The ground realities of lending to small enterprises

Panel discussion with my herosOctober 25, 2016

As part of the global Financial Inclusion week this year, we were able to organize a great panel discussion with my heros who represent the practioners and thought leaders of the small business lending sector in India.

Artoo Ground Realities of Small Business Lending

  • SG: Samit Ghosh, Ujjivan
  • RN: Ramkarishna Nishtala, Vistaar
  • VS: Vaibhav Singh, Capital Float
  • MM: Murli Manohar, Ujjivan
  • ST: Sneh Thakur, Ujjivan
  • SS: Sameer Segal, Artoo


(SG): spoke about how keen the Government is on MSME Lending as it sees it as congruent with its Standup India program which seeks to facilitate bank loans to entrepreneurs among women, Dalits, and tribals.

He also spoke about how 70 percent of GL is small enterprises and about how as they scale, GL doesn’t work effectively for them- also the same staff can’t do this work and that’s why a separate vertical was needed

He also mentioned that GL is maturing and in 5 years he expects a structural shift from 87 percent GL today to 50 percent IL.

RN: said Vistaar’s focus is on the segment in the middle, between the one that is served by banks and the one that is served by MFIs. He also went on to say that this customer has his own unconventional way of tracking his business, which could be on the basis of documents such as a Sales Book for a Kirana Merchant or a Delivery Note for a small manufacturer, etc. He indicated that Vistaar now has a portfolio worth ₹1000 crores across 80000 customers.

VS: introduced Capital Float, which started 3 years back and believes that 75 percent of the market is unpenetrated at the moment. Their underlying hypothesis is firstly that the cost of customer acquisition through the traditional model is too high- it involves too many people and is essentially a brick and mortar business in which it is very difficult to make money because of the high costs. ( Q: Is Vistaar profitable? How profitable is Ujjivan’s IL business? Would it not be a test of Artoo’s customer intimacy to have access to this data, at least for Ujjivan?)

Secondly, it is difficult to do underwriting in the segment as it is not just a question of doing things like throwing away some of the metrics from the list of metrics used by the banking sector. Besides even the old bank statement doesn’t work because ours is so much of a cash-driven economy.

So they believed that those were the two problems that needed to be solved by using technology and they feel that they are lucky to be at the right place at the right time. The pace of digitisation is changing all our lives, it’s the big mega trend today and with digitisation, and “our entire process is digitised, our customers use it and we find that it’s possible to make small ticket loans reasonably profitably. And with digitisation, small businesses tend to leave lot of data footprint in the economy and we work across several ecosystems and we use that alternative data that is available in a certain ecosystem, use it to build an underwriting model bottom up, rather than use more traditional methods .”

We are very young- portfolio size of ₹350 crores with a team size of 350 people.

Are the needs and constraints of the segment really unique? And how so?

SS: We’ve spoken of lack of documentation, especially lack of traditional documentation. Is that all there is to it?

RN: Firstly the MSME segment is not homogeneous…in fact, we may be 3 different organisations here addressing 3 different sub-segments within the MSME segment.

Our analysis is that about 96 percent of the MSME segment does not have any registrations and does not have access to formal finance. Capital Float is catering to the 4 percent of the segment which has the footprint, which considering that there are some 60 million MSME enterprises, is still a very large number. Our customer on the other hand is a kirana store in a very small town. He deals entirely in cash and so the so-called digital footprint is nonexistent because even his electricity bill is paid in cash. So how do you assess him?

The other thing is that many of these people are taking such a large loan for the first time. Therefore a certain amount of credit literacy has to be imparted to him and sufficient diligence on the end-use has to be done. As much as 30 minutes of client education is done during which we don’t tell him about how to run a better Kirana store but about the the consequences of not repaying, CIBIL, etc.

