The practical realities of IoT
an interview with Dan Barthel
Tanja Bezjak of Infinum interviewed Next Mile’s Dan Barthel about the practical realities of enterprise IoT success in May 2023. We’ve cleaned up the spoken word salad and summarized the transcript below. While tech is known for its breathless hype, Dan stays focused on reality here.
How has the Internet of Things changed in the last 5 years?
IoT matured beyond its hype and has been settling into reality. Device- and platform-level solution providers are learning if, where, when, and how IoT can be genuinely valuable, aligning to always-shifting market needs.
What do you regard as the most interesting use of the Internet of Things?
Definitely operation-wide digital twins synthesizing physical and virtual data into a business governance system. These weave production equipment, facility, proof of work, product, and customer data together, culminating in total insight linking a business's customer to their purchased product to their product's production details. We've designed a few of these at Next Mile—they're painstaking but worth it when you see real results in the P&L.
What’s the biggest risk associated with the Internet of Things?
For most businesses, the biggest risks with IoT offerings remain weak return on investment and low end-user value. We founded Next Mile specifically to protect clients from both threats. Both problems manifest after the capital investment phase of a program, most frequently as operations that run a slight loss or minimal profit.
At the market level, any threat to cheap silicon takes new consumer-grade and widespread industrial IoT off the table, re-orienting IoT toward low volume, high value/high cost devices.
What are the security concerns typically associated with IoT?
The two obvious concerns:
- Device-targeted attacks, usually unauthorized remote access. This results in device takeover for botnets at best and physical world consequences at worst.
- Human-targeted attacks, usually capturing identification data or credit card information fraudulent use.
Two less-considered concerns:
- Broadened attack surface and a higher weak-link probability. Even simple IoT solutions require multiple applications, codebases, databases, environments, scaling systems, types of hardware, physical settings, and communication protocols. This goes doubly so for physical product companies unused to thinking of their products as attack vectors into their own business.
- Business continuity.. Inevitably, a critical component of a connected product's architecture either breaks or becomes untenably expensive, leaving the business, device fleet, and customers hostage to it. It's exceedingly rare that any IoT architecture has been properly layered for future interchangeability.
What one factor would most accelerate the benefits of the Internet of Things?
At Porsche Sport Driving School, they say "slow down so you can go faster", meant to help you quiet distracting thought and concentrate solely on your task. In the 2010s, decision makers were fed "speed-to-market" hype resulting (with some exceptions) in a landscape populated with marginally-useful, rushed-to-market curiosities that conditioned normal people to be either cynical or reluctantly hopeful about connected devices.
Instead, "value-to-market" must be the mantra—can you prove that your solution improves your audience's life?
What’s the one piece of advice for a business leader interested in the Internet of Things?
Be deliberate. Successful IoT is a piecewise game that requires you to think around the next corner. Unlike Agile software application development, connected product development doesn't pivot quickly or cheaply, making the market penalty for "wrong" much higher than the penalty for "slow".
For your customers, ask:
- Is your solution compelling enough to cause users to switch away from their current approach to the problem? Sometimes, the dumb solution is smarter than the smart one.
- Does your proposed business model align with the amount and method your customer might pay? We've seen an awful lot of subscription-like OpEx models die!
- Will your costs at scale (meaning product quantity, SKU count, and/or fleet years) really add up the way you think? Or might foreseeable diseconomies force a pricing or terms of service rethink in a way that disrupts customer expectations?
For your business, consider:
- How will I help teams or business units connect and experiment during engineering?
- How will I scale a successful product? How will I sunset a flop? It's rarely as simple as turning a production volume dial—you'll likely have to change the digital ecosystem that brought you to this point.
- How do I deliver continuous value to a potentially large fleet or installed base over several years?
Although they're physical, connected products are still digital products—their development doesn't end when they leave production.
From your experience, what inspires businesses to implement IoT solutions?
Monthly recurring revenue (MRR) is still en vogue, so businesses typically view IoT as a means to earn money beyond the one-time product sale. Trouble occurs when MRR becomes the sole motivator. Basic capitalism still applies, so your solution still has to be valuable to a customer before it earns you anything, and with an added monthly payment, that value provision must now become continuous.
