OnDeck Deep Dive: Demystifying IoT, Navigating Blockchains, and Other Challenges in Commercializing Core Technologies
You can find the original post on OnDeck Daily here: http://www.ondeckdaily.com/post/475
Hi everyone - I'm Shilpi Kumar and currently lead product commercialization at Filament, a venture-backed company at the intersection of industrial IoT and Blockchain technology. Simply, we build hardware that brings long-range connectivity and economic capabilities to existing machines, using distributing networking and embedded contracts. We're currently focused on construction and energy applications. Before Filament, I was an analyst at First Round Capital and worked on early-stage hardware investments for VegasTechFund. Happy to talk about: blockchain hype, IoT business models, hardware development challenges, working with Corp VCs, industrial sales cycles, distributed teams, and moving from venture to a startup.
Q: Pretend you're talking to someone who is a new investor and you have just a few hours to teach them everything you know about UOT investing. Please give a comprehensive overview of the IOT landscape with your opinions on each sector/sub-sector. I know this is like 10 blog posts (read: books) in itself, but, hey, this is a deep dive! I know you take issue with the concept of an IOT landscape, as opposed to IOT in everything. I know you hate these questions. But still.
A: I'll start by saying that I still struggle with the fact that "Internet of Things" has a collective endorsement. Filtering for what fits into the landscape is a subjective matter. To me, it considers the physical digital bridge and the combination of technologies that are expanding the capabilities of computing and information networks beyond computers and smartphones. About five years ago, when I started working on investments in the hardware and crowdfunded product realm, Nest had made the case for the surge of consumer and home devices. The products were appealing to enthusiasts in specific activities (fitness, cooking, watching tv) and makers who now had access to the boards and sensors to easily build their own projects. While some of this was nothing new (I had a clap on / clap off nightlight growing up), much of the mobile internet infrastructure and component developments set up these gadgets with the functional ability to monitor, process, and share events at an affordable price. More importantly, there was an acknowledgment that passive and active collection of this user data from other surfaces could be sold on the backend to other buyers. While big and small companies are still pursuing these products, venture investments have shifted towards compounded effort in creating devices for B2B and industrial environments, with compounded opportunities to monetize the data that’s collected.
I would suggest a new investor in this space start by looking for the company’s existing strengths and competencies — do they have an expertise in a specific industry process? technology to collect different data? Data science and AI modeling that matches this pattern need? Depth of experience in radio communications? Then, look at the “problem” they are approaching. Not just the industry or market vertical, but the actual problem. What do they do now? How much do they pay for it? Is it dirty, dull or dangerous for people? Have they considered other M2M technologies in the past? Why didn’t they adopt them? Is there a regulatory fine for missing this collection? What department’s budget does it come from? How much do they pay for a wire drop or satellite connection? The company should have a deep understanding of the existing standard and proof, even if only with a paper model, that they could ease the customer's pains at a viable price point. While many IoT startups claim “full stack” solutions, there is an insane range of mastery required for any small team to develop and deploy a solution: industrial design, printed-circuit board (PCB) layouts, cryptographic key management, radio routing, power system design, manufacturing and supply chain considerations, distribution and installation UX, firmware reliability testing, certifications, dashboard and data presentation design, app development, firmware flashing and over-the-air (OTA) update support, OS and scripting environments, integrations with existing systems that with inevitably vary by custom, etc. Even if using off the shelf and existing stuff, it must outperform the default behaviors, and anything that GE, Microsoft, IBM can deliver, as they are also going all-in on IoT. I think it's becoming more realistic to determine what parts of the full stack solution a team can build a competitive edge around and partner strategically, while ensuring the value they uniquely deliver is enough of a margin on top of their costs and split with the other partners.
Sectors can be sliced a lot of ways. I’m most impressed with the teams who have done early business cases and large scale device deployments. Even better when their anchor customer(s) are expanding use of the system (devices or use cases) and offering strong referrals. I should mention that this is MUCH HARDER for industrial hardware companies to accomplish than I’ve seen for software. Proof of concepts and pilots require a ton of resources no matter how great the technology is. Some big areas I still see opportunity in:
* Precision agriculture and low-touch farming
* Shipping tracking and chain-of-custody reporting
* Automated energy balancing and battery storage systems
* Transactive machines and secure data sharing between companies
Q: Similarly, pretend you're talking to someone who is a new investor and you have just a few hours to teach them everything you know about crypto investing. Please give a comprehensive overview of the crypto landscape with your opinions on each sector/sub-sector. I know this is like 10 blog posts (read: books) in itself, but, hey, this is a deep dive!
