.png)
4 months ago
27
Why AI-enabled content networks have a retention problem… and how to solve it
We are still in the early innings of AI adoption. I my previous essay, I outlined the evolution of AI applications and their potential for network effects. As of today, AI 1.0 is still very much in vogue — ranging from “vibe coding” products like Lovable to faux social networks like Status. Now we are beginning to see early attempts of AI 1.5 — broadcast networks for user+AI generated content. Most of these networks suffer from the same problem — poor retention.
Before we begin, here’s a quick primer on AI 1.0 and AI 1.5 applications.
This is how I defined AI 1.0 applications:
These applications enable a direct interaction between a user and an AI product, i.e. software.
They represent the earliest wave of AI applications after the emergence of foundational models. I call this category AI 1.0, as users interact directly with AI (created by companies), along the lines of Web 1.0 where users accessed information (created by companies).
There is no possibility of network effects here. Network effects exist when the addition of a user increases the value of the product for all users — this requires a multiplayer interaction, i.e. one between two or more users. Interactions here are singleplayer, i.e. users are interacting with software. So by definition, network effects cannot exist.
In short, these are commodity products. The combination of low switching costs and consensus thinking has led to high competitive intensity. This has also allowed incumbents to compete directly with AI 1.0 startups. Examples include Figma and Canva copying Lovable, Notion replicating Granola’s AI notetaking, Meta taking on a bunch of startups with its own AI ad creation tools, etc.
In contrast, AI 1.5 applications are halfway between pure software and true networks:
There is another category of AI applications where the product is a bit more than “pure software/AI”. Instead, users create an AI-generated content or supply unit, which is then exposed to other users. Crucially, the process of creating AI-generated content is frictionless and takes just a few taps/clicks or a simple prompt.
Network effects do exist here… in theory. User adoption leads to more AI-generated content which other users can interact with. The challenge is that the friction for creating supply (AI chatbots or songs) is very low, and so is the barrier to reaching a critical mass of content.
This category includes companies like Suno and Udio. They allow users to create AI-generated music with a simple prompt. Users can also discover music created by other users. These companies face less competition than AI 1.0. However, early attempts at replicating this model have hit a snag.
Over the last few months, I have come across numerous early-stage AI 1.5 startups. They shared a common theme — they were entirely focused on creators. At first glance, this is shouldn’t be surprising. AI supercharges media creation. Work that previously required painstaking effort now just needs a prompt (and some GPUs).
However, these startups struggled with user retention. To fix this, most founders jumped to the “obvious” solution — making it even easier to create… moar AI. After all, creation is what AI does best. Making that even better should solve the retention problem, right? Wrong.
The truth is that AI-enabled creation is the root of the retention problem. By drastically lowering the barrier to creation, AI causes an explosion in content volume — whether it’s music, images, comics, etc. But quantity doesn’t equal quality. Instead, the end result is high volume of low quality content. This leads to low interest and poor retention on the consumption side of the network. And without a strong consumption side, all that is left is a commodity AI 1.0 tool.
To solve this retention problem, you need to fix the root cause — content quality. Low friction in the creation process leads to a high volume of low quality content. So if you want to increase the quality of content, you need to increase creative friction in the content creation process, i.e. introduce constraints or make it harder to create, not easier.
The solution should require users to use more creativity, not less. As a consequence, the number of creators could decline — users who don’t want to put in the effort will not create. However, the users that do create will put more thought and care into their creations. As a result, the output will generate more interest from the consumption side of the network leading to increased retention.
This creates the conditions for network effects to kick in — creators increase the volume of high quality content on the network, which increases value for consumers and more consumption creates a better incentive for creators to continue creating. This also moves the product closer to a true AI 2.0 network.
In closing, remember that products with network effects defy conventional wisdom. As they are built on a multiplayer interaction, their efficacy depends on the experience of multiple, interlinked groups of users. Standard advice like “make it easier to create” can backfire if it enhances the experience of one user persona at the expense of the other. Instead, I recommend following the user journey of your key personas. Once you do that, be intentional about the trade-offs required to improve the experience for the network as a whole.
1. Reach out: No warm intros required
I invest in pre-seed/seed-stage startups as a Partner at Speedinvest. If your startup is built on network effects and based in Europe, you can reach me here or here.
If you’re not ready to pitch yet, sign up for my weekly office hours to talk through any questions about building network effects or consumer products.
2. Take my Reforge course
Want to learn more about network effects? Take my Reforge course on identifying, bootstrapping, measuring and enhancing network effects. Start the course for free via this link!
3. Take the network effects assessment
Benchmark your network effects by answering a 1 min quiz about your startup or a potential startup investment. Try it here.
