7 Themes — May 2018
As I think about what I’ve been interested in within technology the past few months and what I want to keep exploring, 7 themes have come across on a consistent basis. I thought I’d share to get some feedback to get more conversations going and see how I can collaborate with folks.
Better Data, Better Underwriting
Banking, lending, insurance, and various investment practices still rely on faulty data collection models/guidelines and flawed human intuition to underwrite financial risk. For example, a small commercial bank may keep a lawyer on its advisory board to help underwrite loans to law offices based on industry knowledge. But with various ways of collecting data and building learning models, relying on humans is highly inefficient. Additionally, if a financial institution can plug in to data sources through various APIs, they can continuously build out models for lending and insurance instruments over time.
Petal intrigues me, which creates its own credit model based on the ins and outs of a customer’s bank account (most likely through Plaid) instead of depending on FICO scores. This space probably needs one to pay more attention to the regulatory landscape, but I’m excited to see what people can build.
Coordination and Collaboration
Paradigms like version control and distributed manufacturing have the ability to coordinate complex development processes in a streamlined and efficient manner. Software development is easier with GitHub, but what would a version control system for the production of hardware or durable goods look like? How could the Opendesk model be applied to the construction industry?
Seeing how software can better coordinate various production processes will not only make the operations of current businesses more efficient, but open up opportunities for startups to challenge incumbents.
Service Economy Augmentation and Outcomes Data
Service businesses are highly dependent on human capital and expertise to ensure the viability of their bottom line. Examples of these businesses range from investment banking and real estate asset management on the enterprise side to financial planning and complex care concierges on the consumer end. The human component matters here, and I’m not sure sectors like investment banking will be completely automated in the near future (although people are working on it!), but I do think there are valuable ways to combine human capabilities with outcomes data through various mechanisms.
For example, Grove connects individuals with a trusted financial planner and a dashboard to manage their finances .
Grove arms human planners with internal data moats (decades of research = outcomes data), analytics tools, and dashboards to allow their services to scale across customers and make the whole process more efficient. I’d be interested to see what other sectors this model could make an impact in.
D2C Healthcare
There’s been a proliferation of direct-to-consumer product companies across household items like razors, kitchenware, bicycles, etc. These companies typically start by selling directly online, using advertising channels like Instagram (Attack of the Micro-Brands — Scott Belsky), and offering methods of experiential commerce (Glossier Pop Ups, Stay Floyd). I’m curious to see how healthcare products and services can be deliver through similar channels.
Hims and Simple Contacts are examples of companies I’m talking about. Building a brand around generic medication and undercutting on price, finding healthcare products and services consumers are already paying out of pocket for, or searching for things that are easily reimbursed through insurance, HSAs/FSAs, Medicare, etc. — these are all ways one can find opportunities to build a D2C healthcare platform.
Product Transparency and Breakdowns
This idea is a little more abstract, but I love what Everlane has done with the showcasing of their factories.
It’d be great to see how other companies could be radically transparent with the products and services they’re building. Details you could include are the cost to develop, insights into the product development process, and where things are sourced from. This is probably more applicable for physical goods over software services.
I wonder if there have been any studies within psychology or behavioral economics to show how consumers react when they know more about what they’re purchasing besides the black box of the price tag.
Real World Operating Systems
Managed by Q is an awesome business. They’re building an operating system for a real-world environment by making it easy to manage vendors, process payments, and coordinate services.
However, offices aren’t the only environments that deal with a plethora of local service providers/vendors. Other examples include factories, construction sites, schools, retail buildings, and assisted living facilities. I believe the Q model could expand and shape these other physical spaces as well.
Democratized Access to Machine Learning
What are the ways we can open up opportunities for millions of people to access the power of machine learning, arguably the technology that will have the most impact on human society for the decades to to come?
On the prototyping side, Lobe makes it easy for people to build machine learning models through a visual interface. They launched recently to much fanfare.
Another exciting company in this space is Camelot, which helps people build skills in machine learning and data science, and then helps them land jobs in those fields. These types of companies have enormous potential to become large enterprises in the coming years.
If any of these areas excited you or you're working on a solution in one in of these spaces, feel free to get in touch with me at aashaysanghvi[at]gmail.com. 
