Comcast, who announced a business — MachineQ — with ambitious ideas about creating a nationwide LoRaWAN network, is shifting gears. Lots of problems with the original plan which I’ll touch on below but here is the data:
“When we first started MachineQ, we were out there building networks, strategically, to see what customers and what the market wanted,” said Alex Khorram, MachineQ’s general manager.
The resulting feedback was that many customers desired “a different relationship with the technology” that enabled them to deploy low-power networks focused on their manufacturing facilities, medical campuses or retail locations.
“They just needed the basic tools,” Khorram said. Instead of relying solely on IoT networks built and managed by MachineQ, customers and other potential partners were increasingly interested in more of a do-it-yourself model.
Was it the monthly subscription fees? The weak networking stack? The high costs of exiting a carrier relationship? Security concerns? An unclear carrier value proposition? Hard to say.
“They would rather purchase products and software tools from MachineQ that, in turn, put them in position to install the systems, onboard the IoT sensors at their various locations and then have the ability to manage the whole thing.”
“And the general connectivity could either come from Comcast, if available, or from another service provider.”
By “general connectivity” this implies roaming, which LoRaWAN is notoriously bad at (see below). But reading this, what should stand out is this:
“They would rather purchase products and software tools from MachineQ”
So Comcast is pivoting its LoRa strategy from a conventional carrier paradigm to selling LoRaWAN hardware. This is all probably for the best for any number of reasons including:
– The capital outlays to add LoRa to residential gateways (an ambitious but smart idea, for other reason they seem to be missing) was artificially high. The cost premium of placing a Semtech LoRa gateway radio in Comcast residential gateways is ~$65 vs. using a LoRa endpoint radio at the gateway, as Haystack does. Comcast will certainly get a better deal than the $65 I am quoting here but when the guys in Comcast finance looked at this strategy, I’m assuming they had questions.
– The embrace of LoRaWAN as a networking stack was full of potholes from the beginning. Battery powered LoRaWAN devices are effectively unidirectional and bring high packet error rates, limited or no mobile support, weak signal propagation, and more.
– LoRaWAN is at best suitable for fixed use cases which would include meter reading and the like. These are almost entirely commercial-industrial sector applications with minimal overlap with Comcast’s installed base and distribution strengths. Because of what appears to be overconfidence about the capabilities of LoRaWAN, consumer use cases like asset tracking which are far better aligned with Comcast’s strengths were by definition not available and … a huge opportunity is left on the table.
– Roaming between private networks and other networks using LoRaWAN will require a fully-stocked wet bar. First, in true carrier fashion, it requires a SIM card. Second, this places Comcast at the center of a customer’s security policy, which is likely a reason — perhaps a primary reason for many — for the customer brush-off in the first place. Third, the costs of manually replacing SIM cards on unidirectional LoRaWAN endpoints remains unattractive to many.
– Carrier-based LPWAN strategies are challenging not only with LoRaWAN, but also with NB-IoT, as discussed here.
Comcast entered this space with a LoRaWAN-based offering that was undifferentiated in a marketplace full of LoRaWAN hardware that many customers believe they can deploy themselves. Now instead, Comcast plans to sell undifferentiated LoRaWAN hardware to those same customers.
Get the latest about Haystack by signing up for our newsletter here.