data-supply-chain

Google Data Supply Chain: AI Supply Chains In A Nutshell

A classic supply chain moves from upstream to downstream, where the raw material is transformed into products, moved through logistics and distributed to final customers. A data supply chain moves in the opposite direction. The raw data is “sourced” from the customer/user. As it moves downstream, it gets processed and refined by proprietary algorithms and stored in data centers.

Traditional supply chains

supply-chain
The supply chain is the set of steps between the sourcing, manufacturing, distribution of a product up to the steps it takes to reach the final customer. It’s the set of step it takes to bring a product from raw material (for physical products) to final customers and how companies manage those processes.

In a traditional, and physical supply chain, we move from upstream to downstream, as we go from raw materials to the finished good, distributed in the hands of customers and consumers.

While the physical supply chain is structured by starting from sourcing, manufacturing and logistics. Distribution (intended as the finished product ready to be sold) comes at the end of this process.

There is another kind of supply chain, which for tech and AI companies is integrated within their business models, that is the data supply chain.

From physical to data supply chains

Let’s take a simple example. When Google manufactures its phone, the Pixel, while the phone is designed by Google, its manufacturing is outsourced in China (even though Google is moving it to Vietnam).

This means that up to sourcing and manufacturing, the process is outsourced, where Google instead takes control of the process as the phones produced are shipped to Google and ready to be distributed, either through its online stores, through carriers or other indirect channels.

So the whole physical supply chain might look like the following:

Yet, if we flip the perspective and look at the data side of Google, as an AI company, the whole “data supply chain” changes.

Data supply chains: flipping upside down the physical supply chain

data-supply-chain
In a data supply chain the closer the data to the customer the more we’re moving downstream. For instance, when Google produced its own physical devices. While it moved upstream the physical supply chain (it became a manufacturer) it moved downstream the data supply chain as it got closer to consumers using those devices, so it could gather data directly from the market, without intermediaries.

Where the Pixel phone moves from upstream to downstream, as we saw, it follows a classic supply chain path.

However, once the device is in the hands of customers/consumers/users, they suddenly become the sources of data, and the whole supply chain flips upside down.

A few things to notice here:

  • In the Google’s Pixel case, the whole data supply chain is controlled by Google.
  • Customers become also the sources of raw data.
  • As the raw data move downstream it gets refined by Google’s algorithms and it gets used for several purposes (from products’ personalization, to monetization of its assets through advertising).
  • That data gets stored in the Google data centers
  • And the Google data centers will need to source renewable energies and materials to maintain, and run its facilities, that keep the whole infrastructure going.

From traditional supply chain to AI supply chain

  • AI supply chains start with the sourcing of data. This flips them upside down.
  • Where a traditional supply chain would start with sourcing and manufacturing with a top-down approach, an AI supply chain starts bottom-up.
  • The source of raw data is the customer/user, as the raw data moves downstream the supply chain, it gets processed, refined and stored.

Read: Google Business Model

Read next:

Connected Business Phenomena

Vertical vs. Horizontal Integration

horizontal-vs-vertical-integration
Horizontal integration refers to the process of increasing market shares or expanding by integrating at the same level of the supply chain, and within the same industry. Vertical integration happens when a company takes control of more parts of the supply chain, thus covering more parts of it.

Supply Chain

data-supply-chain
A classic supply chain moves from upstream to downstream, where the raw material is transformed into products, moved through logistics and distributed to final customers. A data supply chain moves in the opposite direction. The raw data is “sourced” from the customer/user. As it moves downstream, it gets processed and refined by proprietary algorithms and stored in data centers.

AI Supply Chain

ai-supply-chains
An AI supply chain starts with the sourcing of data, which is produced by consumers. As this data gets stored on hardware, it goes through a first refinement process via software. Then it’s further refined, and repackaged by algorithms, and stored in data centers, which work as the fulfillment centers.

Backward Chaining

backward-chaining
Backward chaining, also called backward integration, describes a process where a company expands to fulfill roles previously held by other businesses further up the supply chain. It is a form of vertical integration where a company owns or controls its suppliers, distributors, or retail locations.

