Panels get dirty between site visits. The cleaning schedule doesn't know it. By the time SCADA shows the dip, the revenue is gone.
Clean every 90 days — regardless of what's actually on the panels. You over-clean clean sites and under-clean dirty ones. Always.
Unoptimised spendDrive to site, walk the rows, estimate visually. Costs $500–$2,000 per visit, happens quarterly at best, and can't cover every panel.
$2–8K per farm / yearPyranometers measure one spot and miss the spatial variation that causes the biggest losses. $5,000–$20,000 per site to install and maintain.
Blind between sensorsProduction dips appear in SCADA days after soiling takes hold. By the time you dispatch a crew, weeks of revenue loss are already locked in.
Always too late"Cleaning on a calendar means we're either too early or too late. We need to know which farm is dirty right now — not find out after the SCADA reading drops."
Director of Operations, solar O&M company — prospect interview, 2025
200,000 MWh/yr at $40/MWh. Soiling loss data: Wood Mackenzie 2025. SkySense subscription: $18K/yr (Professional tier).
| Cost category | Without SkySense | With SkySense |
|---|---|---|
| Manual inspections / year | $6,000 | $0 |
| Cleaning cycles / year | 4 × $5K $20,000 | 3 × $5K $15,000 |
| Soiling production loss | 5% $400,000 | 2% $160,000 |
| SkySense subscription | — | $18,000 |
| Total annual cost | $426,000 | $193,000 |
Scales linearly with portfolio size. 500 MW portfolio → ~$1.1M in addressable annual savings.
Net annual savings per 100 MW
Share farm boundaries, 12–24 months of SCADA production history, and your current cleaning schedule. Our team handles the rest.
Multi-spectral imagery acquired every 5–7 days via ESA Sentinel-2. Daily cadence available on premium tiers.
ML models separate soiling from cloud cover, shading, and degradation — no false positives from weather.
Dashboard alert with production loss estimate, cleaning cost vs. savings, and portfolio priority ranking.
Every feature exists to recover lost revenue or reduce wasted cleaning spend. Nothing else made the cut.
Full spatial soiling map across every panel row at every site. Identify the exact zones driving production loss, not an averaged reading from one point on the ground.
Every alert includes a cleaning ROI score: estimated production loss per week, cleaning crew cost, and net payback. Stop sending crews to sites that don't need it yet.
One view, all sites, sorted by soiling urgency and revenue impact. Know exactly which farm your cleaning crew should hit first, across a portfolio of any size.
Weekly automated imagery catches rapid events — dust storms, pollen seasons, bird events — before they compound into multi-week production losses. Daily monitoring available on premium tiers.
API and webhook integrations with kWh Analytics, AlsoEnergy, Locus Energy, and custom SCADA. SkySense adds a layer — it doesn't replace what you have.
One-click PDF: portfolio soiling status, production loss attributable to soiling, cleaning actions taken, and trend vs. prior periods. White-label available on Business and Enterprise tiers.
Onboarding uses your SCADA history and cleaning records to establish each site's unique production baseline. Results are specific to your assets — not benchmarked against industry averages that may not reflect your location, panel type, or soiling profile.
You manage multiple sites. Your cleaning budget is fixed. You can't afford to over-clean, and you definitely can't afford to under-clean and miss an SLA. SkySense tells you which farms need a crew this week — and which don't.
You own the assets. Your O&M partner handles cleaning. How do you know they're doing it at the right time — and not just on a schedule that protects them from complaints? SkySense gives you independent, satellite-verified visibility.
Post-construction O&M is a commodity. Every proposal looks the same. Add satellite soiling intelligence to your O&M package and you're offering something your competitors can't — without adding headcount.
The core technical challenge in this space is distinguishing soiling from cloud cover, shading, and natural panel degradation — which general satellite analytics platforms get wrong. SkySense models are trained specifically on soiling signatures and cross-reference solar irradiance data, aerosol optical depth (NASA MERRA-2), and weather event history to isolate soiling with high confidence. False positives trigger you to send an unnecessary cleaning crew — so we are extremely conservative about alerts: we only fire them when the ROI calculation is unambiguous. During your 3-month pilot you'll be able to validate our accuracy against your own on-site observations.
Yes — with an important clarification. ESA Sentinel-2 at 10 m resolution does not resolve individual panels, but it does resolve panel rows on utility-scale farms (50 MW+). Soiling analysis works at the zone level: we identify which sections of a farm are soiling faster, which is the actionable unit for dispatching a cleaning crew. For farms under 20 MW or where panel-level resolution is required, we offer Planet SkySat at 50 cm resolution (available on premium tiers). Your pilot will use the imagery tier appropriate for your site sizes.
Cloud cover reduces imagery frequency — in heavily overcast regions a 5-day revisit cycle may yield one usable image per 2–3 weeks. That said, heavily cloudy regions also have naturally lower soiling rates (rain cleans panels). The ROI for SkySense is highest in the US Southwest, Southeast, Middle East, India, Chile, and Spain — where soiling is aggressive and imagery is consistent. During your pilot we'll give you a honest assessment of expected imagery frequency for your specific sites. If cloud cover makes weekly detection unreliable for your locations, we'll tell you upfront rather than sell you a subscription that underperforms.
SkySense is a precision analytical platform, not a plug-and-play dashboard — and the onboarding reflects that. Before your first soiling report lands, we work through a structured setup together: site boundary coordinates (KML, shapefile, or spreadsheet), 12–24 months of historical SCADA production data so our models can establish each farm's baseline, and a briefing on your current cleaning schedule and any known soiling events. This prep is what makes your results meaningful — our analysis is calibrated to your specific sites, not run against generic solar farm parameters. In practice, most clients complete the data handover in one to two weeks, and first results follow within one satellite revisit cycle of your farms (typically 1–3 weeks depending on location and cloud cover). Our team handles everything on the technical side — you won't need engineering resources beyond the initial data pull.
Your existing monitoring platform shows you production data. SkySense shows you why production is dropping and — critically — whether a cleaning run will recover it. These are different signals. kWh Analytics, AlsoEnergy, and Locus Energy are our integration partners, not our competitors: SkySense soiling data surfaces inside those platforms via API. Think of it as the soiling intelligence layer your monitoring platform doesn't have. The ROI question is simple: if SkySense costs $18K/year and saves you one unnecessary cleaning run ($5K) and recovers 2% of production loss ($160K savings), the math works by an order of magnitude. The pilot lets you verify those numbers on your actual farms before you commit.
Don't see your question?
Talk to our team before you apply →Not a free trial. A real engagement — your farms, your SCADA data, your cleaning schedule — analysed against actual satellite imagery to produce results you can take to an asset owner or board.
We'll confirm fit and walk you through what the 90 days involves before anything starts.
Enterprise portfolios (30+ sites) — contact us for custom scoping.