Will the GPT Store Really Create Millionaires?

And is it the right traction channel for you?

Yesterday, OpenAI announced the release of The GPT Store, a storefront where third-party developers can sell their Custom GPTs.

The concept of the GPT Store, first teased by Sam Altman at the Open AI Dev day in Nov 2023, has often been compared to Apple's iOS app store, both as a platform for third-party commerce and as a significant new source of revenue.

In the excitement since, some have speculated that the GPT Store will "create millionaires" and present opportunities for the next generation of entrepreneurs.

However, as someone that has tinkered with creating several GPTs myself, (including one made just for fun and three more serious ones in progress) I don't believe the App Store is a fair comparison for the GPT Store.

I'm also skeptical about it being the millionaire-maker it was initially hyped to be.

To understand if and how the GPT Store could be a valuable channel for your startup, let's start by looking at how the business model might play out.

How Can GPT Builders Make Money On The GPT Store?

OpenAI's launch today enabled all customers to search, browse, and use community-created GPTs. However, the anticipated "Buy" buttons were missing, and details about payment methods for builders were unclear.

The announcement only mentioned payment once:

As a first step, US builders will be paid based on user engagement with their GPTs. We'll provide details on the criteria for payments as we get closer.

The reference to a “first step” is consistent with Techcrunch’s reporting from November that they “Expect the strategy to evolve a lot, first with a straight revenue share (of unspecified magnitude), and later the possibility of subscribing to individual GPTs if there is demand for that.”

In summary, the final model of how OpenAI will compensate creators for their GPTs is still uncertain.

To evaluate whether the GPT store can truly "create millionaires," we should consider several potential business models that might evolve:

1. The model announced today,

2. The model most builders prefer,

3. The model likely to be most valuable in the long term.

What Was Announced Today: Engagement-Based Revenue Share

It seems OpenAI will pay GPT builders a share of Open AI’s revenues to the GPTs that drive the most engagement (according to some currently unknown measure.)

This is not the same as allowing GPT builders to charge for GPTs, which is what most early speculators seemed to expect when comparing to the iOS App Store’s Paid Apps model.

In fact from a commerce perspective one could argue that a “store” isn’t really the right metaphor at all. With only a select number of makers being rewarded for their popularity, this is more like Spotify than the iTunes Store, let alone App Store.

Any engagement-based rev share is likely to enrich the already rich and famous, more than the next generation entrepreneurs that many were hoping for.

Indeed top listed apps in OpenAI’s release and unofficial listings include big names like Canva, Open Trails and Khan Academy.

Getting seen and getting engagement

A by-product of Apples App Stores has been the new job of ASO - App Store Optimization. Yes, the task of ranking and visibility on the App Store is a full time job.

But while Apple’s various App Stores were fairly well polished from launch, the GPT Store is much less refined, creating even more challenges for getting seen, getting engagement and getting paid.

For proof that success is possible for relative newcomers too, look to Consensus, which allows you to search 200M academic papers, get science-based answers, and draft content with accurate citations.

I’ve had the fortune of meeting Eric and Christian from Consensus a couple of times and have watched as they shipped at incredible speed, launching one of the first GPTs to market, with a use case that fits the medium absolutely perfectly.

I asked Christian Salem, Chief Product Officer of Consensus about their experience of discovery on the platform

Even before the launch today, we were #1 on many "unofficial" lists and twitter threads, so we acquired lots of new users that way.

So it is certainly possible, but I fear for many categories, the GPT store will likely be a force-multiplier for brands and personalities that already have an audience.

This means that the first priority for a builder wanting to make money on the GPT store is not actually building, it’s marketing.

What Builders Wanted: Ability To Charge For Apps

The business model that I suspect most people would prefer would be where makers sell apps through store, with OpenAI taking a cut, much like Apple’s model.

Why This Might Actually Happen

OpenAI’s announcements have left open the possibility of moving to this model in the future, and it does seems likely at some point due to the incentives created by their cost structure.

Unlike Apple, OpenAI incur compute costs that rise with usage, so the very GPTs that OpenAI will pay commission to for engagement, will also be the ones costing them the most money.

This is viable in the short term because OpenAI’s current priority is to encourage users to upgrade to paid ChatGPT plans, and outsourcing the production of GPTs to 3rd parties exponentially increases the value offered to subscribers.

But if/when OpenAI change priorities, sharing revenue on purchases like the iOS store could be a more sustainable model for them.

Did The iOS App Store Even Make Millionaires?

There can be little doubt that Apples various App stores were behemoths of commerce.

There are more than 90,000 paid apps on the iOS App Store, and By 2022, Apple had paid over $260 billion to developers.

Furthermore, the channel is unlike anything else before it, and was probably not dilutive from any other channel. So that’s $260bn of net-new value to the economy.

A little Perplexity searching (my new Googling) reveals there are even some indie app developers that reportedly make millions from the platform - so yes.

But the GPT Store is Different to the iOS App Store

The extreme simplicity of creating GPTs makes them less defensible than iOS apps

Despite newer tools making iOS app development easier, it remains inaccessible to most non-technical individuals without hiring a developer.

