Most telecom energy optimisation proposals are built the same way: measure the current electricity or diesel bill, estimate a percentage reduction from the proposed intervention, multiply by the tariff, and present the resulting saving against the capital cost. It's a clean spreadsheet. It is also, on its own, usually the weakest possible version of the business case — because it ignores most of where the value actually sits.

Energy cost is real, but on a typical off-grid or hybrid site it's rarely the largest number in the room. The business case that gets funded is the one that accounts for what actually drives cost and risk on a distributed infrastructure estate.

Where the value actually comes from

Generator runtime, not just fuel

Reducing generator runtime saves diesel, but diesel is often the smaller part of the genset cost story. Runtime reduction also defers maintenance intervals, reduces the frequency of site visits in areas where a truck roll carries real logistics and security cost, and lowers exposure to fuel theft — a material line item on many African telecom estates, independent of the fuel's market price.

Battery life extension

How a site is operated — dispatch strategy, depth of discharge, thermal exposure — directly affects how long the battery lasts before replacement. A well-designed optimisation programme (better cooling strategy, smarter charge/discharge logic) can measurably extend replacement intervals. That's a deferred capital outlay, and on a large estate it is frequently a bigger number than the energy saving itself.

Site availability

Energy-related outages — batteries depleted before generator start, generators that fail to start, thermal shutdowns — carry a cost that rarely appears in an energy spreadsheet at all: SLA penalties, lost traffic, and reputational cost with the network operator. A resilience improvement that reduces energy-related downtime by even a small margin can outweigh the fuel saving on its own.

Reduced truck rolls

Every avoided site visit — for fuel delivery, battery fault response, or reactive generator maintenance — has a direct logistics cost and an indirect cost in technician time that could be spent elsewhere. On remote or hard-to-access sites, this is often the single largest driver of total cost of ownership, and one that a narrow energy-saving case doesn't capture at all.

FUEL / ENERGY GENSET MAINT. BATTERY LIFE SITE AVAILABILITY TRUCK ROLLS
FIG. 01 — ILLUSTRATIVE COMPOSITION OF VALUE WEIGHTING VARIES BY SITE PORTFOLIO

Why the single-site poster child doesn't scale

Portfolios rarely behave like their best example site. Climate zone, tower type, backhaul load, cabinet design and grid reliability vary widely across an estate, and each of those factors changes which lever matters most at a given site. A business case built on one strong pilot site, then extrapolated across the portfolio, tends to overstate the outcome — the pilot was usually chosen because conditions were favourable. The credible version of the case is modelled at a portfolio level, with sensitivity across site categories, not scaled up from a single result.

The practical implication

A CFO evaluating an energy optimisation programme is really being asked to fund a resilience and maintenance-cost programme that happens to also save energy. Presented that way, against the full cost base it affects, it competes very differently for capital than a pure utilities-saving line item does.

Building the case properly

  • Start from an actual site-level energy and operations baseline, not the aggregate utility bill — the levers that matter differ by site category.
  • Quantify generator maintenance and fuel-theft exposure alongside fuel cost, not as a footnote.
  • Link battery dispatch and thermal strategy explicitly to replacement-interval assumptions in the model, rather than treating battery life as fixed.
  • Price site availability improvement using the operator's own SLA and outage-cost framework, where available.
  • Model truck-roll reduction using real logistics cost per site category, not a flat estimate.
  • Run the model across representative site clusters, with sensitivity ranges, before presenting a portfolio-wide number.

None of this makes the case more complicated to build than the simple version — it makes it complete. The spreadsheet with only fuel and electricity in it isn't wrong; it's just answering a much smaller question than the one capital allocators are actually deciding.