Three to twenty distributed solar and battery sites is the range we see among many operators. Most began with one or two sites and built a manual dispatch routine that worked. As they expanded, they kept using it at ten sites, even as strain became clear. Revenue per site fell, dispatch errors appeared in settlement reports, and the operator could see the process was failing without knowing exactly where.
Manual dispatch is not inherently unsuitable at small scale. Its failure appears when coordination grows, usually around ten sites. Identifying the source of that failure helps determine when to change the process.
The daily coordination requirement grows faster than the portfolio
A one-site routine uses much the same inputs each day: review tomorrow's day-ahead prices, set the overnight charge target, then log in after the morning session clears and adjust if necessary. For one site, this takes perhaps twenty minutes of focused attention daily.
A second site does not necessarily double the workload when both share a price zone and have similar battery configurations. The price signals are common, so the decision for site 2 largely follows site 1. Coordination rises modestly rather than proportionally.
The failure is nonlinear but predictable. As the portfolio grows, sites in different locations produce different solar profiles on the same day. The overnight charge target for site 3 in Setubal may therefore differ from the target for site 7 in Faro. Different grid zones can impose different export limit constraints. Batteries installed at different times also vary in chemistry and degradation state, so round-trip efficiency assumptions diverge.
At ten sites, one uniform dispatch decision is no longer a reasonable approximation. There are ten separate decisions, each with its own inputs and its own opportunity cost when wrong.
The intraday session problem multiplies by site count
In MIBEL zones, the intraday market has six discrete sessions each day. Each can revise prices for hours when sites are still charging. If session 3 closes with higher-than-expected prices for 18:00, an operator managing ten sites must check whether each site's SoC and charging path are suitable for that window.
That creates approximately sixty decision points daily across a ten-site fleet, before follow-on adjustments. Most are simple individually, but they arrive throughout the day, cannot be batched because they rely on real-time SoC readings, and carry a direct revenue cost when missed.
Operators managing ten sites manually report the same behavior: urgent signals get handled while others wait. Attention usually goes to sites that recently caused a problem, not those offering the most available spread that day. Human attention is limited, but this prioritization is poor for revenue optimization.
Price zone divergence creates silent revenue losses
In normal MIBEL conditions, Portugal and Spain clear at the same price, allowing one price feed to guide every site. During congestion, the zones separate. Portuguese and Spanish prices can diverge by 15 to 30 EUR/MWh for sustained periods.
Applying Spanish day-ahead prices to Portuguese sites during congestion means using the wrong signal. Batteries then charge and discharge against a price that does not represent their market clearing. Unless zone-specific clearing prices are checked, the mistake may stay hidden in day-ahead data. It appears later as lower settlement revenue, often blamed on "intraday volatility" instead of a price feed mismatch.
Manual processes seldom catch this. At 13:00, after day-ahead clearing, the operator may not know MIBEL zone separation occurred or may lack time to retrieve zone-specific data for every site. A dispatch system with per-site zone assignments sends each site to the correct price feed automatically, without operator intervention.
SoC data staleness accumulates across the portfolio
Manual dispatch relies on the operator's latest view of each site's state of charge, which is always somewhat old. At 14:30, the operator logs into site 4's inverter portal and sees 60% SoC. When a MIBEL session clears at 15:45 with attractive pricing for 17:00, they assume site 4 remains near 60%. Yet solar generation added 15% and household demand used 8%, leaving site 4 at 67% SoC.
With one site, this staleness is manageable. Across ten sites, each with its own SoC trajectory, solar generation rate, and load profile, the operator's mental picture increasingly separates from the fleet's actual state. Decisions based on stale SoC create inferior cycles, especially when a site assumed to be low is actually charged for the target window.
The scaling threshold and what it implies
Ten sites is not an exact threshold. Some operators handle eight manually with careful processes and few misses. Others encounter failure at six because their portfolios differ more or occupy price zones that require separate tracking. The figure approximates the point where coordination reliably exceeds the operational capacity of one person giving dispatch a reasonable amount of daily time.
The practical implication is clear. An operator with seven sites planning to reach fifteen should not wait for manual dispatch to fail visibly before adding automation. Silent losses, missed intraday sessions, and stale SoC decisions already occur at seven sites, then grow with the portfolio. Establishing dispatch infrastructure before the process breaks avoids a period of inferior performance that would otherwise be paid for during the transition.
We are not claiming automation becomes the right choice at one fixed portfolio size. The economics depend on battery capacity per site, the available spread in the relevant market, and the automation platform's actual cost. The case for manual dispatch does weaken predictably as the portfolio expands, so evaluate the change before coordination becomes a daily crisis, not afterward.