The other cold reality is one aspiring IoT users should embrace: in IoT, most value comes from savings, not revenue. Financially, it's better this way, as a new saved dollar (whether from fixed or variable cost) has a greater impact on operating margin than a new revenue dollar. That may not be the business goal, though.
How do you see the role of AI in the IoT space?
Broadly-defined AI (to include ML) and IoT are symbiotic. Connected device sensors collect the data that can train a model, and AI helps an IoT device react and respond to changes in its specific environment based on the model trained by thousands of others.
What are the biggest challenges preventing companies from implementing IoT solutions?
It's been our experience that IoT solutions have three stages of difficulty, each exponentially harder than the previous:
- Getting connected, meaning acquiring data from a sensor, formatting it, and sending it to a remote database. There are a lot of solved problems here, so designers and engineers usually integrate off-the-shelf tech to achieve this at the PoC stage.
- Creating and maintaining a mechanism of value delivery. Whether it's an implementation of a pre-built tool, a custom mobile app, or a series of interfaces, everyone creates an experience and an ecosystem that requires customization. It's always more complex than anyone promises. Additionally, in most product companies, IoT is owned by the product expert who knows the most about digital. At best, that person understands a part of digital, not the whole picture, but the systemic nature of IoT invites the whole picture into their world!
- Fleet ownership. Now you've got dozens of SKUs, millions of connected devices, and plenty of customers who depend on your offering's reliable function. It's delicate and the stakes are high. Can you run an OTA without bricking critical gear? Does your QA process account for enough edge cases when angering 0.5% of your users would result in thousands of 1 star reviews? The multiplicative nature of fleet operations doesn't just make mistakes likelier, it increases their consequences.
Can you name some of the industries which have benefited the most from IoT?
Certainly older process controlled systems (SCADA, PLC, DCS used in factory automation) are the all-time winners. They've been valuable in manufacturing for decades!
Recently, wireless remote monitoring is the next most beneficial use case. Any user with assets somewhere other than where its administrator is can reduce waste through some combination of the classic use cases (reduced truck rolls, predictive maintenance, recordkeeping, etc.) Locally, they let integrators bypass hardwired retrofits (and often software integration) and still reap the benefits.
How do you think IoT will change shape in the years ahead?
We're exiting the hype era where cheap, ubiquitous connectivity was supposed to link everything together into techno-utopia.
In the near term, IoT will crawl toward solutions that create genuine utility, so you'll see fewer wandering, inscrutable IoT offerings and more short paths to value. I also expect that business models for connected offerings will become more humane, with lower variability and fewer fees.
What are some of the emerging trends in IoT that enterprises should be paying attention to?
In the field, you'll see more modular solutions that combine a standard, a protocol, and mutually-compatible computing and sensing hardware into packages integrators can assemble as needed, simplifying integration. Successfully retrofitting pre-existing capital will continue getting less far-fetched and less expensive.
In the office, digital twins are just getting started. Look for them to grow beyond equipment and facility monitoring and control, and ask yourself what a digital twin of your entire business might look like.
Above the fray, pay close attention to regulation. All it takes is another high-profile security breach or scandal that trods on normal peoples' rights or privacy to spur reactive, industry-shaking legislation. Companies with IoT offerings are every bit as exposed to force majeure from government action as they are to hackers.
What does it mean to be ahead of the curve in IoT technology utilization?
In IoT, “ahead of the curve” means:
- Actively-curated solution design: exactly the right teams, tools, and technology, in the right amount, doing the right thing at the right time.
- Antifragile systems that can withstand changes and minimize change cost.
- Organizational support for the full scope of IoT: early engineering, production and scaling, and fleet ownership.
You're behind the curve when:
- Disruptive surprises keep occuring.
- Results don't match your effort and expenses.
- You can't pinpoint why the connected version of your offering will deliver consistent, recurring value to its buyer and end-user.
Infinum and Next Mile both exist to keep our customers ahead of the curve.