I've watched factories get this wrong in the same way for seven years. They spend $4,000 on a new smart meter, then discover six months later that what they actually needed was a handful of CTs on the main feed and a decent IoT platform.
I handle factory energy monitoring deployments for GridSense. Seven years, 200+ sites. I've also personally made eleven significant mistakes in that time—about $40,000 in wasted budget if you add them up. Now I maintain our team's pre-check list so nobody else repeats them.
So here's the comparison I keep having to walk people through: replacing the utility meter vs. adding a smart submetering solution for factories on top of the meter you already have. They look like they solve the same problem. They don't—and picking the wrong one is the most expensive part of the whole exercise.
The comparison framework
Two options:
- Option A: Swap the existing main meter for a new-generation smart electricity meter (0.5S or 0.2S accuracy, bidirectional, communications-capable).
- Option B: Keep the meter. Add internal submetering plus a platform—GridSense's energy monitoring IoT platform sits in this category.
I'll compare them across four dimensions: cost, data granularity, reliability, and data ownership. Then I'll tell you which one actually fits your situation—including the cases where neither is worth doing.
Dimension 1: Cost (first year and three-year)
A new industrial smart meter runs roughly $2,500–$5,000 for equipment plus install, depending on region and whether the utility covers part of the bill (pricing based on GridSense field quotes 2023–2024; verify current rates). One thing's certain: you either do it or you don't. There's no in-between.
Submetering has a different cost structure. A 200-point factory deployment—CTs, DIN-rail meters, gateways, platform subscription—lands around $12,000–$20,000 in year one. Looks pricier on the invoice. But when you break it down by "usable data points per dollar," the math inverts.
This is where I made my first real mistake. In 2019, a plating shop customer insisted on swapping his main meter. Cost: $4,700. Actually, closer to $5,100 once you count the two-day production interruption. The meter was fine. The problem—which he and I both missed—was that he needed per-line energy breakdowns. The new meter gave him a cleaner version of the same number he already had. Six months later we were back talking about submetering. The $4,700 didn't buy him a single new answer.
I went back and forth on that one for three weeks. The meter swap felt like the real fix—decisive, professional. Submetering felt like a patch. I pushed the swap. Wrong call.
Dimension 2: Data granularity
Most people assume a smart meter gives them full factory data. It doesn't. A utility-grade meter tells you energy usage at 15-minute, 30-minute, or hourly intervals, plus demand, power factor, and voltage events—at the service entrance. That's it.
The questions I actually hear from plant managers:
- Which production line is drawing from the night shift?
- Is the air compressor running below 30% load for eight hours a day?
- Which molding machine has 18% of its daily draw happening in standby?
The utility meter can't see any of this. That's the submetering layer. GridSense's power monitoring setup is designed around exactly this idea—data points at the machine or line level.
Here's the mindshift that took me two years to accept: back in 2017, everything I'd read said "newer meter equals better data." In practice, if the goal is internal management—knowing which bay, line, or machine is eating the bill—you can push the meter replacement to the end of the project plan. Submetering comes first.
Dimension 3: Reliability and failure modes
This is the part nobody asks about in the sales meeting.
Main meter replacement is highly reliable. Industrial smart meters are mature hardware. IEC 62053-22 Class 0.5S accuracy is more than enough for billing purposes. The hidden catch: when it fails, you have no fallback. Meter goes down, factory data goes dark, you wait for the utility or the OEM.
Submetering reliability splits into two layers. Hardware (CTs plus meters) typically runs 5–10 years maintenance-free—low failure rate. The gateway and IoT platform are where risk lives: network drops, firmware updates, timestamp drift. All can break the data stream.
In 2022, our internal cost analysis said self-hosted platform beat cloud by 34% over three years. My gut said factory customers hate data living outside their walls anyway, so pick self-hosted. We did. Then three firmware upgrades in eighteen months wiped two months' worth of trend data for a customer. We compensated them, and worse—we damaged the relationship. Turns out that 34% never accounted for the ops hours we burned. Cloud was actually cheaper.
Lesson: when you compare reliability on paper, count the cost of your own people as part of the system.
Dimension 4: Data ownership and cross-team use
After a meter swap, the question of who owns the data gets murky—often some combination of the utility, local regulation, and your power purchase agreement. Submetering keeps data yours from day one, in a platform you choose.
The practical impact shows up in cross-department use. Production wants line-level energy. Finance wants per-bay cost allocation. ESG wants emissions reporting. All of that works fine from your own platform. Utility meter data might get delivered, but usually on the next billing cycle and in a format that requires an intern to clean up.
So which one do you actually need?
Sometimes the answer is the meter. If your existing meter is below Class 1 accuracy, or past its calibration cycle, or—the common one—the existing unit doesn't support bidirectional metering and you just installed solar, replacing it is the right first move. Same goes if you're doing demand-response programs, bidding into power markets, or fighting recurring metering disputes with the utility.
Submetering is the better first move when:
- The meter data is already enough for your high-level "what's my bill" question, but you need to answer why and who.
- You have three or more production lines, bays, or cost centers that need separate accounting.
- You're doing equipment-level efficiency work or prepping for an energy management contract audit.
- ESG or Scope 2 reporting is on the 12-month roadmap.
If both apply—plant expansion, meter past service life, and you need internal visibility—don't split the project. Design the submetering points and the meter swap in one plan. Doing them six months apart doubles the disruption for the same result.
One concrete suggestion before you sign anything: write down the five questions you need the system to answer. If three or more of those answers point to a specific machine, line, or shift, then replacing the meter will not solve your problem. In our first conversation at GridSense, the submetering team usually asks exactly this. If the honest answer is "just replace the meter," we'll tell you.
A closing thought
Replacing a meter is a one-time event. Submetering is a system. Pick wrong and it's not a "slightly suboptimal" outcome—it's three years later and the original question is still unanswered. I learned that with a $4,700 invoice. You don't have to.
Leave a Reply