AI: Startup Vs Incumbent Value and The Power User Curve: Understanding Engaged Users

AI: Startup Vs Incumbent Value and The Power User Curve: Understanding Engaged Users

In the world of technology and innovation, the distribution of value has always been an interesting phenomenon. When it comes to AI, there have been notable differences in how startups and incumbents have captured value.

Looking back at the first wave of the internet, we saw that most of the value went to startups such as Google, Amazon, Paypal, Ebay, Salesforce, Facebook, and Netflix. However, incumbents like Microsoft, Apple, IBM, Oracle, and Adobe also managed to extend their franchises onto the internet and capture a significant portion of the value. This resulted in a relatively balanced split of value between startups and incumbents, perhaps around 60:40 or 70:30.

When it came to mobile, the story was different. The majority of the value went to incumbents like Apple and Google, with every mobile version of an incumbent's app gaining significant traction. However, startups like Whatsapp, Uber, Doordash, Instagram, and Instacart also managed to capture a notable share of the value. This time, the split leaned more towards incumbents, with a ratio of around 20:80.

Interestingly, in the world of crypto, startups have dominated the capture of value. Companies like Bitcoin, Ethereum, Coinbase, Binance, and FTX have seen tremendous success, while existing financial services and infrastructure companies have had limited participation in value creation.

To beat an incumbent as a startup in the AI space, you need to either build something dramatically better that overcomes the incumbent's advantages, or focus on a brand new customer segment or distribution moat that the incumbent cannot serve. In general, a 10X better product is required to disrupt the market.

One possible reason why incumbents have historically won in the AI space is their data advantage. However, this advantage might be diminishing as companies leverage the broader internet as an initial training set and shift towards models that work effectively with smaller data sets. This creates an opportunity for startups to level the playing field and capture a larger share of the value generated by AI.

The current wave of AI feels different from previous ones for several reasons. The speed of innovation across various areas is remarkable, making it easier to create products that are 10X better than existing solutions. While GPT-3, a breakthrough AI model, has not yet led to the emergence of startups building big businesses on it, a model that is 5-10X better could pave the way for a new startup ecosystem while augmenting incumbent products.

Unlike previous AI waves, there are now infrastructure-centric companies that have gained widespread adoption and rapidly growing usage. These companies, including OpenAI, Stability.AI, Hugging Face, Weights and Biases, and others, provide startups with access to essential technologies and create more opportunities for innovation.

Additionally, there are specific use cases where AI can make a significant impact. Highly repetitive and highly paid tasks such as coding, marketing copy, and website images can benefit from workflow tools that incorporate AI features. The ability to summarize or generate text and images in a high-fidelity manner opens up new possibilities for product applications.

However, it is crucial for startups to avoid the trap of becoming a hammer looking for a nail. The key is to identify actual end-user needs and untapped markets that can benefit from the advancements in AI technology. Focusing on the needs of the end-user should always be the priority.

In the world of product engagement, understanding the behavior and preferences of your most engaged users is essential. The Power User Curve, also known as the activity histogram or the "L30," provides valuable insights into user engagement. It is a histogram that shows users' engagement based on the total number of days they were active in a month.

Power users, who are highly engaged and contribute significant value to the network, drive the success of many companies. While the common metric of DAU/MAU (daily active users divided by monthly active users) provides a broad measure of engagement, it fails to capture the nuances of user behavior. The Power User Curve, on the other hand, reveals the variability among users, distinguishing between slightly engaged users and power users.

Successive Power User Curves should ideally show users shifting towards the right side of the smile, indicating a hardcore, engaged segment that returns to the product every day. This level of engagement is particularly advantageous for social products that can monetize through advertising.

However, not every company needs to have a smile-shaped Power User Curve. Some product categories, like LinkedIn, may have a lower daily usage rate, but they have business models that are not tied to daily engagement. Analyzing the Power User Curve in different timeframes, such as 7 days for SaaS/productivity products and 30 days for others, can provide further insights.

Plotting the Power User Curve for different cohorts of weekly or monthly active users can also be insightful. In network effects products, newer cohorts should gradually improve as network density and liquidity increase.

The Power User Curve can be based on core activity, going beyond app opens or logins. It reflects the heterogeneity among a user base and allows for a deeper understanding of different user segments.

As the CEO or product owner of a platform, it is crucial to design the platform in a way that gives everyone a chance to succeed. Understanding the nuances of user engagement and leveraging the Power User Curve can help optimize product design and drive growth.

In conclusion, the landscape of AI value distribution is evolving. While incumbents have historically captured a significant portion of the value, advancements in AI technology and the rise of infrastructure-centric companies are creating opportunities for startups to claim a larger share. Additionally, understanding user engagement through metrics like the Power User Curve is crucial for building successful products. By focusing on actual end-user needs and leveraging the insights provided by the Power User Curve, startups can navigate the AI landscape and thrive in this exciting era of technological innovation.

Actionable Advice:

1. Build a product that is 10X better than existing solutions to overcome incumbent advantages and capture market share.
2. Identify untapped markets and customer segments that incumbents cannot serve effectively, and focus on meeting their specific needs.
3. Leverage the insights provided by the Power User Curve to understand and optimize user engagement, ultimately driving growth and success.