There was no honest playbook for building in Africa. So we wrote one.
There was no honest playbook for building in Africa. So we wrote one.
The Method
A loop for building a company where the data is thin, the benchmarks do not apply, and being wrong is expensive. Each cycle names one constraint, tests what would move it, and keeps only what the evidence supports.
Each turn is one cycle. Each begins from the evidence the last one produced, so each runs tighter than the one before.
Principles
One constraint at a time.
At any moment, one constraint limits the company more than all the others. Work elsewhere moves little until it moves.
Beliefs are hypotheses.
Going to market and raising capital are propositions to be tested, not beliefs to be argued. Form a view. Try it. Keep what the evidence supports.
Every stage ends in an output.
A stage is finished when its output exists, not when a calendar says so. Every output is written down and can be checked.
The loop is handed over.
The work continues until the company can run the loop without us. That is the exit condition.
Stage 1
What is actually holding this company back?
The business is mapped in full: market, numbers, team, constraints. Most companies find their binding constraint is not the one they were working against. The diagnosis is done with the founder, not to them.
Procedure
Output
Gate
The founder accepts the diagnosis. If not, the engagement ends here.
Stage 2
What would have to be true for this to work?
Most companies stall here rather than at the product. Something works, and there is no repeatable way of putting it in front of the people it is for. Build turns hypotheses into experiments and runs them in the real market.
Procedure
We believe [this change] will move [this metric] from [baseline] to [expected result] at a cost of [time and money].
Output
Gate
Continue, change the plan, or stop.
Stage 3
What did the evidence actually say?
Each experiment is measured against the metric it set out to move. The gap between result and expectation is not a failure. It is the material for the next round.
Procedure
Determinations
It proved itself. Written into something the team can run alone.
It half worked. Narrowed and run again.
It failed. Written down so it is not repeated.
Output
Gate
Continue, change the plan, or stop.
A constraint stops constraining a company once it is resolved, and another takes its place. The next cycle starts from measured evidence rather than assumption, which is why each runs faster than the last.
Exit condition
The company runs the loop without us. We aim at this from the first week. Not every company gets there.