SG: What Capital Float is doing in terms of customer acquisition is extremely efficient. For us, the traditional way is extremely inefficient, the productivity is very low, customer acquisition is very difficult, we need to do much better marketing and sales for customer acquisition. Our practices at the moment in terms of acquiring and servicing customers are extremely inefficient and primitive - cold calling, etc. A lot needs to be done here to make the process more efficient. ( What do you think he has in mind?Would above-the-line advertising not be a more cost-efficient way of creating the ” pull” for a brand like Ujjivan or Vistaar? Is anybody in the sector using that sort of marketing-mix strategy? How important is the word-of-mouth factor? I am half-tempted to help these guys:))

The other thing is that we need to understand each industry and its structure. For instance, there may be 3 different levels in the footwear industry in Agra and there may be a network of relationships between them. Vistaar does a good job- from what I hear- of doing industry studies.

RN: My view is slight contrary to Samit’s. One thing we have discovered is that anything that comes easily is likely to go wrong and blow up in your face in the long run. E.g we developed a very good sectoral understanding of the dairy industry and that gave us a good sense of the nuances of it such as the fact that, if the cow standing next to the borrower nuzzles him or not might tell you if he is the true owner or not…

The downside was that once we had developed this understanding entire branches became dairy branches because sourcing became easy, there was no need for leaflets or cold calling till we had to pull them up to go back to the rigour and discipline of counting the streets, generating the leads in order to diversify their portfolios.

VS: RN is right, the segment is heterogeneous. That said,interestingly our fastest growing segment is Kirana stores and the way we’ve been able to find some relevant data is that some 200,000 of them work with platforms like Suvidha, DTH, recharge services, remittances is a huge part- we’ve gone ahead and partnered with a lot of them. Again this Kirana profile is different from the one that Vistaar targets but we believe these guys have a lot of data and the business relationships are more established, and we think we can look at the data and give him a short term loan. One of the things we ask is that a borrower has to download a CF App if he’s using a smartphone…now a lot of them don’t have smartphones so we started a scheme for smartphone financing and now that has resulted in us getting a lot of data about their lives and businesses …we’ve been surprised ourselves at how much digital data there is below the surface, you just have to scratch it…

SS: The smartphone penetration ( inaudible)… So in effect is there anything homogeneous at all about the segment, say their aspirations…

VS: We find that most SMEs are one-man-shows if you burden them with a lot of process, then they fall off. We tend to believe in data rather than things like gut feel and our data shows that 70 percent of applications come outside business hours- and this cuts across sectors and therefore the financing has to take place outside business hours- this is a clear and compelling requirement.

RN: There are segments- the first one are the contented Satisficers. The second category are fence-sitters who will possibly come into the financial mainstream and are asking the question, why can’t I also take a loan and grow faster…third is saying why should I grow so slowly. E.g, we know this powerloom operator/ laborer. He started with 2 power looms, which became 28 looms over a 10 year period. He said I want to take it from 28 to 50 and I don’t want to take 10 years over it. My son has passed, there is demand but I can’t go to a bank, I don’t have the documents they want. So this is the segment that wants to leapfrog or telescope, often propelled by the next generation, they want to crunch what took 10 years into a few months! This is the segment that has the hunger and this is the segment we’re going after… ( I found this part of the discussion slightly disappointing as so much of the discussion seemed to be only on small functional motivators- surely the motivations of these customers have to do with the need to be successful. RN was the only panelist who seemed to segment customers on the basis of how they relate with their businesses. But there is also the possibility of segmenting on the basis of how they relate with lenders and lending…)

MM( Murli Manohar): We are moving our focus from ex-GL customers to open-market customers.And what we find is that a lot of them struggle with things like taxation and accounting, which in turn also complicates their case with us. Also there are these sharks in the form of DSAs and Loan Brokers/ Consultants who indulge in sharp, even corrupt practices because of their target pressures, they game the system, take kick-backs, etc.

VS: Yes, we are trying to discipline these guys but they form only a small portion of our business at the moment.