A; With crypto, there is a new thing to decentralize and subsequence investment opportunities all the time so any advice I have is likely irrelevant. I can share my personal method for digging through the white papers and companies popping up. Very few of them have finished (or started) building a robust product or system, so assessing the technical viability of the idea getting off the page is part of the challenge. I similarly try to figure out the friction of getting the idea to displace or pose a robust alternative to the centralized system(s) in place, and look for evidence that the team is paying heed to those efforts. I lean heavily on the cryptography and distributed networking experts on Filament’s team and suggest everyone find similar counsel. All this said, to make money on crypto right now is not necessarily about forecasting the value and more about predicting how the market of block-heads will lurch towards it or attack it on Reddit. I was in denial of Ethereum for 2+ years (and still have my concerns about their claims!), but ignoring the hype doesn’t make it inconsequential. Despite the conferences and corporate endorsements, it’s definitely still wild west so hard to know what’s going to go up in smoke or fall victim to a slackbot hack.
Q: With all the hype around ICOs, which ones as of recent do you believe are solid projects with high potential? What blockchain innovations do you think should be created?
A: The nature of tokenization is actually a fascinating and sensical model for building marketplaces and facilitating seamless value exchange. The way that ICOs have gained momentum, the eventual value of the token seems less consequential than its volatility in the crypto trading markets. Grid+ and Blockstack’s tokens are both interesting to me. Grid+ building a peer-to-peer energy economy as an alternative to an entirely disjointed and problematic system for transmitting and retailing energy in the US, cutting out middlemen and companies that existed mostly to move money around. Blockstack is offering a way for internet users to engage their data with applications without trusting any authorities or presiding companies to take advantage of them. For any of these ICOs to be truly decentralized, the organization cannot be the sole giver and arbitrary of tokens.
I don’t think a blockchain will create any solutions on its own, but rather feed into the information innovations and secure communications protocols that open up what type of interactions between people and machines are possible. Frankly, I think micro-transactions and smart contracts are much more interesting if they are able to execute on machine-time-scale and machine-decision-making, rather than replacing the transactions people are familiar with. I see a need for a trustable system for device identity and dynamic engagement between machines of different owners, that works in the field where there is no power and low compute available. Fortunately, this is exactly what Filament is building 😉
Q; What did your transition from VC to startup look like? What things from First Round do you think help/hinder you at Filament?
A: I loved working on the investment side, but would often get really deep into research or diligence in a particular area and struggle to move away from it. I was working researching Filament for their Series A round and got hooked by the vision and elegance of their vision. First Round didn’t invest, but I managed to talk my way onto the team. I’ve definitely struggled with the focus and patience required to get hardware deployed at scale. Investing experience has kept me skeptical; careful to approach new markets and relentless at pushing customers to articulate their actual interest in our technology. I still wish we could magically ship our entire roadmap today, but now know intimately that it's much harder to ship things than simply assess the viability of them.
Q: There are always security concerns with IoT, especially “always-on” sensors and recorders; what aspects of these concerns are unfounded consumer panic, and which are real issues that companies are having a hard time solving?
A: Great question. IoT and security headlines are often referring to lapses in both security and privacy. Security of IoT systems must be proactively accounted for at the hardware, network, gateway, cloud storage, and application layers, as well as all other exposure during the lifetime of the device. I’ve seen enterprises becoming much more opinionated about the level of security precautions and establishing best practices after the Jeep / Target / Equifax hacks. The “always-on” thing for consumer and enterprise devices is more of a privacy tradeoff to me. Are you willing to trade information about your activity, floorplan, crop yield, for better recommendations and insights? Are you okay with it knowing that they will continue to share and monetize the information without giving you any cut? Generally, consumers don’t seem to care; or know how to care until something viscerally threatens them. Filament wrote a paper on IoT Security that gives a deeper dive into my perspective on the matter: https://filament.com/assets/downloads/Filament%20Security.pdf
Q: I’m curious to hear more about distributed teams! Whenever I’ve worked fully remote, it’s been a mixed bag; between working with different time zones, feeling the need to schedule video call meetings that end up wasting time, or having to be “always on” with neverending slack messages, it’s hard to strike a good communication balance. How do you approach this?
A: Filament’s team is more “decentralized” than fully distributed, so our main challenge is making sure we stay synced up across functional departments. We have a slack brainstem for ongoing conversations but also are intentional about creating artifacts from meetings and decisions for longer-term planning. There is a constant rebalancing, but being spread out helps us to use meetings when they are absolutely necessary and make sure all relevant details are captured. The key for me is to never assume that someone knows what I’m thinking. It helps that we all get together in person every quarter to smooth out anything that's welled up. We all have budget to jump on a plane whenever we need to work through something in person, which helps a lot!
Q: How has starting in VC/investing helped you be a better startup builder?
A: I think I answered this in response to Dwight's question. In short - quick learning, fearless curiosity, and shameless questioning gave me a lot of the confidence I needed to understand the industries we're working in.