Nanosense engr 245 lean launchpad stanford 2019

Editor's Notes

• #4: As you saw in our video intro, we designed and prototyped custom chips that would enable edge sensors in low-power. We saw all kinds of benefits from these devices (click) so we assumed this was a good enough business model (click through) and that we would have no problems commercializing this technology
• #5: As an initial sample application we targeted smart home devices (click) that need always-on sensing You might want these devices to be battery operated because you don’t always have a plug conventinet to where you want to place them. In that case, the battery would need to last for a long time because users would not want to replace the batteries or re-charge their devices regularly So we concluded that our product could be a wake-up mechanism for these products
• #6: These smart home devices, however, have many other functionalities. They are complex systems and our product (click) would be a part of this system
• #8: So we built our initial business model canvas based on these observations and made some guesses on who our initial customer segments might be
• #9: Here is a simplified version if it. We hypothesized that our customers would be system companies like Amazon, Google, or Apple. Warehouse automation players and larger semiconductor manufacturers who might be interested in licensing our IP to add our functionality to their larger System-on-chips (or SoC)
• #10: During an office hour with Steve Blank we were encouraged to think big and go broad for a moment to make sure we wouldn’t miss an important market segment. We laid out key vertical markets that came up during our brainstorming sessions.
• #11: First, we consider portable electronics like smartphones and wearables. They surely require low-power and low-latency in their always-on sensing modes. However, we found out that for a small semiconductor startup like ourselves in these segment we would have to compete with many established and low-cost players in a crowded space. (click) In addition, we realized that we would have to rely on a distribution channel that was very cost-structured and would therefore require that we ship high-volumes at very low margins. This seems unfeasible for a small chip startup and was ultimately (click) an unsatisfying value-add for our team. (click) For smart cities we found that city departments use a similar OEM model by purchasing from large integration vendors.
• #12: We also learned that the city decision-makers are not as tech-savvy as in other industries and sales cycles are slow and depend on long-term relationships with people like the chief of police and the office of the city mayor
• #13: In the drone and autonomous driving space we found out our solution wasn’t particularly well suited because the sensing system was not a significant portion of the energy or complexity bottleneck
• #14: In the Smart Home application space we had a mixed bag. The market segment is very concentrated and tough to break into. We would need a partnership with a large and established player (like Nest or Honeywell) We did find promising partners willing to continue working with us to determine the value of our products. We also found a key learning (click): that real value-add and therefore margins would not be as high unless we were able to provide a solution that was more full-stack This brought us to our first crisis on our journey.
• #15: We were having a hard time getting enough interviews with relevant players and were debating about where to focus our efforts. We lost some hope that we had a compelling product. We could also see the stark difference between business and engineering and discovered we had different goals for the outcome of this project and the class. As these differences became clear (click) we had a team member decide to withdraw from the team.
• #16: From this experience we clearly saw the importance of having the right team dynamics for a start-up to be successful. We worked to… This ultimately motivated our estranged classmate to return to the team since we had more clarity on roles and responsibilities and what each of us were able to contribute to the team
• #17: The second important crisis we hit was to define our value-add. George John gave us sage advice after one of our team presentations (click), we needed a hammer From our interviews until this point, we were getting a lot of responses that amounted to “you have some really compelling tech but we don’t have an immediate use for it”
• #18: As we examined our initial assumption about our value add we felt lost in a complex system with many moving pieces
• #19: We struggled to see the real value of our product
• #20: This crisis led us to our next learning: to continue grinding. We had some idea of what we had learned to far. Here is what our updated value proposition looked like Also, in order to keep our product as full-stack as possible without undertaking a major engineering and logistic effort (click) we would look for problems that required simple solutions
• #21: This is when we came to an interview with Robert Locke, one of the class mentors. He told us about how one of the biggest payouts of large insurance policies is water damage from failed pipes. He also told us that the previous attempts at solving this issue hadn’t worked out because they were either complex and cumbersome to install or too simplistic and not accurate enough. It hit us that this particular application (click) checked off all of the boxes that we were looking for… (click through) This piqued our interest and we continued to brainstorm (click) about other applications in similar segments that would be addressed by our solution.
• #22: The other hammer to hit us in the face was at the ISSCC conference we attended. Chris gave a great presentation on his smart imager research and later that night we attended an industry showcase event where several companies demonstrated custom silicon to enable interesting applications. (click) This is where we met the PsiKick team. They are a provider of steam trap monitors that work without any batteries because they harvest the energy from the heat of the pipes to power their ultra low energy temperature monitors. Their business model is to essentially give away the monitoring hardware which is relatively low cost and instead charge for a monthly subscription fee
• #23: This model is working well for them and co-founder Ben Calhoun told us that if all they did was sell steam trap monitoring they could run a healthy business
• #24: This led us to an important conclusion about the value of disruptive sensing We realized we can leverage the data that sensors collect to provide actionable value to our customers that they care about
• #25: We quickly iterated our MVP from a standalone chip to check that our algorithm could detect the sound of running water by writing a quick app.
