How To Develop Loan Product ?

This month marks my 2 years tenure in Amartha. After almost decade of experience in telco industry and straight outta Manchester Business School Innovation studies, in late September 2017 i jumped headfirst into fintech startup world. It was challenging and exciting ride with constant learning every single day of the week.

One of my key learning from leading Amartha’s Business Development team is how to design loan product. Amartha core business is micro-loan for rural women and as bizdev, my task is to explore new loan product opportunities. In this post, i will brain dump my learning in developing loan product for micro-business segment.

Hope its useful!.

Solving Pain Points

Like any other products, Loan Products should be tailored to solve problems for specific segment of users. Good loan products is align with user’s business and financial cycle.

Lets take a look at three different micro-business profile below.

First one is warung kelontong or small shop. These business typically have high frequency of small amount transaction in daily basis. Mostly sell packaged foods, cigarettes and sembako (groceries). Since the transactions are happening frequently, warung kelontong needs to constantly restock whenever goods are run out. However, typically warung income is the sole income of household. Hence when theres sudden big financial needs such as Lebaran or School entry day, warung owners could run out of money to replenish their stock. With no stock come no sales hence the need for loan.

Second one is a small fisheries (tambak). Typically fish-farmers need capital upfront to buy seedlings, food and vitamin. Harvesting cycle for small fish is usually 3 months. Once fully grown, fish then sold to agents (tengkulak, bakul). Biggest cost for fish farmers are fish foods. In this 3 months, they need to constantly spend without having income. If suddenly they run out of money, fish then are sold half grown which is a loss for the farmers. For this users, loans are to buy capital to start production or to sustain their current fish batch.

Third one is a home-based handcraft business selling hand-made doormat (keset). Owner usually are housewives in rural villages wanting to get income during their spare time. These ibu-ibu make the doormat using spare cloths which can be easily found near textile factories, sold for near to nothing. Doormats then sold to agents, picking up the mats in their house on a weekly basis.

Now we can see that each of these business have different characteristics, financial cycle and loan needs. Common SME loan products from bank offer very high ticket size (more than 10 million), require not only bank accounts but also 3 months transaction history, and most of the them are on monthly repayment basis. Serving all three with generic loan product will result in low take-up rate (user not interested) or high non-performing loan (cannot pay installment ).

Loan Features (T.R.I.P)

Loan products usually have four core features, tenor, rate, installment and plafond. Below are simple explanation on each :

  • Tenor : How long does the loan last ? in weeks, months, or years ?. The longer the loan, the higher the risk. But the longer the loan, repayment become cheaper which makes it more attractive to loan users.
  • Rate : How much is the interest ? Defining interest rate is a balance between market attractiveness, profitability, risk. Unbankable segment is by default riskier segment due to lack of financial records, hence higher interest. How loan operator deliver the product determines cost to serve. Serving low-tech segment which needs manpower and paper could make the interest high. However if the interest is too high, user wont take it.
  • Installment : Frequency of user to pay the loan. It can be daily, weekly, monthly, quarterly or even bullet payment, which all the loans and interest are paid in bulk at the end of tenor. Loan user generally prefer more frequent installment mode because amount to pay on each installment become lower. However the more frequent installment modes could bear processing cost for the operator. Installment also need to be aligned with borrowers income cycle. So lender could charge installment as soon as the borrower get their money which lower the risk of default.
  • Plafond : How big is the loan ? the bigger the loan bigger the risk.

Now lets try to implement these TRIP to our business case.

In my experience, small warung kelontong typically have 500k to 1.5 mio revenue per day. Quite high turnover but in the next 2-3 days they need to restock with similar amount. Product suitable for needs like this is short-term loan with plafond enough for once or twice restock. Bullet payment can be used due to very short tenor.

Fisheries loan can be quite risky. Capital raised can be high upfront and fish-farmers might not have income for the next 3 months. Two options for type like this, bullet payment at the harvest period or monthly interest-only payment with principal paid at the end.