Exchange took place here between SG and VS on how Capital Float is partnering with most of the e-commerce websites and financing the working capital needs of sellers on these sites, how it has become hyper competitive with banks entering the fray and how Capital Float is now making a full and final offer in real time to some of these sellers, giving them a certain kind of customer experience, a feeling of “instant gratification “which has helped it maintain a good share in this sector.

How can we perceptibly, remarkably improve the customer experience based on greater empathy for the customer?

MM: I have observed that the basic needs have remained the same over the years: the need for a simple process, not too many documents, low TAT, not too much hassle… So we go to the customer’s doorstep ( everyone does that now), we keep our requirements simple, not too many documents, relying heavily on personal discussion and reference checks.

The early pioneers in this space like Vistaar have evolved the credit process, simplifying it to the point where we can leverage on it. ( Vistaar is the kind of demanding, thought-leader customer with whom Artoo’s customer intimacy is likely to work best. No pressure but it appears to be a must-have customer for you guys)

Technology is the only thing that can help speed up the process. Like when we partnered with you, we were able to reduce the number of processes, simple things like being able to pull out bureau reports right there and then, running scenarios, giving possible solutions to the customer e.g., if you go for this structure, this will be the loan amount, this will be the tenure, etc. More or less decision gets taken at customer’s doorstep, even if the final approval still comes from Credit.

We are also interested in getting into channels that can be approached through e-commerce because we strongly believe that that’s the direction to take.

RN: I have a contrary view again. I keep going back to the 96 percent. We recognise that in this market the interests of the customer and the supplier may not be completely aligned. The customer would be delighted if within a few minutes of applying for a 10 lakh loan, the money hits his bank account. But what would be a delight for him would be a disaster for us. We have studied the expectations of our customers and we find that 14 days is a very good processing time. Because his options are no loan, or a rejection from a PSB after 5 months, and of course the private sector bank won’t even entertain him. Of course we try to reduce the 14 days but I don’t think we’ll ever get to the point where we’ll be giving him instant gratification.

Sometimes we go back for discussion and find a different reality. So for instance there was this Kirana Merchant who wanted a 10 lakh loan to expand his stock when his monthly sales were about 7 lakhs. His CIBIL score was good, the cash flows but we had a smart credit manager who smelt a rat, going by the scale of his house, his car, etc. So unannounced he went back and chatted with the man’s wife, only to discover that the Kirana store was just a side-show and that his real business was sand smuggling! Now, no internet, no smartphone could have found this. Only an intelligent Credit Manager could have.

SG: There was a time when if you asked the customer to wear a yellow pajama if he wanted a loan, he would wear a yellow pajama. Over time, their behaviour has become similar to everybody else’s. I was with Bajaj Finserve the other day and they were talking about a TAT of 3 minutes! Whether we like it or not, I feel that’s the direction we are headed in. We are now working on a GL TAT that could be as little as 48 hours , much to my discomfort. In GL we have collected a lot of data on customers, their occupations, etc and based on that, we now have the data to do rule-based lending( What’s that?)No need for backend or front end work - we have to live with that reality, whether we like it or not.

MSME is in its early stages so it’s not yet there but over time it is inevitable…Credit decisions will have to be based on data and analytics will be the name of the game. The process has to be made simpler and much more efficient. In MSME, over time you can definitely collect enough data to segment customers and determine behaviour from that, think in terms of clusters..

VS: There is a popular theory that has taken roots in Eastern Europe and parts of Africa which says that this tree is in this part of the forest because it’s more conducive here but over time, when you relate it to lending, you start with a certain segmentation and with a certain set of rules but you recognise that over time people migrate, so effectively the nature of the forest changes over time. The parallel with lending is that borrowers’ profiles change over time so you can build a decision tree on top of that forest which allows you to recognise that the customer has now moved from this vicinity of the forest to another…some of this theory has been put into practice…it may take time…now whether it takes 10 years or 20 or 5 remains to be seen…( I must admit I did not fully understand this and when I googled it, I got a lot of technical material and very little explanation in layman’s language. However I was struck by how a guy like Samit Ghosh was saying that this is the future reality we are all gonna have to live with, much to his discomfort. 2 Questions arise: is this another subprime crisis in the making? And what implications does this have for the Artoo model? Should Artoo be developing a strategic point of view on this?)

