FEATURE ARTICLESable clients. These clients often take larger loans toexpand <strong>the</strong> working capital of <strong>the</strong>ir businesses or tofinance asset acquisition. The MFI will make mostof its profits on <strong>the</strong>se larger loans since <strong>the</strong> cost ofadministering a loan is almost <strong>the</strong> same irrespectiveof its size. These long-term customers should alsobe better credit risks—although this is subject todebate. And crucially, <strong>the</strong>se larger-loan clientsallow <strong>the</strong> MFI to finance its smaller loans to poorerclients. The last thing that an MFI, with its sightsset on financial sustainability, wants to see is <strong>the</strong>graduation of <strong>the</strong>se precious clients. Instead, MFIsshould retain <strong>the</strong>m by seeking to meet <strong>the</strong>ir needsthrough a range of client-responsive financialservices.Conclusions for <strong>the</strong> <strong>Microfinance</strong> IndustryThere is compelling evidence, not just fromBangladesh, to support <strong>the</strong> contention that mostdropouts occur because MFIs do not meet <strong>the</strong>needs of <strong>the</strong>ir market. Dropouts are expensive forMFIs, in terms of money already invested that islost when <strong>the</strong> member leaves, and <strong>the</strong> lost potentialbusiness from that member in <strong>the</strong> future. MFIsseeking to develop permanent sustainable organizationsshould improve <strong>the</strong>ir services to reduce clientdissatisfaction and thus desertion. Such a strategyis likely to prove cost-effective.For MFIs committed to creating permanent institutions,graduating <strong>the</strong> most experienced and affluentclients into <strong>the</strong> formal banking system is not adesirable strategy as it implies <strong>the</strong> loss of <strong>the</strong> mostvaluable and cost-effective clients. Indeed, MFIsshould tailor <strong>the</strong>ir services to ensure that <strong>the</strong>y retain<strong>the</strong>se high value customers.For all <strong>the</strong>se reasons, MFIs should pay (and indeedare paying) close attention to <strong>the</strong> nature and qualityof <strong>the</strong>ir financial services. The trade-off between<strong>the</strong> quality of <strong>the</strong> services and cost of providing<strong>the</strong>m is clear, but finding <strong>the</strong> right balance isdifficult. To date, MFIs in Bangladesh have put toomuch emphasis on trying to implement standardized,inflexible, low-cost, credit-driven systemswhen <strong>the</strong>ir clients are asking (and willing to pay) fora broader range of quality services.The irony of this situation was that <strong>the</strong> genesis ofmicrofinance in Bangladesh was originally driven byan extensive program of careful market and operationsresearch designed to understand <strong>the</strong> needsof <strong>the</strong> clients. Professor Yunus’ work with his studentsat Chittagong University in <strong>the</strong> village ofJobra in 1976 was quintessential market research.It is to <strong>the</strong> fundamentals of market research andproduct development that MFIs must return if <strong>the</strong>yare to retain clients and build sustainableinstitutions.Graham Wright is Programme Director of MicroSave-Africa, Chair of CGAP’s Savings Mobilisation WorkingGroup and a Research Associate at <strong>the</strong> Institute ofDevelopment Policy and Management, University ofManchester, UK. He can be reached atGraham@MicroSave-Africa.com.ReferencesBrand, M. and J. Gerschick. (2000). Maximizing Efficiency in<strong>Microfinance</strong>: The Path to Enhanced Outreach andSustainability. Washington: ACCION International.Hulme, D. (1999). “Client Exits (Drop outs) from East AfricanMicro-Finance Institutions.” Kampala: MicroSave-Africa.Khan, Md. K. A. and A.M.R. Chowdhury (1995). “Why VOMembers Drop Out.” Dhaka: BRAC.Matin, I. (1998). “Informal Credit Transactions of Micro-CreditBorrowers in Rural Bangladesh.” mimeo, Dhaka.Mustafa, S, et al. (1996). “Beacon of Hope: An ImpactAssessment Study of BRAC’s Rural DevelopmentProgramme.” Dhaka: BRAC.PromPT (1996). "Financial Services for <strong>the</strong> Rural Poor -Users' Perspectives." Dhaka: PromPT.Ru<strong>the</strong>rford, S (1995). ASA: The Biography of an NGO,Empowerment and Credit in Rural Bangladesh. Dhaka: ASA.Wright, G.A.N. (2000). “<strong>Microfinance</strong>: The Solution or aProblem?” in MicroFinance Systems: Designing QualityFinancial Services for <strong>the</strong> Poor. Dhaka: University PressLimited and London: Zed Books.16 MICROBANKING BULLETIN, APRIL 2001
