146 customers, 7.3 % repaid: when the best seller is the worst for cash
In May 2023, at Bboxx, I designed the reward programme for the commissioned sales agents of nine African markets. Management wanted a target of 30 sales per agent per month. Out of 1,245 agents measured, 35 reached it, 2.8 %.
Three months later, a new column appeared next to the sales figures: repayment at three months. One agent showed 146 new customers and 7.3 % repayment. A star on volume. On cash, the most expensive kind of sale.
This page is about that column. In a business that sells on credit, an incentive that pays on the sale alone buys defaults.
The group sold solar kits, televisions and phones on instalments, through agents paid on commission. Most unit-economics models expected 80 % repayment over the life of a contract. In October 2023, the portfolio review showed that nearly half of customers were already below 80 % after three months. And it showed something more useful: 60 % of the good sellers, those with more than ten sales, were above 80 %. The problem was not the best agents. It was the incentive that rewarded the others for volume.
An agent is paid on the day of the sale. The company is paid over three years. Any reward built only on the first date pushes the risk onto the second, and nobody in the field sees it.
1 · Size the reward on local pay, not on the marketing budget.
Average monthly pay for an agent went from 100 dollars in the DRC to 450 dollars in Kenya, 215 dollars on average. I set the reward against that, country by country. The interviews with agents gave the mechanic: the only metric that mattered to them was the cash they had at the end of the month. So the reward was cash, not an object.
2 · Reward progress, not level.
The grid paid on the gain over the previous month: 15 dollars for ten extra sales, up to 200 dollars for sixty. The cost per extra sale fell as the steps went up. Every agent could enter it, including the many who sold little.
3 · Say no to the target, with the number.
I declared the 30-sales target impossible, with the distribution in hand: 35 agents out of 1,245. The rule applied became 20 sales and more for an agent, 30 and more for a shop supervisor. Both conditional on more than 80 % repayment at three months.
4 · Put repayment next to volume.
From August 2023, the monthly winners file carried the three-month repayment rate beside the number of customers, and a second list sat beside the winners: the agents with high volume and low repayment, named, to coach. The same document rewarded and warned.
5 · Read the shop, not the country.
The gaps inside a country were wider than the gaps between countries. In Kenya, 86.2 % repayment in Kisumu against 53.9 % in Masara. In the DRC, 90.4 % in Lubumbashi against 32.7 % in Walungu. The lever moved from the national level to the shop.
Nine markets, May 2023 to March 2024.
- Management target of 30 sales per agent replaced by a reachable rule, after showing that only 35 agents out of 1,245 hit it
- Three-month repayment placed beside volume in the monthly recognition, from August 2023
- First quarter of 2024: 4 winners out of 27 possible places; the best performer, in Ghana, signed 173 new customers at 92.4 % repayment, and nobody else held both conditions
- Programme run on nine markets; where it stopped, the cause was stock outs, cash or a clash of commission schemes, never the mechanic
- Redesigned in February 2024 by a growth associate from my team: the programme outlived its designer
Four winners out of 27 says the criterion bites. I think it may also sit too high, and I would reopen that question today. The lesson holds either way: in a credit business, a sales incentive is a finance instrument. If it pays on the signature alone, it pays for the defaults that follow.
Does your sales incentive pay for the sale, or for the cash?
The first call30 minutes. We talk about your traction and what is blocking it. You leave with 2 or 3 moves you can act on. No pitch.
The same system on three markets, three opposite results
Constraint: same system everywhere, local conditions nothing alike.
Read the case studycountry budgets, one group budget, one rule
Constraint: ten markets, ten budget formats, no line authority.
Read the case studyconversion, in a five-month Ghana pilot
Constraint: off-grid customers, low bandwidth, five months to prove it.
Read the case studyBefore, there were two segments: those who spend and everyone else
Constraint: a till database that was not a customer database.
Read the case study