By Adi Sudewa
As an impact investor, each year in the past five I have to evaluate hundreds of pitches from well-intentioned impact enterprises with various background and approaches. Like any other private equity investor, I had to evaluate each proposal based on the usual criteria: Is the market big enough? Is this a scalable business? Are the entrepreneurs and their senior team capable in execution? What are the risks? How would we get our exit? These are hard questions, but at least I have solid footings with my MBA education, management consulting and banking experiences, and a vast literature on these topics.
On top answering the above questions, I have to look after the mandate from our LPs (Limited Partners, investors in the fund) and ensure that the impact enterprise under evaluation has a business model that engage with the low-income communities in a fair and transparent manner. The company under evaluation usually either source input material from the poor (e.g., smallholder farmers), or sell affordable product and services to them. How the entrepreneurs engage or plan to engage the poor is a good indicator of how likely their business will be successful.
There is also limited background work on understanding to conduct business with the poor. Researches prior to 2000s were mainly focusing on the effectiveness of aids. I could only do so many field visits; more often than not, their area of operations are quite remote. Over the years, I have a developed a “mental model” of which kind of businesses are more likely to be successful, and which ones that are not. This mental model, or sometimes known in another term “heuristics”, is developed from my own experiences, my colleagues’ experiences in South Asia and Africa, and case studies from books about social enterprises and development practices. One of the best books in the field is “Poor Economics”, written by Nobel prize winning academics Esther Duflo and Abhijit Banerjee. I read the book in 2014 when I was new to impact investment, and I re-read it recently after the 2019 Nobel Prize recipients were announced.
Duflo and Banerjee won the awards because of the “RCT” (randomized controlled trial) experiment model that they pioneered in the field of poverty reduction. However, I found that the findings from their RCT experiments are more applicable in my field than the methodology itself. In particular, findings like how the poor are not very rational in making decisions, their desires for small luxuries, their aversion to risks, their inability to save money, and how almost all of them wants stable livelihood, all resonate with my experiences so far in the field. To me, the book was not only about economics, but more so on how human psychology is amplified in poverty situations. This book forms the basis of my mental model of poverty.
Drawing from my experiences and learnings from others in the field, these are the criteria that a business model should fulfil for it to be attractive to the poor:
- Affordable — entrepreneurs usually understand the importance of affordability, although there is a lot of tendency to put too much emphasis on this factor while neglecting others
- Offering immediate and steady benefits
- Responding to clear demand from the community
Hence the following types of business models unfortunately do not work well with the poor:
- Enterprises selling products requiring a lot of customer education and/or cost-benefit analysis. Examples: savings and insurance products, no matter how simplified, is a hard sell
- Enterprises selling products that will give benefits only in the long term. Examples: primary or general secondary education, preventive healthcare, sanitation, business management training
- Enterprises selling risky products. Examples: sales of most new agriculture equipment like drip irrigation
I am not saying the enterprises doing the above activities can’t be successful. I would love to be proven wrong by these enterprises. From my experiences, their solutions can only work with really extraordinary teams, extraordinary circumstances, and/or the right amount of good luck.
This leaves us to the five business models that I think are more feasible and would merit further evaluation.
1. “Liquidity” Model
In this model, the enterprise provides liquidity to the poor. Most common business model here is the microfinance model, but there are other less-known liquidity providers for the poor: Education loans, housing loans, and remittance. I would also consider Ergos, a provider of grid of micro warehouses as a liquidity model, because they enable farmers to get financing based on the receipts issued by the warehouse.
This model works well with the poor because of a simple reason: Almost everyone need credit in some phase in their life.
2. “Necessities” Model
You might have read the famous book by the late CK Prahalad, “The Fortune at the Bottom of the Pyramid”. The book was published in 2004, well ahead of the surge of microfinance and the rise of impact investing. It advises that corporations can profit from doing business with the BoP (bottom of pyramid, a common term used to address the poor). There are several case studies discussed in the book, but the most often cited one is about Unilever India penetrating the BoP market by offering products with single-serve packaging and super-affordable pricing.
I include this model here because it is a valid impact business model and many poor people around the world still need access to these necessity products. With very low prices (hence low risk) and little education on how to consume the products, this model has a very good chance of being successful in the market.
3. “Aspirational” Model
While most impact enterprises are trying to supply what the poor need, e.g., clean water, education, jobs, I want to argue that it is equally impactful if impact enterprises can provide what the poor want, e.g., “small luxury” products.
Mapan, an Indonesian social enterprise who recently was sold to Go-Jek, understands this very well. They have a business model to facilitate housewives to pool money through “arisan” mechanism to buy affordable products. But products in their catalogs are a mix between productive (like sewing machines or cooking stoves) and non-productive products like hair dryers, fashion items, and BMX bicycles. They curate the products based on the combination of what their customers need, what they want, product affordability, and how well the manufacturers support the products in case of breakdown as the manufacturers often don’t have service centers in rural areas.
Another example, also in Indonesia, is Sorabel, an e-commerce player who provides affordable fashion items to low-income segment, probably the only tech player specifically doing so in the country. They are using AI to predict the fashion trend of the segment to minimize unsold inventory, and mitigate buyer’s risk by allowing for cash upon delivery payment option, which is not universally provided by other e-commerce companies here.
4. “Livelihood” Model
This is a dominant model in the portfolio of many impact investors, covering all enterprises who source from the poor. It can be agri/fishery products, craft products like furniture, or even waste from the wastepickers. It is the best business model for the poor as it enables them to achieve a decent income level and, in many cases, through increasing volume and quality, should offer a path out of poverty.
This model does not automatically imply that it is a low-risk situation for the poor. They still need to invest to buy inputs to production, often in higher spec/prices than the ones they usually have. Selling to a new party can also complicate their relationships with existing offtakers, who are more embedded in the area. Therefore the enterprises should outline how they plan to address it to attract the suppliers.
5. “Productivity” Model
This model covers vocational training, provision of agri-input product and agri-equipment, as well as loans for business expansion. It is easy to mistake fake productivity with real productivity potential. For example, training for baking or car repairs is useful only if there is a skill gap in the community. Otherwise it would just produce a glut of skilled workers who compete with one another. The best combination is to pair productivity model with livelihood model, where an offtaker has certain quality standards to be met and the communities must upskill themselves in order to fulfill their demand.
My mental model does not aim to determine which engagement model gives the broadest or deepest impact to the communities. That would be a different topic to be discussed in a different article. It is worth pointing out here that if the poor does not want to engage with you, there will be zero impact, regardless of how well the intention of the entrepreneurs.
This mental model is super simple, far from scientific, and also clouded with personal biases. It is also not future-proof. New digital technologies like smartphones, AI, blockchain, big data analytics, drones, sensors/IoT, will open new possibilities in engaging the poor. However, I found this model useful at least in the near future, because in its heart it incorporates the elements of human psychology, and change in human psychology would only come very very slowly.