I’ve built my plugin on top of the functionality provided by PluginLab.
Looking at the analytics after the first few days… I’m thinking it would be great and best for all parties involved if OpenAI did revenue sharing with plugin developers.
It is the stated goal to make it easy for people to install and try plugins. Rightfully so!
On the other hand, if we want to have a thriving ecosystem then there has to be some way to monetize. After all, running the backend for the web services that the plugins expose does have costs … not to mention the development time etc.
PluginLab solved the issue of no friction installation and monetization by using quotas: it is free until a given threshold is reached and then it asks for a subscription.
Fine… but what is the right amount to charge?
When most people only use a plugin 4-5 invocations a week… what are you going to charge them? 10$/month? 5$?..1$? It doesn’t really make sense unless you find a niche where the plugin would be used frequently enough to justify the cost.
So unless we have some kind of crypto micropayments the best thing would be revenue sharing.
BTW, selling subscriptions and revenue sharing are not mutually exclusive. Look at X (Twitter) that has both. I would like to see OpenAI adopt a similar model.
It’s definitely an idea and may be worth a discussion how much monetary value the plug-ins are creating for OpenAI.
Let’s think this through.
We consider only the plus users and the 20$ bills. We deduct the cost of running the business and offering the service or a fraction of it. Then we look at all messages/input-output tokens per month. Then we relate this number to the tokens directly generated by a plug-in. Then you deduct the cost for administering the money transfers, the fraud checks etc…
As a developer you may be eligible for a few cents at max. This guesstimate is also based on my experience in the PPC online advertising space.
Now, a few cents are not nothing, right? But it’s nothing compared to a properly paying customer base that regularly uses your plug-in. And these customers, I suggest, you charge yourself.
there’s the YouTube business model where they share ad revenue based on amount of views which is interesting. There could be a world where sharing based on times used per month could lead to enough revenue to build a business around.
I feel like OpenAI’s main focus is making the models better and I don’t see any inkling that they would want to have to hire to deal with something like this.
I think if there’s more distribution, like plugins to all users not just plus, it would be easily feasible to build a subscription business with a handful of team members taking salary.
In an ideal world i’d personally like a credits solution where a user buys X amount of plugin credits and each time they use a plugin XX% of that credit goes to that plugin.
We really miss a simple monetization solution here. Plugins cannot be monetized efficiently externally unless they are just a gateway to an already existing, established product. OpenAI’s plugin store and support is inferior and I don’t see them progressing ahead with it. Will 3rd party plugins be abandoned entirely at some point?
I am trying to better understand the draw for developers. I enjoy using the Wolfram plugin as a plus user. But how does Wolfram benefit? Revenue sharing might work (taking a small amount from the plus subscription for plugin devs). I am also curious how this translates to the new gpt bot actions. I searched and did not find Wolfram AI actions that would be used for bots.