Congratulations to DevicePilot customer POD Point on their acquisition by energy giant EDF. We're proud to play our small part in their journey.
A lovely definition of product-market fit is Sequoia's: "The customer's hair is on fire and all you have is a brick – so they'll try to put the fire out with the brick". Over the last few years DevicePilot has transformed from a brick to a fire-blanket for companies like POD Point who are scaling and need to monitor their connected services, and we continue to evolve thanks to all the feedback from our customers. This month we released two more powerful KPI features and also documented our simple-but-powerful Deep Linking capabilities to improve your service monitoring workflow.
Respected IoT analysts Beecham Research released a useful report in conjunction with several IoT players into why IoT projects fail. The findings that particularly resonated for me are:
I laughed when I saw that the percentage of "successful projects" according to technical people was about twice that of the same metric according to business people – a classic case of "it works...but isn't useful"?
You've doubtless heard the outrage that Sonos provoked when it announced it would no-longer support hardware that was on-sale as little as two years ago - and then had to somewhat retract that position. Clumsily-handled, but I could empathise as I went through similar anguish at my previous company AlertMe (eventually hardware just becomes too-old to support). But also perhaps a failure to think ahead, especially for expensive, substantial, well-made things like speakers (many of us have Hi Fi speakers we bought 30+ years ago which still work perfectly-well). Perhaps if they'd had a recurring revenue model they might have planned for a "smart hub, dumb speakers" architecture, which would have been more efficient to keep up-to-date.
Foundries.io are ex-Linaro folk who offer open-source, pre-built Linux images with a CI/CD framework for managing them in the field (e.g. upgrades). Reminiscent of Balena, except they are open-source, can't see your data, and claim much better security. This is exactly the kind of technology component that a hardware company should buy rather than try building themselves, because of opportunity cost and because there are so many ways to get this stuff wrong.
Other IoT news:
Finally - if you haven't already - please do complete our CMM model questionnaire to test your readiness to deliver at scale.
Until next month.