I used to pick cellular modules by unit price. I was wrong. The cheapest quote almost always ends up costing more after engineering hours, carrier certifications, firmware updates, and downtime.
The lowest unit price is rarely the lowest total cost. If your procurement process only compares invoice totals, you're not comparing total cost.
For context: I'm the office administrator for a 180-person logistics company. I handle all the connectivity hardware ordering—roughly $150,000 a year across 14 vendors. I report to operations and finance, so I see the same conversation from both sides. Operations wants it to work. Finance wants it to be cheap. The hard part is proving that those two are often the same thing.
How I ended up in the Sierra Wireless EM7345 project
In 2024, we were refreshing a few hundred field laptops and needed new 4G LTE modules. The generic module was $18 cheaper per unit than the Sierra Wireless EM7345. At 250 units, that looked like $4,500 in savings. My first reaction was to take it.
The first phone I ever chose for a company order taught me to hesitate. It was the cheapest model on the list. It worked in the office, but it hated elevators, parking garages, and basically any building with thick walls. We replaced three of them in six months. That phone cost twice as much the second year, but it showed up on the expense report as a different line item. Nobody connected it to the original price decision. So I started paying attention to TCO before repeating the same mistake.
We did a short evaluation with both modules. The generic one showed decent speeds on the bench. In a real drive test, it dropped the connection when the laptop moved between towers. The Sierra Wireless EM7345 didn't drop calls in that same test. I'm not going to pretend that's a lab-grade study—it was a single afternoon with a few devices and a lot of frustration—but it was enough to make me look deeper.
Firmware is part of the purchase
The deciding factor was firmware. The EM7345 uses Qualcomm's Snapdragon X7 LTE modem, and the Sierra Wireless Qualcomm Snapdragon update for the EM7345 improved handover behavior on one of our regional bands. That update mattered. The generic module didn't have an equivalent update path. When I asked when a fix might arrive, the answer was vague. Honestly, I've never fully understood why some vendors abandon firmware so quickly after a module ships. My best guess is their roadmap moves to newer, more expensive parts.
Meanwhile, our IT specialist spent a week building a workaround for the generic module's network drop issue—roughly $3,000 in loaded cost. Then we replaced ten modules that refused to activate, each eating a couple of hours of troubleshooting. The $4,500 in planned savings was gone before we even deployed the remaining units.
A better way to compare quotes
Since that project, I use a simple TCO framework. I still compare unit prices, but I add the following:
- Shipping and import handling, not just the base price
- Technical integration and certification time
- Firmware update history and future commitment
- Band support for the specific carrier areas we operate in
- Failure rate and RMA process (the cost of a dead module is never the module itself)
- Downtime when a device doesn't work as promised
Some of these items are hard to quantify. A lot of procurement advice pretends you can assign a dollar value to everything. You can't. But even rough estimates beat pretending the quote is the whole story. I'd rather say "this might cost us $5,000 in engineering time" than tell finance I have no idea.
Infrastructure hype doesn't replace device compatibility
I've also learned to ignore debates that don't help me. People love to compare infrastructure providers—Crown Castle vs another tower owner, for example—as if one name automatically means perfect coverage. From where I sit, the network is only as useful as the device that connects to it. A strong tower that can't negotiate the modem's band combinations still looks like a weak tower to the user.
I do pay attention to vendor track record, though. Sierra Wireless has been around in the cellular module space for a long time. That's not a guarantee, but it does mean their drivers and firmware tend to stay available. When you're buying modules for a multi-year deployment, a vendor's long-term behavior is a procurement decision, not a support issue.
What about the budget?
I get why someone would pick the cheaper module. Budgets are real, and it's hard to walk into finance with a more expensive invoice. I've had that conversation. But I've also had the harder conversation after a "savings" turned into an RMA project. That one is worse. If I can show finance the full expected cost—including integration hours and likely failure rates—I have a much better case for a slightly higher invoice.
Earlier this year, we had 48 hours to pick a module for a customer proof-of-concept. Normally I'd run a full evaluation. There was no time. I went with Sierra Wireless based on the EM7345 evaluation we already did. So glad I did—the demo didn't fall over. We almost ordered the generic module just because it was in stock. Dodged a bullet.
Final thought
One more thing: I've learned to be skeptical of broad claims. Per FTC guidelines (ftc.gov), marketing claims have to be substantiated. When a vendor says "works on every network" but won't show carrier test results, I treat that as a red flag. The Sierra Wireless EM7345's datasheet lists supported bands and carrier certifications. That's not as sexy as a marketing line, but it's a lot more useful.
This isn't a blanket endorsement of Sierra Wireless. They've had supply chain delays, and I've had to RMA another product before. But the EM7345 project changed how I buy. It made me look past the unit price and ask what a module would cost over the next two years of a deployment. That answer is where the real numbers live.
Prices and carrier band lists change fast. This was accurate as of Q1 2025, so verify current specs before you make a quote decision.