'My app says zero'
A customer rings: their app has shown no generation for two weeks. They only noticed because their bill did not drop. The inverter has an error code. Your team logs in to the manufacturer's portal and there it is, flagged as offline for sixteen days, among dozens of other alerts nobody looked at. Some are wifi dropouts. Some are real.
You have systems from three or four inverter brands in the field, each with its own portal and login.
Why faults go unnoticed
- Each inverter or battery manufacturer has its own monitoring platform.
- Alerts go to an email address set up at install, which may be a person who has left.
- Most alerts are communication dropouts, so real faults are buried in noise.
- Nobody compares a system's output against what similar systems nearby produced that day.
- There is no routine for who looks at alerts and when.
What late fault detection costs
Customers who find out first and feel let down. Lost generation for the customer while the fault continues. Reviews that mention the fault and the wait. And missed chances to offer after-care services, because your only contact with customers is when something has gone wrong.
One view across monitoring platforms
- Alerts and daily generation are read from the monitoring platforms you use, through their APIs where the manufacturer provides one and you have access.
- Each system is linked to your job record: customer, address, equipment and install date.
- Communication dropouts that fix themselves are separated from alerts that persist.
- Each system's output is compared with others nearby on the same day, so a system producing far less than its neighbours stands out even without an alert.
- Systems that look wrong go onto a triage list with the evidence, for a person to review.
- From the list, the office can contact the customer, open a warranty claim or book a visit.
- Customer consent to monitor their system is recorded, and customers who decline are left out.
The view flags systems that look wrong. Diagnosis stays with your qualified engineers.
Triage list examples
| What the system shows | Likely category | Next step |
|---|---|---|
| Offline for a day, back online | Wifi dropout | No action |
| Offline for several days | Comms or power issue | Contact customer |
| Online, output far below neighbours | Possible fault or shading | Engineer review |
| Inverter error code persists | Fault | Engineer review, warranty check |
Calling the customer before they call you
Real faults are spotted within days rather than weeks. Customers hear from you first. Your engineers see the evidence before they visit. And after-care becomes something you can offer and deliver, not just a line in the proposal.
Comparing neighbours, not only alerts
Some of the most useful signals are not alerts at all. A system can be online, show no error code and still produce much less than it should, because of a failed string, a tripped isolator or new shading. Comparing each system's daily output with others you installed nearby, on the same day and under the same weather, makes those systems stand out.
It is not a diagnosis, and it will flag some systems for innocent reasons, such as a customer who has had a tree grow or a system with a different orientation. That is why the list goes to a person who knows the installs, not straight to the customer.
Are faults finding you late?
- Customers tell you about faults before you know.
- Monitoring alerts go to an inbox nobody watches.
- You have several monitoring portals and logins.
- You cannot tell dropouts from real faults without logging in.