The Storm Season Readiness SOP for Maintenance Plan Customers
Purpose
A maintenance plan customer bought one thing: a generator that starts when the lights go out. The outage season is when that promise is tested, and it is the worst possible time to discover a dead battery, a unit that stopped exercising in the spring, or a propane tank at a quarter. Once the storm arrives, every plan customer with a problem calls at once, the roads are bad and the parts houses are empty. This SOP works the plan list in the weeks before the season so the problems are found while there is still time to fix them in order of risk, and so the shop goes into the first storm with the parts staged and the on-call rotation published.
Scope
Covers: every active maintenance plan customer, run once a year ahead of the region's main outage season (hurricane season, the winter ice season, or the summer storm peak, whichever the shop's service area lives with). The shop's default is to start this SOP 60 days before the season's start date.
Does not cover:
- The service visit itself, which is the Annual Generator Maintenance SOP.
- Remote monitoring setup, which is Remote Monitoring Setup and Customer Handoff Technique; this SOP only reads its status.
- Response to an outage call, which is Unit Failed to Start During an Actual Outage, Response.
- Non-plan customers, who may get a marketing message but no readiness call.
Roles and responsibilities
| Role | Owns | Hands off to |
|---|---|---|
| Service manager | Steps 1, 6 and 8; the cut-off date | Office, with the ranked list and dates |
| Office | Steps 2 to 5 and 9: data pulls, calls, customer notices | Service manager, with every flag filled |
| Parts lead | Step 7 | Service manager, with stock against the list |
| Technicians | The visits step 6 schedules; on-call per step 8 | Office, with visit results back in the record |
Procedure
Pull the plan list with the fields this SOP needs. Export every active plan customer with unit size and fuel, install date, last completed service date, remote monitoring status, battery install date, and whether anyone at the address depends on powered medical equipment. Acceptance: one row per active plan, every field filled or marked unknown. Wrong looks like: rows with no last service date or no battery date, which usually means the visit was done but not recorded; stop, have the office reconcile against the work orders before ranking, because a blank ranks as safe when it is not. Hazard: none at this step, it is office work.
Flag overdue service. Mark every customer whose last completed annual service is more than 12 months before the season start date. Acceptance: each row marked current or overdue against that date, with the count recorded. Wrong looks like: a visit scheduled but not completed, counted as current; only completed visits count. Hazard: none at this step.
Read remote monitoring status. For each monitored unit, record whether it is online, the result of the last four scheduled exercises, and any active fault or warning. Where the platform reports run hours, record them too. Flag a unit offline for more than 7 days, any missed or failed exercise in the last four, any active fault, and any unit whose hours since its last oil change are near the maker's hour interval, as after a long outage; that one counts as overdue service. Acceptance: each monitored row carries online or offline, an exercise count out of four, and a fault entry or "none". Wrong looks like: an offline unit treated as fine because nothing has alarmed; offline means unknown, and an unknown unit is flagged, not passed. Hazard: none at this step.
Flag battery age. Mark every battery at or beyond 3 years from its install date, which is this shop's default for a pre-season load test (tune it to your climate and battery type; heat shortens life), and any battery that failed or was marginal at its last load test. See Choosing a Standby Generator Starting Battery for the type and rating behind the default. Acceptance: each row marked by age band, with the count of aged batteries recorded by group size. Wrong looks like: a battery date copied from the unit install when the battery was replaced since; check the last battery work order. Hazard: none at this step.
Call every propane customer for a gauge reading. Ask the customer to read the percentage on the tank gauge and nothing else: no valves, no regulator. Flag any tank below 60 percent, the shop's default trigger to arrange a fill ahead of the season (the supplier fills to its own maximum, commonly about 80 percent; a reading alone does not prove cold-weather runtime, which How to Check a Propane Tank for Generator Runtime and Vaporization covers), and remind the customer that the supplier sets and fills the tank and to call the supplier, or authorize the shop to, for a delivery. Acceptance: a gauge reading and date for every propane customer, and a delivery date for every one below 60 percent. Wrong looks like: a customer who cannot be reached after two calls on different days; stop trying by phone, send a written notice asking for the reading, and flag the row unknown for step 6. Hazard: tell any customer who smells gas at the tank to leave the area and call the supplier or 911 from a distance, not to look for the leak.