MM: We are looking forward to developing other products that will complement and make it a complete product suite -other than term loans, people need working capital loans more than term loans…

VS: We are actually doing a product that mimics a line of credit especially in certain ecosystems like the more traditional supply chains like IT distribution - we assign a limit which he draws down as and when he requires and that becomes a bullet loan

SS: We’ve spoken about instant gratification, can we now turn to product innovation and how that might differentiate between lenders and improve the customer experience…

RN: inaudible - but he seemed to speak about lower interest rates, cutting operating costs and about how about 2 percent of their base migrates regularly to larger NBFCs…and how they have to compete with them straight on interest rates

VS: I agree that it is an extremely price-sensitive market but there is one more aspect is the point of engagement, the current interaction is purely transactional. However the customer is spoilt for choice and a lot of virgin segments are closing quickly…the key question is how do you engage with the customer more deeply not just as a one time lender but as a financial partner or advisor by adding more value to his business… ( Check out ” Building Brands without Mass Media ” by David Aaker for a good discussion on customer engagement)

SS: I think you did a conference recently on this, right?

VS: Yes, the purpose of that event was to provide advisory in terms of how capital, logistics, etc can help the second-tier of e-commerce sellers to become more savvy in the way they run their businesses

What is the role of technology in MSME lending? Will it make the field agent redundant one day?

VS: I think this progression will be a curve- for the smaller ticket sizes in some segments, the day is not far when we’ll have replaced the field agent; But for the larger ticket sizes, I’m not sure. I think that we at Capital Float are for enhancing rather than compromising credit quality and we believe only in automating the 80 percent of the activities that are low-risk, low reward and low value. We believe human beings are smarter than machines and will always have a role to play in doing the high-risk, high-reward, high value work.( This reminded me of Aravind Eyecare ‘s revolutionary process innovation through which they succeeded in transferring 80 percent of the low-end work to inexpensive paramedical staff, releasing the surgeon’s time for high end work and thereby generating step-gains in volume which radically reduced costs without the slightest compromise in quality. Related question: if micro enterprises are the ones that take small ticket loans, then will those be taken over by the fintechs, making the assisted model relevant mainly for the larger ticket, medium enterprises ?)

SG: I think technology will help in making the process more efficient ( interesting how they all are still only talking about efficiency, not effectiveness). I believe the loan officer’s time and input are both extremely valuable, so if technology can reduce or even eliminate the grunt work- the leaflets, the cold calling etc- I mean 9 out of 10 cold calls fail, so if I can use technology to remove the 9 so that the loan officer can spend his time on the 1 rather than the 9, then that’s really helpful. I still believe that in financial services, TAT is extremely important, much more than even interest rates…( he seems to understand his customers and their main anxieties)

SS: ( inaudible for a large part) but I think his question was given that you see technology’s role as removing the 9 out of the 10, what would your mandate to tech providers be?

SG: I think analytics will play an extremely important part..