Home-base handcraft business usually can be started with very small capital. When Amartha first started, loan size was only 500k rupiah. Material cost is very low with majority of the cost is actually the time and manual effort in making this. This product have very low margin hence loan product needs to have long tenor with weekly repayment, to make each installment as cheap as possible.


Once the product are designed, lender need to determine who to give the product which give the best chance of profit (loan repaid with interest). Process of assessing, scoring and giving the credit to borrower is called underwriting.

Underwriting process, in a very simple way, is a process to measure probability of loan repayment from 3 main criteria :

  • Credible applicant : Making sure that the applicant is real people, with real address and real business. Simplest way is to ask borrowers to fill forms. Lenders then need to verify given information via third party validation. Most commons is using Dukcapil to verify borrower KTP information (name, date of birth, address).
  • Ability to pay : Second is to measure their financial capability. Three data points needed to do this : Income, Asset and then Expense of the applicant. This process can both very simple and complex depending on the capability of loan operator / lender. General rule of thumb of good ability to pay is their expense for paying installment is around 30% of their disposable income.
  • Willingness to pay : Financially able to pay doesnt necessarily guarantee that the applicant is willing to pay. To predict willingness, we measure their past behavior of taking loans. If they have history of not repaying loans, applicant is considered high risk. In Indonesia, conventional credit data are centralized in OJK system called SLIK. If they are connected to SLIK, lender could check whether applicant has good or bad history in taking loans in other places.

Underwriting process relies on availability and credibility of information related to loan applicant / borrower. For people who dont have bank account or small business without any official financial records, this could pose significant challenge. Hence fintech targeting this segment increasing relies on alternative data instead of financial data to do underwriting.

I discuss lack of data as one of the challenge in serving financial services for unbanked in my previous essay, read it here.

Disbursement & Repayment

Once underwriting is approved, then borrower will get their loans. This process is called disbursement. Design choices of how borrower could get their loan money could also have impact on risk. Common method to disburse loans are to transfer it to bank account or giving it directly in cash. However if its productive loan, especially for micro-business, theres a risk of side-streaming : loan are not utilized for production but for consumption. Misuse of the fund make borrower’s business not growing, while the next week they need to pay the installments. That borrower are doomed to fail.

Lower risk options could be disburse it directly as goods. For example, Cupang fish farmers buy all their production needs in a store and then the store convert it to loan. With farmers receiving all goods and no cash, possibility of side-streaming is low hence repayment probability is high.

Repayment is also another aspect that needs to be design carefully around borrower’s profile and behavior. Repayment is the most frequent process which borrower have to go through during loan process. Hence any friction or inconvenience in repayment process which came from misaligned design could increase probability of non-repayment.

For example, giving options to pay the loans in ATM even though borrower live where there are no nearby ATM in 30 minutes drive radius. Another is to force unbanked borrower to deposit their repayment in their designated savings account, which bank then auto-charge it every first of the month. If the borrower’s business is cash heavy (such as warung kelontong), cashing in their income to account could become burden for borrowers as well.

Reasons to Pay

Last but not least, disincentives need to be designed for borrower who have bad intentions (not paying the loan). This disincentives needs to be strict enough to deter bad behavior, but not harsh enough to scare away potential borrowers.

Common practice for loans is to deploy field collector team to deal with bad borrower. However for micro-business loans, where loan size is very small, deploying collector can cost more than the profit margin itself.

This is why for micro-business loan, it is crucial to design loan product who fit the need and have maximum impact to their business. Because if the borrower business are successful, its easier for them to pay. While if borrower’s business default, it cost more to recover.

Better alternative is to align disincentives with their business continuity. For example, we disburse loan for fish farmers through their seedlings provider. Hence if they dont pay the loan, they cannot buy seedlings again unless their debt are settled. With this schema, borrower have more incentives to pay the loan instead of not.

Further Readings

If you’re interested to know more about developing loan product, heres couple of reference materials i recommend :

This article are driven by the countless of discussion with my inspiring current and former colleague in Amartha. Kudos to Szagibella Aprilia, Giovanny Agnes, Budhi Siswoadji, Adityo Putranto and Aria Widyanto for giving me base idea for this essay.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.