FEATURE ARTICLESExodus: Why Customers LeaveKim WilsonA customer is an asset. It is hard to disagree withthat statement. Yet <strong>the</strong> microfinance industry hasnot caught on to a customer’s value. While manymicrofinance institutions pay close attention tominimizing <strong>the</strong> costs of delinquency, <strong>the</strong>y seem todisregard <strong>the</strong> expense of losing good customers.This article proposes that <strong>the</strong> microfinance industrylook at deserting customers with <strong>the</strong> same passionand precision as it examines delinquency. Tworeasons compel us to do so. First, lost customersplace our social agenda in peril. How can <strong>the</strong>empowering benefits of microfinance take place ifour customers flee after one loan cycle? And,worse, we may be losing our poorest customers,those whom we most want to serve. Second, lostcustomers cost money. They cost additional marketingdollars to attract and prepare new customersfor loans; and <strong>the</strong>y cost us in lost profits.An Example of Using Dropouts to InformMarketing DecisionsThe first step in stemming <strong>the</strong> flow of dropouts is tounderstand who is dropping out and why. At Mikra,a new MFI in Bosnia-Herzegovina, <strong>the</strong> ExecutiveDirector, Sanin Campara, and his staff pay specialattention to data on deserting customers. They usedata to draw conclusions and make changes in <strong>the</strong>program. Campara believes that information suppliedby deserting customers is sometimes far morevaluable than data from existing customers.Deserting customers have nothing to fear inanswering questions. They have already made <strong>the</strong>decision to leave, so information is likely to be valid,absent any bias a current customer might have.Current customers may fear losing services altoge<strong>the</strong>rif <strong>the</strong>y are honest in answering questions.Deserting customers can afford to be honest.Moreover, deserting customers offer a marketsavvyMFI more than information about <strong>the</strong>mselves.They may offer clues as to why some customersmay not be interested in <strong>the</strong> MFI in <strong>the</strong> first place,as well as insights into <strong>the</strong> relative attraction of <strong>the</strong>competition (where applicable).To understand deserting customers better,Campara and his team interview every singledropout. To simplify <strong>the</strong> procedure, <strong>the</strong>y check offreasons for dropping out on a one-page sheet thatdisplays <strong>the</strong> most commonly cited answers.They found that in one town, Kakenj, businessfailure caused 30 percent of <strong>the</strong> dropouts. Kakenjhad suffered many economic set backs—<strong>the</strong> closingof a coal mine, <strong>the</strong>rmo-electric plant, and <strong>the</strong>state-owned salmon processing facilities. Mikraimmediately put this information to use by tailoringservices to <strong>the</strong> local environment. In Kakenj, <strong>the</strong>loan officers started to help village banks becomemindful of potential business problems due to afailed economy.In ano<strong>the</strong>r example, 7 percent of <strong>the</strong> dropoutsindicated that <strong>the</strong> loan terms were too short. Whileonly a small percentage, this figure became alarmingwhen staff discovered that <strong>the</strong> clients in thiscategory included some of Mikra’s best customerswho left as <strong>the</strong>y approached <strong>the</strong>ir fourth loan. Mikraresponded by leng<strong>the</strong>ning <strong>the</strong> term in <strong>the</strong> fourthcycle from six to eight months. Was this <strong>the</strong> rightmove? Mikra staff polled two village banks in <strong>the</strong>fourth cycle and learned that 13 percent would havedropped out if Mikra had not leng<strong>the</strong>ned its loanterm. When asked how many would have droppedout if competition had offered longer terms, elevenof twenty-five members in one village bank respondedthat <strong>the</strong>y would have defected.Mikra is now turning its attention to <strong>the</strong> relationshipbetween delinquency and desertion. Are delinquentcustomers dropping out when <strong>the</strong>y would prefer tostay? Does an increase in tardy payers who areforced to leave <strong>the</strong> program signal <strong>the</strong> need for aproduct that better matches <strong>the</strong>ir cash flow? AtMikra, approximately 13 percent of <strong>the</strong> dropoutswere delinquent. Staff is fine-tuning <strong>the</strong>ir dataga<strong>the</strong>ring and database filters to determine who isboth delinquent and dropping out. Do dropout/delinquent customers have larger loans? Sincelarger loan customers are more profitable, changing<strong>the</strong> loan product to meet <strong>the</strong>ir needs may be worth<strong>the</strong> effort (if it does not increase <strong>the</strong> organization’scredit risk). If new customers are deserting, thissuggests <strong>the</strong> need for adjustments to <strong>the</strong> marketingand screening processes.Mikra has one of <strong>the</strong> lowest desertion rates of allCRS-sponsored programs. Over <strong>the</strong> past 30months, with 2,500 customers, Mikra enjoys anannualized desertion rate of less than 11 percent. 13Is this because Mikra’s attention to deserters isexactly what prevents customers from deserting?At this early stage, we can only speculate.13This is calculated as follows: (number of dropouts last month /number of clients this month) x 12.MICROBANKING BULLETIN, APRIL 2001 17