What we monitor — the widest independent view in Pakistan
This report is not a survey of opinions. It is grounded in what BijliBachao actually watches every day. Through Solar Performance Cloud we independently monitor around 7 MW of commercial and industrial solar across Pakistan, from 2 MW installations down to single rooftops — measured live, at the level of individual panel strings, across seven different inverter manufacturers.
A note on honesty, because it matters in a report: these are systems we monitor, not “projects delivered,” and the 7 MW is a conservative floor — several capacities are cautious estimates from each site’s own measured output, so the true figure is higher. Around 55 private homes are monitored separately and are never named. No individual client’s figures appear here.
- The spread is deliberately broad — flour and rice mills, steel, textiles and dyeing, pharmaceuticals, cold storage, footwear, poultry and dairy, retail, co-working, hospitality and more.
- That breadth is the point: the same performance patterns show up whether the load is a spinning mill or a cold store, and across every major inverter brand.
| Metric | Value |
|---|---|
| Commercial & industrial solar monitored | ~7 MW (a conservative floor) |
| Sites | 43 |
| Inverters, watched string by string | 70 |
| Inverter brands supported | 7 |
| Commercial & industrial clients | 20 |
| Range | 2 MW installations → single rooftops |
The performance gap is real — and measurable
Solar rarely fails all at once. It bleeds output slowly, through mechanisms a monthly electricity bill hides completely. The global data is now unambiguous: analysis of monitored fields by Raptor Maps found that average PV project power loss doubled over five years, and that equipment underperformance cost U.S. solar operators roughly $5,720 per MW in 2024 alone.
NREL’s field data puts hard numbers on the routine drag — availability and performance-loss factors that quietly separate what a system should produce from what it does. The categories are well understood; the problem is that, unmonitored, they are invisible until the savings simply fail to arrive.
| Loss source | What it is | Typical signature |
|---|---|---|
| String faults | A weak or dead string among many | Now the single largest loss category (Raptor Maps); invisible in the plant total |
| Soiling | Dust and smog on the glass | Gradual daily loss; steep in Lahore’s air |
| Degradation | Panels aging over years | Median ≈ 0.5% per year across decades of data (NREL) |
| Inverter downtime | Trips and outages | A large share of unplanned downtime for a small share of hardware cost |
Why the gap costs more in Pakistan
The same lost kWh is not worth the same everywhere. Pakistan’s industrial power tariff is among the highest in the region, and for a textile mill, energy is 12–18% of total input cost. When electricity is that expensive, every unit a solar system fails to produce is bought back from the grid at a premium — so underperformance here punishes the balance sheet harder than in most markets.
The environment makes it worse. A Lahore-specific study measured soiling losses reaching roughly 0.8% of output per day at typical tilt without cleaning — among the highest rates recorded anywhere — which means an un-inspected system in Punjab drifts away from its potential faster than one in a cleaner climate.
And the 2026 net-billing reform changes the maths again: a self-consumed unit now avoids the full retail tariff, while an exported unit earns only a low buyback rate. Keeping every unit flowing, and using it on site, has never been worth more.
Why your inverter app doesn’t show it
Most solar owners believe they are already monitoring, because the inverter came with an app. But an inverter’s own app reports its own hardware’s numbers, one brand at a time, and mostly tells you what already happened. It shows “green lights” — the system is on — not lost revenue.
The failure mode that costs the most is precisely the one an app misses: a single string quietly underperforming while the plant-level total still looks normal, so nobody investigates. That is the difference between reactive monitoring (a dashboard you glance at) and active inspection (a system that continuously compares expected against actual and tells you where the gap is).
How the gap gets closed: independent, string-level inspection
Closing the gap requires three things an inverter app cannot offer. First, independence: the monitoring layer must not be owned by the same hardware whose faults it is meant to catch. Second, multi-brand reach — real portfolios mix Huawei, Sungrow, Solis, Growatt, GoodWe and others, so the inspection layer has to read them all from one place. Third, string-level depth, because that is where the faults hide.
This is exactly what Solar Performance Cloud does: it reads every major inverter brand, grades performance down to the individual string, and pairs AI analysis with solar engineers who verify each issue before an owner ever hears about it. Doing that across ~7 MW and seven brands is, as far as we know, unmatched in Pakistan — and it is why this report can exist at all.
What good performance looks like
The international standard for measuring PV performance, IEC 61724-1, formalises the discipline — the metrics (such as Performance Ratio), the sensors, and the data quality a credible assessment is built on. Against that yardstick, “good” is simple to state: a system producing close to what its size and site should yield, month after month, with any shortfall found and named rather than absorbed.
The practical test for any solar owner is to compare output per kW against a healthy past month or a similar nearby system. If it has dropped beyond seasonal variation, the energy — and the money — is going somewhere. Independent inspection is how you find out where, and independent inspection at scale is how a market finally learns what its solar is really doing.
