The Double Moat Strategy

Featured Photo by Zoe Jane on Pexels‍

In an era where numerous artificial intelligence (AI) companies and tools are emerging which claim to be powered by GPT-3.5/4 or similar technologies, it is the exceptional few that truly understand the nuances of their domain. To be successful, companies must distinguish themselves with laser focus on addressing specific problems and domains — this is what we refer to as a ‘Double Moat’ Strategy.

Let me start by explaining the concept of a ‘moat’. A moat, in the context of this article, refers to the initial layer of protectable competitive differentiation that a company builds, often through smart models and data collection mechanisms. This is not unique to AI technology, but is particularly important in this realm. A moat ensures a continuous flow of unique valuable data, optimizing proprietary AI models and enhancing user experiences.

The second moat, and the essence of the Double Moat Strategy, stems from a company’s deep understanding of users and their pain points. By addressing real pain points rather than offering mere superficial solutions, companies fortify their competitive position. They own a uniquely deep connection to the users in their market and their workflow challenges. By combining both moats, a company establishes a robust and differentiated position in the market, making it difficult for competitors to replicate their success and further enhancing their overall competitive advantage.

Companies that immerse themselves in user research, empathize with their target audience, and develop an in-depth understanding of their pain points will have a solid foundation for success.

Building a Moat through Smart Data Collection

An AI-based startup’s moat should not solely rely on inventing new special neural network architectures. Instead, a robust foundation can be established by integrating a smart data collection mechanism into the product from day one.

This approach ensures a continuous flow of relevant and valuable data, enabling the refinement of AI models and the delivery of enhanced user experiences. Moreover, by anonymizing collected data and prioritizing user privacy, trust is built, further strengthening the moat versus competition.

The true power lies in combining both a moat that incorporates smart and secure proprietary technology as well as a moat around an amazing end-to-end product that directly targets user pain points.

Understanding Users and Pain Points

Beyond unique technology and data collection, a moat can be fortified by deeply understanding users and their needs. Especially in this climate of constrained resources, a product must be a painkiller, addressing real acute challenges rather than being perceived as a mere nice-to-have (vitamin). Companies that immerse themselves in user research, empathize with their target audience, and develop an in-depth understanding of their pain points will have a solid foundation for success.

Combining Moats for Success

By integrating intelligent data collection mechanisms and leveraging their profound understanding of user pain points, companies can create a compelling solution that meets the needs of its target audience.

The true power lies in combining both a moat that incorporates smart and secure proprietary technology as well as a moat around an amazing end-to-end product that directly targets user pain points. This Double Moat Strategy is where companies shine and truly establish barriers against new entrants.

A Few Examples of Double Moats

While many companies focus on developing cutting-edge neural network architectures or new data collection techniques, I often observe a lack of deep understanding of users and their pain points. That said, there are several companies out there doing a great job developing their defensible Double Moats. Below are a couple examples.

When it comes to companies that have successfully implemented a Double Moat Strategy, Atlassian stands out as a prime example (full disclosure: I work for Atlassian). The first moat for Atlassian lies in their robust suite of products, which includes Jira, Confluence, and Trello, among others. These tools are not just standalone applications; they seamlessly integrate with each other, creating a powerful end-to-end solution for teams and organizations. The workflow and interoperability between Atlassian products make it difficult for users to switch to alternative solutions, solidifying Atlassian’s market position.

The second moat for Atlassian stems from their customer-centric approach and deep understanding of user pain points. Atlassian has established a strong relationship with their user base by continuously listening to feedback and actively incorporating it into product development. By addressing real pain points and continuously improving user experiences, Atlassian has cultivated a loyal customer following. This deep connection with users forms a second layer of defense, making it challenging for competitors to entice Atlassian’s base away.

Another example of an effective Double Moat strategy is the user research focused startup, CoNote (full disclosure: I am an advisor and equity holder). As a first priority, their tech is a proprietary AI engine, which combines several models in an ensemble strategy. While they do leverage some APIs, it would be hard to replicate the functioning of the entire engine. This is their first moat.

Beyond the tech, they have an unwavering focus on addressing user pain. The team comprises experts not just in data science, but in design and product. In fact, the founders were inspired to build the solution because they needed this type of tool for their own user research. This firsthand experience, and deep connection to users and customers in the space, enables them to create a product that addresses a real problem. Their tech is the star only insofar as it is required to solve the user pain point.

Pursuing a Double Moat Strategy can enable AI and other tech-forward companies to defend their positions long enough to make lasting impacts on the market. All the new developments in AI are cool. But I believe they are only useful to the extent they are implemented in ways that acknowledge how humans need to use them.

‍