• #26: By re-examining the vertical segments we were previously addressing, we could see which of them would fit into this new model (click) and saw that with a small tweak to the smart home segment we could (click) group the three segments on the right into a new vertical that can be called...
• #27: … for the commercial and industrial segments that is uniquely enabled due to our innovative sensors
• #28: So how are we different than the rest of the market? We provide real-time monitoring in otherwise data-rich sensing modalities such as water leakage, alarm monitoring, and facility occupancy. We can provide this where others cannot (click) because we have re-designed the traditional sensors so that they don’t rely on heavy computation engines to make detections Because the installation is easy (click) and the maintenance is very low, we can cover hard-to-reach application that may not have access to wired power Finally, by having visibility into the signal of interest we provide relevant analytics (click) to our customers so they can make real-time decisions (click) based on the monitor signals.
• #29: In addition, we can re-think our customer funnel because the subscription model allows for application-specific hardware (click) so the specialization cost can be amortized over the lifetime of the subscription. (click) We can also improve our customer retention and create an up-sell channel by providing new analytics based on the data generated by our installed sensor modules
• #30: This is our final value proposition box in the business model canvas (click) and here are a few interview quotes that support these value props
• #31: And this is our updated full canvas along with the boxes that changed significantly
• #32: Moving forward Chris and Dan plan to continue working on this venture to further explore the demand behind the updated market segment. Roy and Mindy will do an internship in the summer and keep a close eye on the project progress in case there is an intersection of needs upon graduation. Sophia is planning to take on a project management role at Amazon. We would like to thank our amazing mentors Radhika and Rekha who believed in us and gave us advice and encouragement with plenty of war stories from their time as engineers in the Valley
• #35: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #36: Solves latency and memory problems
• #37: Solves latency and memory problems
• #38: Those devices do not have room for sophisticated heavy processors, GPUs They don’t have the battery to do so
• #39: Those devices do not have room for sophisticated heavy processors, GPUs This was validated in the PhD work of my teammates Chris and Dan
• #40: Those devices do not have room for sophisticated heavy processors, GPUs This was validated in the PhD work of my teammates Chris and Dan
• #41: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #42: OEMs, Licensing, Defense, Safety
• #43: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #44: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #45: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #46: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #47: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #48: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #49: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #50: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #51: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #52: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #53: Describe what your team will do tomorrow and this week Include anything else you feel is pertinent towards understanding the direction your team is heading
• #55: We predict the selfie of the future would look like this, using the imager of nanosense. Dan and Chris at the Booth.
• #56: MVP is required this week. What did you change? What was it testing? What did you learn?
• #57: We optimize the complete smart sensor processing chain, not by accelerating the compute, but by changing the system to sense "smart features" - this requires cohesive algorithm/hardware co-design.
• #58: Our sensing methodology is to perform feature extraction in the analog domain before the digitization to allow power savings due to lower sampling and computation requirements
• #59: Our sensing methodology is to perform feature extraction in the analog domain before the digitization to allow power savings due to lower sampling and computation requirements
• #60: Those devices do not have room for sophisticated heavy processors, GPUs They don’t have the battery to do so
• #61: Our sensing methodology is to perform feature extraction in the analog domain before the digitization to allow power savings due to lower sampling and computation requirements
• #62: We optimize the complete smart sensor processing chain, not by accelerating the compute, but by changing the system to sense "smart features" - this requires cohesive algorithm/hardware co-design.
• #63: The gap between what we were doing and what other groups in our class were doing.
• #64: The gap between what we were doing and what other groups in our class were doing.
• #65: We decided to pivot
• #66: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #68: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #69: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #70: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #71: *Update weekly For [customer segment] who [key need or opportunity], [company name] is a [type of product or service] that [key value proposition that solves key need or opportunity]. Unlike our competitors, we [differentiation].
• #72: More lucrative to focus on insurance and industrial IoT application to have a recurrent revenue model with a very self-contained technological solution that is managemeble for a startup, instead of selling 1-off chip to a startup. *Update weekly Each customer segment needs a matching value prop. Use a different color for each customer segment. Order of Validation: 1. Customer Segments 2. Value Propositions 3. Channels 4. Customer Relationships 5. Revenue Streams 6. Key Activities 7. Key Resources 8. Key Partners 9. Cost Structure