Assembling Automation

From data to subsystems and, ultimately, platforms

I don’t have a crystal ball to say when autonomous vehicles, UAVs, delivery robots, or human-lite warehouses will be commonplace. I do have an understanding, though, as to how we might see the buildout that will get us there. It’s an assembled solution that brings together the efforts of many. The most successful companies will extract value from data, piece solutions together into subsystems, and eventually aggregate subsystems into a platform.

Data is the Atomic Unit

Consider tasks that are currently performed by humans. A human worker will account for 100s of inputs, explicitly and implicitly when completing a given task. For example, truck drivers account for the size of their vehicle, tonnage, surface conditions, inclines/declines, weather, pick up/drop off points, traffic, dock conditions, and so much more. And this is just the tip of the iceberg given that the act of driving itself is one that requires constant observation, attention/alertness, and processing to stay safe, avoid delays, and capitalize on opportunities for improved service.

In an autonomous vehicle, much of this data needs to be captured and explicitly included. The raw data however is just the atomic unit — important, but needs to be applied constructively in order to unlock its value. In other words, data needs to be productized. Think about geospatial data. Static coordinates and location descriptors alone aren’t particularly useful, but applied in such a manner as Google Maps, it becomes invaluable. Another example, real-time weather (like with DarkSky), can greatly inform our decision making when it comes to travel, what to wear, and how to equip vehicles. The most successful products are able to:

• Clearly improve outcomes/decision making. For example: improving on-time delivery by 10%+ or optimizing travel time by rerouting people or packages
• Inherently improve the accuracy of information through use. There are currently products in use today that capture information regarding road conditions, wait times at docks, and live weather data.
• Leverage updated information or data in a timely fashion. This might include traffic data surrounding a vehicle, driver behavior or even congestion at port.

Over the years, we continue to see products and services both capturing and tapping into existing data and information. These products and services are ultimately brought together as key cogs in various subsystems.

Piece Together Subsystems

In an effort not to swallow the elephant whole, things are generally engineered in components. It might be a crude way to think of it, but a subsystem is essentially the specific combination of various products/services to achieve a specific goal.

In terms of autonomous vehicles, subsystems include an electric transmission, anti-collision systems, connectivity services and more. They generally require various inputs and combine several smaller products and services to achieve a goal. Examples that come to mind — interfacing with the cloud, driving a car forward, and preventing a car from hitting other objects.

A subsystem tends to be complex in nature and, as such, generally subscribes to many products/services or aggregates value from multiple companies. For example, anti-collision systems pull together location, lidar, computer vision, weather data, vehicular inputs and more to achieve a specific outcome.

You might be able to tell from the aforementioned example that these are some of the major operations in getting a car or truck moving on it’s own — where the car, truck (or warehouse) are essentially platforms.

Aggregate Subsystems into Platforms

Platforms are the holy grail for founders and venture investors — value tends to aggregate to them. They generally foster an environment that allows others to build on their efforts and address various problems and opportunities. The car or truck is viewed as the next mobile platform and dollars are flowing to support that thought.

Departing from the vehicular references, let’s view the warehouse as a platform. At a high level, such an example might comprise of certain key subsystems such as receiving, storage, inventory, security, and outbound distribution, where the value of a company corresponds directly with how much of a subsystem they own. The key here is that there is compatibility among subsystems as the warehouse or physical platform is digitized. For example, a receiving solution from Company A might employ robots that wait at a dock, unload trucks, and scan received goods. These goods are then handed off to another solution from Company B that routes the items to the proper location in the warehouse facility.

Value exists across the assembly

Observations would lead me to believe that the most successful products are able to shift into subsystems and ultimately platforms.

There are a variety of pockets where value might emerge. By understanding our desired outcomes in an automated supply chain, founders, customers, and investors can create and derive value — whether it’s at the data, subsystem, or platform level. Stay tuned for part two that fleshes these potential opportunities out further.

How do you anticipate the shift towards an autonomous supply chain to look? Share your thoughts and questions in the comments, and give this post a ‘like’ so more readers can be a part of the conversation.