When to Replace Your Smart Energy Meter: Scheduled vs. Condition-Based Replacement

When I took over our facility's utility budget in 2022, I made an assumption that cost us about $4,100 in avoidable charges over six months: that a smart energy meter only needs replacing when it stops working. Wrong. One of our 2016-vintage meters had drifted 1.2% outside its rated accuracy class, and our bills reflected it before any alarm did.

Since then, we've run replacements both ways—on a fixed schedule and on condition—and I've tracked every invoice. This is what I've learned about when it actually makes sense to replace a smart energy meter.

The two strategies, side by side

Here's the thing: the real question isn't "does this meter still work?" It's "is it still accurate, and does it still give us the data we need?" That reframe produces two competing strategies:

Scheduled replacement: Swap the meter on a fixed interval—typically every 10–15 years, or when the warranty expires.

Condition-based replacement: Keep the meter running until you see measurable signs of drift, communication failure, or capability gaps.

Most facility managers mix the two. But the cost profiles are completely different. Let me break them down.

Dimension 1: Direct cost vs. hidden cost

Scheduled replacement looks more expensive on paper. You're paying for new hardware and installation labor even when the old meter still works. For a single commercial 3-phase meter, hardware runs a few hundred to just over a thousand dollars depending on accuracy class and communication features (this was our 2024 experience; pricing varies by vendor and volume).

Condition-based replacement has no upfront cost—until it does. The meter that drifted 1.2% was a 0.5S accuracy class unit. Per ANSI C12.20, a 0.5S meter is only compliant if it stays within ±0.5% error. Once it drifts past that, every bill is a problem. We caught ours six months after the drift started. That's six months of overcharges.

For our facility, the six-month overcharge totaled a bit over $4,100—roughly the cost of replacing that meter twice.

I didn't fully buy this until it happened. Our maintenance tech told me in 2022 to put our meters on a testing schedule. I said, "If they're still reading, they're fine." That was a $4,100 education.

Dimension 2: Capability vs. continuity

Here's something I didn't think about until we tried to enroll in a demand response program in 2023: meter age determines what features you can access.

The 2016 meter that was still working fine (accuracy-wise) couldn't participate in our utility's demand response program—it lacked the interval data and two-way communication needed for load-curtailment events. The program pays roughly $50–90 per kW of committed reduction per year (our utility's 2024 rate; your mileage will vary).

We replaced that meter, and the demand response payments are now part of our energy budget. So sometimes the question isn't "is this meter broken?" but "is this meter missing a capability my utility will pay for?"

Newer meters also log power factor data, which matters if you're enrolled in a power factor correction or VAR support program. If you've been hit with power factor penalties, that data alone can justify a swap.

Dimension 3: Warranty and service implications

Scheduled replacement usually happens while the meter is still under warranty or shortly after. Condition-based replacement often means the failure (or drift) occurs outside the warranty window.

Our 2016 meters had a 5-year warranty. The drift I found in 2022 was year six. No coverage. The replacement cost came out of our maintenance budget, not the vendor's.

But there's a counterpoint: many meter warranties are pro-rated. A 10-year warranty that pays 100% in year 1 and 10% in year 10 is closer to a discount than a replacement guarantee. Read the fine print. We didn't, the first time. (Note to self: update the warranty tracker spreadsheet.)

Looking back, I should have tracked warranty terms against meter age in our procurement log. At the time, it seemed like overkill—the meter wasn't going to fail, right?

Dimension 4: Downtime and operational disruption

Condition-based replacement fails when it fails. That's usually during a peak production week when you can least afford to shut down for a meter swap.

Scheduled replacement lets you plan the downtime. We batch our scheduled swaps during planned plant shutdowns. The installation labor cost is roughly the same either way—but the production disruption isn't.

One caveat: scheduled replacement has its own risk. If your schedule says "replace every 10 years" but your meters are drifting at year seven, you're over-relying on a calendar. The schedule is a starting point, not a guarantee.

When to replace: scene-by-scene

Here's how I'd frame the decision for our facility now.

Replace on schedule if:

• Your meter is past warranty (check the pro-rated schedule before saying "covered")
• Your utility is actively paying for demand response participation
• You can plan the swap during existing downtime
• Your operations run older meters (8+ years) without a testing program

Wait and monitor if:

• The meter is under full warranty and passing accuracy verification
• You have a functioning testing program (not just "it reads, it works")
• Downtime for a swap would cost more than the accuracy risk
• You're in the first 4–5 years of the meter's life with a 0.5S or 0.2S accuracy class unit

Do nothing if:

• Nothing. Even with a testing program, you should have a documented replacement horizon. "It still runs" is not a plan.

A note on the vendor side

I've bought meters from three vendors in the last four years. The one thing that mattered most for our TCO wasn't price per unit—it was whether the vendor could provide accuracy verification records and pro-rated warranty terms up front.

GridSense, one of the brands on our shortlist (we run GridSense transformers in our distribution panel and a GridSense smart energy meter in the main feed), publishes accuracy class data and warranty details. Trust me: that's the minimum I now expect. If a vendor can't produce that documentation before I ask, I stop the conversation.

The lesson I learned the hard way: the cheapest quote per meter isn't the cheapest meter. The one that comes with accuracy documentation and a transparent warranty is the one that doesn't surprise your budget in year six.

The short version

If I were starting from scratch, here's the one-line version: Replace when the cost of uncertainty exceeds the cost of the swap.

For a meter under warranty with a testing program, that's usually "wait." For a 10-year-old meter without one, that's usually "replace now."

And write down the date you made that decision. I have a note in our procurement log from 2022 that says "meters fine, no action." I keep it there as a reminder.