Unlocking distribution for a green energy startup in East Africa

Context

A Series-A distributed solar company in East Africa with a strong technical product, solid capital, and a growing installed base had hit a ceiling. Monthly new activations had been flat for four consecutive quarters. The product worked — customers loved it, default rates were in line with category benchmarks, field ops were competent. And yet.

The founders — sharp, technical, immigrant-returnee profile — believed they had a distribution problem. Their internal diagnosis was that the sales team needed to grow from 40 to 120. They had raised partly on that plan. The bet was that activation was linear in field sales headcount. Hire more feet on the ground, plant more systems.

We took the engagement because we suspected they were wrong.

Moved Fast

Week one, we rode along. Every senior B3n engagement with an operational company starts by shadowing the lowest-level staff doing the actual work. Two days with field sales agents in three counties. Two days with installation teams. One day in the call centre. By day five we had seen three things that were invisible from headquarters.

First, field agents were spending 60 percent of their time on the KYC and credit-check phase — tasks that could be done in the customer's home via a phone agent but were being done in the field because the app required an in-person interaction. Moving KYC to a remote-plus-field handoff doubled effective field time overnight.

Second, the agent compensation plan was optimised for installations, not activations. An installation that later churned paid the agent the same as one that stuck. Agents were — rationally — signing marginal customers and moving on. We redesigned the comp plan to delay 40 percent of the commission until the customer hit ninety days of consistent payment. Alignment reset in week two.

Third, and most important: the company had never meaningfully pursued a partnership channel. Two obvious partners — an agricultural cooperative with 120,000 members and a microfinance institution already doing last-mile credit — had been in "conversations" for eighteen months with no execution plan. We brought them both to signed pilots inside four weeks.

Failed Forward

The initial diagnostic deck we presented in week two had an error in the unit economics model. We had pulled the default rate from the company's internal reporting rather than back-testing it against their actual cohort data — which was materially different. The model showed a break-even CAC that was 30 percent too high.

The founders caught it. We corrected the model within 48 hours, re-ran every recommendation against the fixed numbers, and — critically — half of the recommendations got stronger in the corrected model while two got weaker and were dropped entirely.

The near-miss was a reminder that in operational work, the most dangerous number is the one everyone agrees with. We now bake a "rebuild the base data" step into every engagement, no matter how trustworthy the client's internal numbers look.

Built

By end of week six:

• KYC workflow redesigned and shipped. Field agent productive time rose from 2.5 hours per day to 5.2.
• Agent compensation plan redesigned and launched with the existing team. Three months of retention-linked compensation created natural selection — the agents who were selling marginal customers churned out. The ones who stayed were 40 percent more productive on ninety-day-retained customers.
• Two partnership channels signed, one co-branded with the agricultural cooperative and one embedded as a credit-check-plus-upsell flow with the MFI. Both were projected to contribute $2.4M ARR within twelve months. Both beat projection.
• A revised board narrative replacing "we need to triple the sales team" with "we need to double the operations team and halve the partnerships-to-signature timeline."

The founders cancelled the planned 80-agent hire. Redirected the capital to field operations and a partnerships function. Activations 3x'd inside two quarters.

Compounds

The work compounds because the partnerships layer we installed is structurally advantaged. Every additional agricultural cooperative, MFI, and faith-based network the company signs into the partner channel now plugs into the same playbook — compensation model, KYC handoff, onboarding runbook. The marginal partner adds without linear operational cost.

The deeper compound is harder to measure but more important. The company's leadership now asks different questions in strategy meetings. "Do we need more sales?" became "Does this need more sales, or is this a process problem masquerading as a sales problem?" That diagnostic discipline is the compounding asset. The partnership channel is just the first thing it produced.