Big data is such a huge innovation

Safaricom had 24,206,341 active subscriptions on its M-Pesa platform at the end of September. FILE PHOTO | NMG

Earlier this month, this newspaper led with a headline that Safaricom’s Fuliza product lent Sh6.2 billion in its first month after launch. In case you’re one of those laggards that hasn’t entered the M-Pesa universe yet, Fuliza was launched by Safaricom in January 2019. Its objective is to help the M-Pesa user avoid that embarrassing “oh-no” moment when goods or services that she wishes to purchase are literally in hand but the funds to pay are not.

I signed up for the product following an SMS blitz by Safaricom as soon as it was launched for no other reason than to just stop the confounded messages coming through. Two weeks later I stood at the supermarket till purchasing items via M-Pesa. Lord help me because I came up short, Sh434.74 to be precise.

Usually I would give a sheepish grin to both the cashier and to the visibly irritated customers behind me and mumble something about “please let me withdraw from my bank” and have to wait several nail biting, interminable minutes as my bank’s mobile app chooses to be slow on that day at that moment.

But the M-Pesa app immediately prompted me to Fuliza – which, by the way, means “continue” in Swahili. In seconds I had been allowed to overdraw my M-Pesa by that amount, the transaction was completed, I got an update that I was charged the princely amount of Sh4.35 for the overdraft facility and I now owed Sh439.09 due in 30 days.

Most importantly, the fellows standing in line behind me never knew that I had run out of funds. At all. The next day I withdrew funds from my bank into M-Pesa. Again I got a message in a split second, the outstanding amount had been automatically deducted from my funds. And my available limit was back to the Sh12,000 that I had automatically been awarded when I signed up.

Fuliza is a testimony to those two words you see being bandied about miscellaneously: “big data”. Big data are extremely large data sets that may be analysed to reveal patterns, trends and associations relating to human behaviour and interactions. CBA Bank, the creators of the first M-Pesa-based lending product Mshwari, used M-Pesa usage data to feed into their credit algorithm that calculated how credit-worthy the loan applicants were. It soon became apparent that about 58 per cent of M-Pesa transactions failed where the user was sending money to another beneficiary. But about 85 per cent of the same transactions would be repeated within two days, that is, payment to the same beneficiary because funds were now available.

It doesn’t take a rocket scientist to see that the data was speaking to a funding gap that would be eliminated within 48 hours as cash came in. In banking-speak this is what an overdraft does: provide a short term cash bridge pending arrival of funds. In the example above, my overdraft interest rate was one per cent for a 30-day facility.

If you were to ask the over 400,000 customers that are using Fuliza daily as to what the annualised interest rate (12 per cent!) is, they’d tell you that they didn’t care. I certainly didn’t at the point where I was standing at the till with a basket of goods already packed and carefully perched on the counter ready for my hasty exit. Actually neither do the millions of Mshwari customers who are ready to pay a flat fee of 7.5 per cent for a 30-day loan (again, if you annualised that you would get 90 per cent).

The Fuliza product currently endures a default rate of less than 1 per cent. So for every Sh100 that are lent out, less than Sh1 is lost. The automatic limit that I received of Sh12,000 was done without my asking and without my knowledge. The bank just used my data to generate a very important product for me.

To the Kenyan legislature, the lesson here is this: a little bit of research would have led you to see how you could force banks to use the reams of data that they have about their customers to provide a better and differentiated pricing which would have achieved the goal of lowering the cost of loans. Instead, the largely uninformed interest cap route taken has ended up drying credit supply.

What these mobile loan applications are telling parliamentarians is that at the end of the day, the retail client is indifferent to the price. He just wants to “fuliza” his life!

PAYE Tax Calculator

Note: The results are not exact but very close to the actual.