Decoupling

decoupling
According to the book, Unlocking The Value Chain, Harvard professor Thales Teixeira identified three waves of disruption (unbundling, disintermediation, and decoupling). Decoupling is the third wave (2006-still ongoing) where companies break apart the customer value chain to deliver part of the value, without bearing the costs to sustain the whole value chain.

Entry Strategies

entry-strategies-startups
When entering the market, as a startup you can use different approaches. Some of them can be based on the product, distribution, or value. A product approach takes existing alternatives and it offers only the most valuable part of that product. A distribution approach cuts out intermediaries from the market. A value approach offers only the most valuable part of the experience.

Disintermediation

disintermediation
Disintermediation is the process in which intermediaries are removed from the supply chain, so that the middlemen who get cut out, make the market overall more accessible and transparent to the final customers. Therefore, in theory, the supply chain gets more efficient and, all in all, can produce products that customers want.

Reintermediation

reintermediation
Reintermediation consists in the process of introducing again an intermediary that had previously been cut out from the supply chain. Or perhaps by creating a new intermediary that once didn’t exist. Usually, as a market is redefined, old players get cut out, and new players within the supply chain are born as a result.

Scientific Management

scientific-management
Scientific Management Theory was created by Frederick Winslow Taylor in 1911 as a means of encouraging industrial companies to switch to mass production. With a background in mechanical engineering, he applied engineering principles to workplace productivity on the factory floor. Scientific Management Theory seeks to find the most efficient way of performing a job in the workplace.

Poka-Yoke

poka-yoke
Poka-yoke is a Japanese quality control technique developed by former Toyota engineer Shigeo Shingo. Translated as “mistake-proofing”, poka-yoke aims to prevent defects in the manufacturing process that are the result of human error. Poka-yoke is a lean manufacturing technique that ensures that the right conditions exist before a step in the process is executed. This makes it a preventative form of quality control since errors are detected and then rectified before they occur.

Gemba Walk

gemba-walk
A Gemba Walk is a fundamental component of lean management. It describes the personal observation of work to learn more about it. Gemba is a Japanese word that loosely translates as “the real place”, or in business, “the place where value is created”. The Gemba Walk as a concept was created by Taiichi Ohno, the father of the Toyota Production System of lean manufacturing. Ohno wanted to encourage management executives to leave their offices and see where the real work happened. This, he hoped, would build relationships between employees with vastly different skillsets and build trust.

Dual Track Agile

dual-track-agile
Product discovery is a critical part of agile methodologies, as its aim is to ensure that products customers love are built. Product discovery involves learning through a raft of methods, including design thinking, lean start-up, and A/B testing to name a few. Dual Track Agile is an agile methodology containing two separate tracks: the “discovery” track and the “delivery” track.

Scaled Agile

scaled-agile-lean-development
Scaled Agile Lean Development (ScALeD) helps businesses discover a balanced approach to agile transition and scaling questions. The ScALed approach helps businesses successfully respond to change. Inspired by a combination of lean and agile values, ScALed is practitioner-based and can be completed through various agile frameworks and practices.

Kanban Framework

kanban
Kanban is a lean manufacturing framework first developed by Toyota in the late 1940s. The Kanban framework is a means of visualizing work as it moves through identifying potential bottlenecks. It does that through a process called just-in-time (JIT) manufacturing to optimize engineering processes, speed up manufacturing products, and improve the go-to-market strategy.

Toyota Production System

toyota-production-system
The Toyota Production System (TPS) is an early form of lean manufacturing created by auto-manufacturer Toyota. Created by the Toyota Motor Corporation in the 1940s and 50s, the Toyota Production System seeks to manufacture vehicles ordered by customers most quickly and efficiently possible.

Six Sigma

six-sigma
Six Sigma is a data-driven approach and methodology for eliminating errors or defects in a product, service, or process. Six Sigma was developed by Motorola as a management approach based on quality fundamentals in the early 1980s. A decade later, it was popularized by General Electric who estimated that the methodology saved them $12 billion in the first five years of operation.
Scroll to Top
FourWeekMBA