MM: The other area would be CRM, some sort of management technology, I mean the plain vanilla is available in the market but it may not be useful for everybody…( Is this still Artoo’s main chance therefore? Is the product alternative still Salesforce rather than fintechs based on Big Data because-come to think of it- they are Artoo’s customers’ competitor, rather than your own? Or do we have two product alternatives- not ideal from the point of view of developing a sharp positioning statement )

SG: What Murli is talking about is a scenario we envisage in which the loan officer carrying a tablet can open accounts paperless, they can do loan applications and appraisals, they can do collections, transactions except deposits. That is what we are going to use.. the only thing I am scared of is overloading the loan officer too much…but that is what we are going to use and it will simplify a lot of things .. for instance for account opening the form had some 80 fields, now with e-KYC, all they have to fill up is 10 fields! And then it’ll all have biometric, over 80-90 percent of our customers have Aadhar cards,

MM: Also I feel that there is an opportunity for tech providers to educate the client- you see, there are so many different vendors with different sorts of capabilities and sometimes there may be redundancies and clients need to be educated on all this…especially after implementation, there may be unsolved problems that are being ignored ( I think he seems to be crying for a tech provider who can play a disinterested role as an integrator?)

Sneh: Today we are using a model which is completely a descriptive and a diagnostic model, going forward, with greater understanding of customer segments, we will need to look at platforms which will enable predictive scoring and modelling. It has to be a part of the stack as MSME evolves ( are these some sort of multiple regression techniques she’s referring to which will predict default given a bunch of explanatory, independent variables?)

SG: So if everything works out for us.. we have a lot of work to do…we should be able to open accounts in 10 minutesflat…

RN: I think the field agent in our scheme of things is here to stay for at least the next 10 years…his role is going to be as important or perhaps it might even become more important…technology will make some parts of the process more efficient but it won’t be able to dispense with other parts…e.g., we won’t sanction loans on the basis of 1300 “likes” garnered in 10 minutes . For us the Credit Manager going out into the field and doing credit assessments will remain a core activity…

Where will technology come into all this…let’s look at the loan officer…will the loan officer not have to make cold calls? ..cold calls are necessary because the cold calls are helping us figure out who are the people who have expressed interest …we have some basic filters- firstly the guy must be interested in taking a loan…I don’t want to give a loan to somebody who is not interested in taking one…how do we find out for- remember our 96 percent segment , not the 4 percent- there is little substitute for many years to come other than go and ask him. We also have another filter for all our loans more than 1 lakh which is 98 percent of our business, he has to give property as collateral. Firstly he has to have property and secondly he has to be willing to give it. This can happen only through a the cold calls are necessary to figure out who is interested…subsequently then we also are looking at deploying technology heavily for many of the subsequent processes which means that we will also have credit scoring ( inaudible) which will help us to identify the clear “noes” from the clear “yeses” which will be fast-tracked and those who are in between.

The loan officer himself is not a very organised and systematic guy… if he meets 10 customers on a beat, 2 or 3 of them will say I have to consult my brother or my father because the property is jointly owned so please can you come again next Monday …today we have very little clue whether this guy will go back on Monday or not…this is a need which has been garnered with a lot of effort does he have the mechanism to track this… therefore technology in the form of a CRM will offer a TAP? for the loan officer…that’s the direction in which we are going so that all our 1200 loan officers will be equipped with TAPs… How else will TAPs help? These people also have to collect …there is a bolt-on collection system which can also use TAPs…so at the click of a button he can open up and see that he has 6 tasks for the day- 2 collections, 3 new prospects, 2 where additional documents have to be collected. In addition, 10 cold calls have to be made so an automatic route map gets created for him…so that’s the way in which technology will make his life more efficient.

Now let’s move on to credit management. One of the strengths we have is, we have information on the performance of 125000 customers garnered over the last 6 years. So industry statistics on how powerloom operators are doing is one thing- but that is only of limited value. On the other hand let’s take a credit manager in a small town meeting a power loom operator- what if we were to give him 3 or 4 alerts- what if were to tell him that the master weaver this power loom operator is supplying to has 25 customers and 3 of them are non-performing… Or that the average output for a saree is₹ 4500, which this person is claiming to be ₹5700. ..therefore do you want to take it to be ₹ 4500 or 5700? This is technology in real time, it works in the time in which he needs to take the decision. Thus even for the Credit Manager, real value addition can be made though such prompts, which are based on the data, the intelligence we have garnered. Thus technology will never dispense with the loan officer or the credit manager but it will make them a whole lot more efficient ( even he is still using the word” efficient ” rather than ” effective” to describe the technology proposition. And Vistaar seems to be the kind of prudent lender that HDFC is, the one that eventually ends up with the much stronger balance sheet. Or are they being ultra-conservative and might end up missing the bus?)

