users found it difficult to ignite the stove, even after using it

for 21 days. Still an issue but less so than with other stoves, users were bothered by the difficulties in

cooking large quantities of food in bigger pots. Specifically, 11/24 said it takes longer to cook (on ICS

than traditional stoves) with large vessels and large amounts of food, and some said that big pots

cannot be used as the flame does not spread (8/24), but again less than with other ICS. Unlike other

ICS, few other problems were mentioned.

SUGGESTED IMPROVEMENTS: As with other stoves, the vast majority (20/24) said the stove should

be larger to accommodate bigger pots and bigger families. Related to the cooking capacity, many

(15/24) wanted the fuel chamber to be bigger/wider to accommodate more and bigger wood,

something that needs to be addressed but not by making the chamber bigger (which would

negatively affect the stove’s efficiency).

Less than other stove users, only a few suggested the opening for entering the wood in the stove be

changed (slanted) to prevent the wood pieces from falling out of the chamber (4/24), and ash and

flame size were not particularly problematic with the Eco‐Chula.


ATTRIBUTES NAMED BY

ENVIROFIT USERS WHO

PREFERRED ICS OVER

TRADITIONAL STOVE: More than

half of users said the Envirofit

stove uses less fuel, looks nice

and keeps the house cleaner.

Some, but not a majority,

mentioned that it emits less

smoke and is well manufactured.

PREFERENCE RATES: More than

half (11/20) preferred the

Envirofit to their traditional stove

after three days, but this dropped

to less than half after three weeks, to 6/20.

PROBLEMS: Whether they preferred the Envirofit to their traditional stove or not, many experienced

some problems with the stove. Users were most bothered by the difficulties in cooking large

quantities on the stove. Specifically, 16/24 said it takes longer to cook (on ICS than traditional

stoves) with large vessels and large amounts of food, and that big pots cannot be used as the flame

does not spread (12/24). Related to this same issue, a few (4/24) specifically addressed problems

cooking rice in large quantities. Some (9/24) mentioned chopping wood as a problem. Lastly, some

(6/24) mentioned problems of ash buildup and wood pieces slipping out (4/24) due to the slant of

the opening. Just 2/24 mentioned problems with stability of the stove when stirring pots.

SUGGESTED IMPROVEMENTS: As with the other stoves, the majority (15/24) said the stove should

be larger to accommodate bigger pots and bigger families. Related to the cooking capacity, many

(17/24) wanted the fuel chamber to be bigger/wider to accommodate more and bigger wood,

something that needs to be addressed but not by making the chamber bigger (which would

negatively affect the stove’s efficiency).

Some suggested the place for entering the wood in the stove should be modified (specifically,

slanting inwards) to prevent the wood pieces from falling out of the chamber (9/24), but fewer than

most other stove users. Ash and tipping were not of particular concern to Envirofit users due to the

stove design, but flame size was perceived to be too small and users suggested the flame should

reach the vessel bottom and spread (7/24).

understanding consumer preference and willingness to pay

Documents