December 15

Jake Kendall

Does ’the Cloud’ Have a Silver Lining for the Poor? New Sources of E-Transaction Data Hold Promise

Can new sources of electronic transactional data improve financial services for the poor?

Financial services companies in wealthy nations depend heavily on data to design their products, segment the market, and tailor individual products to individual consumers. A very basic example is using credit scores, derived from past financial history, to set interest rates and target credit offerings. More advanced approaches use customer-level analytics that identify customer life events (births, moves, entering school, receiving a bonus, etc.) to sell financial services offerings.

To date, this data-driven approach has been largely impossible for any banks that would pursue it with populations at the bottom of the economic pyramid. Most of the poor use informal cash-based financial services and have no formal relationships (not even regular utility bill payments) and thus have no financial data to mine.

In a previous post, I alluded to the idea that mobile money systems are platforms over which a lot of financial and non-financial services can flow. A major benefit is that these systems take cash handling costs out of the system. Another potential benefit (still somewhat theoretical at this point) is that clients will create a financial history, and a useful (and valuable) trove of data for financial service providers.

Though there are few current working examples of data from the cloud being used to drive business models that reach the poor, I did meet with a number of players in the financial space who believe the potential is large – and several are gearing up to launch. Here are some examples:

Pesapal, a billing aggregator, plans to roll out a credit scoring product based on the regularity and size of clients’ utility bill payments. They still haven’t linked credit default data to their billpay data to create a scoring system but believe they can develop one in short order as their customer base grows. Customers would opt in to allow their score to be shared with a lender (who would purchase the data), which could lower lenders’ risk in credit decisions.

First Access is a start-up operating in Tanzaniathat willdo something similar, but using data on prepaid airtime purchases as a primary source of information.The basic idea ofusing pre-pay airtime purchase data as a proxy for income stability has been discussed by mobile network operators for many years. But First Access has managed to develop a business modelthat they believecreates value for all parties involved.They have looked into the idea of doing credit scoring based on mobile money transactions, butcurrently believe a model based on payments between individuals would be too easy to game by savvy clientssending money back and forth. Down the road, as formal income streams like salary or pension payments start to get paid outover mobile money, this could become feasible.

Kopo-Kopo has created an SME merchant transactional data interface. With parallels to but geared towards merchants, they pull a merchant’s transactional record from Safaricom (M-PESA’s parent company). They then use this data to create a platform for merchants to access and manage their payments. They believe this will be a powerful source of data on which to base merchant credit scoring down the line (merchant credit could be a large and lucrative market).

Our team spoke to a number of banks and other players who are also eyeing the space but either had no specific plans or asked us to keep their ideas confidential. Clearly, market players believe the potential is there.

However, there is reason to be cautious in predicting a transformational impact from new data being captured on poor clients’ financial behavior. Most of these approaches are relatively straightforward attempts to measure income or ability to repay credit through observing clients existing payment behaviors, and have fundamental limits in the extent to which they can predict credit worthiness. More advanced targeting of credit will have to be based on well-run credit bureaus, which have their own challenges.

Additionally, none of the players we talked to were actually implementing their strategy and it remains to be seen that their operational issues can be worked out. Finally, some markets will feature regulatory hurdles that will limit the ability to share and sell data. Perhaps the most telling indicator, regarding the perceived value of data in the market, is the fact that the majority of banks working in developing markets don’t effectively use the data they already have on their clients (even some of the ones who profess to be excited to mine mobile money data!)

That said, it’s hard to imagine that these issues won’t be overcome in the long run, which could result in more efficient and more tailored financial services offering for the poor.

Base of the Pyramid, financial inclusion, mobile banking, technology