Rank and schedule the visits by risk. Every customer with any flag gets a visit or a call-back. Rank in this order: medical dependency with any flag, active fault, monitoring offline or missed exercises, overdue service, propane unknown or below trigger, aged battery only. Schedule highest rank first, all visits before the season start date. Acceptance: every flagged customer has a visit or call-back date on or before the cut-off, with the highest ranks in the first week. Wrong looks like: more visits than the technicians' days can hold before the cut-off; stop and take it to the owner, who decides between overtime, a subcontractor, or telling the lowest-rank customers in writing that their visit falls after the season starts. Hazard: none at this step; the visits carry their own procedures.
Stage the parts the list predicts. Count from the list: batteries by group size, oil and filters by engine, air filters, spark plugs, controller and sense fuses. Batteries are staged against the shop's historical failure rate on aged batteries rather than one per aged battery. Acceptance: stock on the shelf at or above the predicted count for each part, recorded. Wrong looks like: a part the supply house cannot deliver before the cut-off; stop, find a second source or move the affected visits later in the schedule so they are not booked without the part. Hazard: batteries are heavy and hold acid; lift with the back straight, carry them upright, and strap them upright on the truck shelf, never loose in the cab.
Publish the emergency call rotation. Set the on-call schedule for the season's peak weeks, with a primary and a backup for each day, a triage script that routes failed-start calls to Unit Failed to Start During an Actual Outage, Response, and the ranked list loaded on every on-call phone or tablet so the medical-dependency customers get the first truck. Acceptance: rotation published to every tech, each on-call truck stocked from step 7, the list accessible offline. Wrong looks like: a rotation with no backup, or a list only on the office computer that goes dark with the power; stop and fix both before the cut-off. Hazard: storm calls mean downed lines and flooded roads; every tech treats any downed line as live, stays well back from it and anything it touches, reports it to the utility, and does not drive through moving water.
Tell every plan customer where they stand. Send each a short notice: their plan status, the visit date if one is booked, what to expect in an outage, and the emergency number. Acceptance: notice sent to every active plan customer, with bounced or undeliverable contacts followed up by phone. Wrong looks like: a contact that no longer works; update it now, because it is the number the office will need during the storm. Hazard: none at this step.
The record this produces
A season readiness sheet per customer, kept on the customer record: each flag with its value and date, the propane gauge reading, the rank, the visit or call-back date, the visit result, and the notice sent. A summary for the season: counts per flag, visits completed before the cut-off, parts staged against used, and calls taken during the season. The service manager reads it after the season to set next year's defaults; the on-call tech reads the customer row at 3 a.m. during the storm.
Worked pass: 140 plan customers, 60 days out
Step 1. 140 active plans pulled. Four rows had no battery date; the office found three replacement work orders and marked the fourth unknown.
Step 2. 18 customers past 12 months since their last completed service.
Step 3. 96 units monitored. 11 offline more than 7 days, 6 with one or more missed exercises out of the last four, 4 with an active fault.
Step 4. 37 batteries at or beyond 3 years, plus the one unknown: 38 to test.
Step 5 failed its acceptance for 5 customers. Of 52 propane customers, 47 gave a reading on the first or second call; 9 of those 47 read below 60 percent and had deliveries arranged. The other 5 could not be reached after two calls on different days, so there was no reading for them. The stop rule triggered: written notices went out and those 5 rows were flagged unknown. Three replied within a week, and two of those three were below 60 percent, so those two had deliveries arranged too. The last two were visited in person as part of step 6.
Step 6. Flags overlapped heavily; the unique visit list came to 41 customers, 3 of them with medical dependency. At 2 technicians doing 4 plan visits a day each, that is 8 visits a day, and 41 / 8 = 5.1, so 6 working days. Season start was 60 days out, so capacity was not the constraint. The 3 medical-dependency customers were seen in the first two days.
Step 7. Last season about one in three aged batteries failed its pre-season load test. 38 / 3 = 12.7, so 13 batteries staged, split by group size per the list. Oil, filters and plugs staged for the 18 overdue services.
Step 8. Rotation published for the eight peak weeks, primary and backup each day, list synced to every tablet for offline use.
Step 9. 140 notices sent, 6 bounced, all 6 updated by phone.
After the season. 12 of the 38 aged batteries failed their load test, about one in three, so 12 of the 13 staged were used and the rate stays for next year. The shop took 23 outage calls during the season; none came from a customer whose unit had been seen on a readiness visit.
References
- See related: Remote Monitoring Setup and Customer Handoff Technique
- See related: Annual Generator Maintenance SOP
- See related: Unit Failed to Start During an Actual Outage, Response
- See related: Choosing a Standby Generator Starting Battery
- See related: How to Check a Propane Tank for Generator Runtime and Vaporization