A financial model is a good servant and a poor map. It holds capex, opex, revenue, and a schedule, and it treats the world as a set of line items that resolve on time. The factors that most often decide whether an East African project lands on schedule are not line items. They sit between the cells, and they are usually filed — if they appear at all — under “risks,” as though they were exceptions rather than the terrain.
What sits between the cells
Permitting and approvals. Consents run at both national and county level, and the two do not always move in step. Environmental assessment, sector licences, and county approvals each have their own process and their own duration, and a plan that assumes they run in parallel usually discovers they run in series.
Land, wayleaves, and access. For anything linear — fibre routes, transmission lines — the real critical path is often the right-of-way, not the construction. Continuous access across many parcels and jurisdictions is a sequence of negotiations, and the corridor moves at the speed of its slowest hold-out.
Interconnection. Connecting to the grid is a project inside the project, with its own studies, works, and queue. It is routinely the difference between mechanically complete and actually operating.
Import, currency, and logistics. Long-lead equipment, port and inland transport, and currency or import friction stretch procurement timelines in ways a domestic Gantt chart tends to compress.
Community and local stakeholders. Local acceptance is not a formality to be secured late. Where it is treated as one, it becomes the thing that stops a project that was otherwise ready.
Why this is analysis, not colour
The temptation is to treat local knowledge as texture — useful anecdotes to sprinkle around a model built on standard assumptions. That gets the relationship backwards. In these markets the local factors are frequently the base case: they set the real schedule, and therefore the real return. A permitting sequence, an interconnection queue, a wayleave corridor — these are quantifiable, and they belong in the timeline the model runs on, not in a paragraph after it.
Knowing how a project actually gets built here — which approval really gates the start, which corridor is the true critical path, how long an interconnection has taken in practice — is not local flavour added to a technical read. It is part of the technical read. Left out, the model is precise about the wrong number.