Audience Question: What do you think of extremely short term loans made in the same way as a stock trader does day trading?

RN: The problem with extremely short term loans is that they are extremely short term. Therefore they don’t make money. Not only is it financially not a good idea to give a term loan to somebody who has a working capital requirement but inherently we find that in this sector if you give people longer term loans so that they have more time to repay and you set up the discipline of repayment every month, it makes more money for us and works out better for them too. This is because the sector is not disciplined and it’s definitely not organised.

MM: What may work is if we can set up a limit for the borrower against which they can draw… but if it’s a one-off, the cost of setting it up may not be justified.

VS: This is out there, it’s just that it’s not business-focused, it’s consumer-focused. There is this whole line of lending called Pay Day lending which thankfully the regulator in India has frowned upon from day one. The problem with very short term loans is that the IRRs at which they get made are a few hundred percent which is detrimental to the lifecycle of that customer. Anyway it works in the consumer sector because hopefully a businessman will plan for his funding needs a bit better.

SG: In the market we have these day lenders, the moneylenders come in the morning to the mandis and their rates are extortionate. These are pushcart vegetable and fruit vendors and we lend a lot to them and they really don’t need to borrow on a daily basis and once we become a bank, we can offer them an overdraft facility which is much more economical for them than taking a term loan.

As a bank we can offer the whole range of financial services, one of the most important being savings. The second most important is remittance. In our mind, true financial inclusion comes when the customer has a full range of financial services. So that’s the benefit to the customer and that’s what drives us. And frankly, a safe place to save is more important than a loan because the poor lose out so much of their savings to these fly-by-night operators…but it’s gonna be difficult because this segment is quite used to saving in the unorganised sector and getting them to change and move to a organised sector will be difficult… that’s why we’ve studied very closely what the unorganised sector offers them…people say bad things about Sahara but we went out to understand what explains Sahara’s tremendous customer loyalty and we found that they do a lot of good things which the organised sector does not provide them and we are trying to incorporate a lot of those things into our offerings.( What could these be?)

VS: E-commerce sellers are about 30 percent of our business and as the ticket size gets larger in that segment, we add more and more documents like tax returns and balance sheets and so on and don’t just lend on the basis of the e-commerce data.

RN: ( answering a question on how their underwriting processes have evolved over the years to the benefit of their customers) : Because we now have a lot of data and information we have developed Sector Guides and these enable us to reduce the number of questions we ask our customers to a standardised 7-8 questions which give him great insight into what is actually going on in the customer’s business. So the personal discussion time has got reduced and we use a lot of supplementary information from the people the customer trades with e.g., there is a network of master weavers we depend on for rich information on power loom operators.

It is about systematic capture of information and insights into that particular sector which helps us to make the credit decision more efficient than earlier. Our TAT is down from 25 days to 9 days and we give a guarantee which entitles the customer to a reduction in the processing fee in case the TAT is higher than promised.

SS: ( Summary) Huge opportunity, the criticality to the economy, the time spent on customer education is well spent, role of technology is to eliminate low risk, low reward activities so that the field agent moves up the value chain. Exciting times, government doing its bit and we are all at the epicentre of it.

Sameer is excited about working at the intersection of technology and impact. An idea that sparked from an internship, resulted in a company, Artoo, that enabled US $1 billion loans to small businesses and impacted 5 million lives. He is fascinated by the elegance of a well designed system and the satisfaction that a maker gets by seeing the awe in the user.

Sameer loves experimenting with software and technology in general. He is an aspiring triathlete and loves taking his daughter on Dadventures.