This barrier to entry means that even a straightforward app with limited functionality but good user experience is worth customers paying a few dollars, as most customers are unlikely to create it themselves.

Contrarily, custom GPTs are extremely straightforward to create, fully lowering the barrier to entry. This not only allows for rapid replication by competitors but could also discourage users from purchasing at all.

If a user appreciates the value of a GPT, they might choose to create their own version instead of paying for one, especially if they can further customize it or use it long-term.

I asked

Charlie Guo

of

Artificial Ignorance

about this point:

I don't think custom GPTs are defensible as a standalone business case, period. Not only is the barrier to entry low, but we've already seen evidence that custom GPTs are just as susceptible to prompt injections that leak the custom prompt, which is usually the entire value of the GPT.
–Charlie Guo

Unlike iOS Apps, Moderation of GPTs is Limited

The iOS app store is known for requiring pre-approval for every app release. Although this can be a hassle for developers, it's crucial in building consumer trust in the store's apps.

OpenAI's release states: “in addition to the existing safety measures we've built into our products. The review process includes both human and automated review. Users are also able to report GPTs.” However, Techcrunch speculates the reviews will be minimal.

While this might initially seem to make the GPT store more accessible and profitable, it could lead to increased caution among users. This, in turn, could make user adoption of GPTs more challenging for new brands.

How To Make Money In An App Store Style Paid GPT Model

If OpenAI does move to an iOS style paid app model in the future, there are a few ways you can increase your earning potential.

Include Proprietary Data:

Each custom GPT offers the option to upload proprietary data documents or connect to custom APIs. This enables the builder to train the GPT on distinct data and generate unique responses that no other language model (including ChatGPT) could produce.

If the data source is valuable and attractive to customers, the GPT will deliver a value-added experience worth purchasing.

Apply Proprietary Methods:

This is not about what the GPT does, but rather about how it does it. A significant aspect of developing a GPT involves customizing its actions and incorporating specific proprietary methods. Techniques for writing or evaluating data can be integrated into the GPT to yield informed outcomes without necessarily disclosing the method.

This could work hand-in hand with…

Add “Celebrity” Appeal:

A cooking GPT created by a celebrity chef or a popular restaurateur will be more desirable and valuable than one created by unknown app developers.

Similarly, influencers, popular creators, and known brands could also succeed with branded versions of simple functionalities that are well promoted. This might be more successful than superior functionalities from unknown creators.

Consider Category and Use Case

Just as the iOS App Store has certain categories that have proven a perfect fit for the form factor of iOS, apps, and devices so we might see certain categories that best fit the form factor of GPT.

Consider how natural a fit for your customer use case the GPT experience is.

I see productivity being the biggest category. We're seeing some evidence here already - the most popular third-party GPT is Grimoire, which is focused on software development. ChatGPT Plus, which is required to use GPTs, is $20/month. That's a no-brainer from a productivity standpoint, but difficult to justify from an entertainment standpoint when iOS games and social media are free. –

Charlie Guo

Don’t Leave Discovery To Chance:

As previously mentioned, bring your existing audience to the GPT by leveraging other channels such as your email list, social media or signage on key touch points. Earn a place on the leaderboard for your category then let the GPT Store boost usage for you.

Enable Exclusive Integrations and functionality:

GPT Builders have the option to integrate with other systems including in-house proprietary systems via API, opening up many unique capabilities. This is probably the most powerful step of all, as Charlie Guo again explains:

The GPTs that have a shot at being defensible are ones that incorporate third-party data or APIs - like a travel agent GPT that can actually book flights and hotels for you. But in that case, you already have a standalone business, and GPTs are simply a new distribution channel.

Which leads me to last category of ways to make money with the GPT Store, which is to not limit functionality and monetization to the GPT Store alone.

What Makes The Most Money: Using the GPT Store as One Traction Channel Among Others

Indeed, focusing the discussion on the iOS App Store solely around paid apps misses a significant source of app revenue. Instead of selling the app for a price, offering a free app that requires a subscription can generate more income.

A report from Sensor Tower revealed that in 2020, global subscription app revenue from the top 100 subscription apps alone rose 34% YoY to $13bn, a $3.3bn increase from 2019.

Therefore, we should consider the GPT store in a similar way. Instead of solely focusing on the revenue from customers purchasing GPTs, we should integrate it into a broader strategy.

7 Ways To Monetize Custom GPTs Without GPT Store Payments

Here are some ways a Custom GPT could deliver ROI without relying on revenue share or buy buttons from ChatGPT

Develop a “Loss-leader” GPT to fuel discovery of your core business

Custom GPTs include a clickable link back to the developer, which can guide users back to your main site or app if your GPT is discovered. Consider creating an app in a popular location using promotional mechanics, such as a Product Hunt launch or AI newsletter.

This inclusion can drive traffic to the GPT and kickstart its ranking. Then, allow GPT users to discover your brand through your free GPT.

Create a “taster” product for a more valuable core experience

Just like luxury brands offer affordable products such as lipsticks, fragrances, and sunglasses to introduce consumers to their brand, you could offer a sliver of your functionality in the form of a free or low-priced GPT.

This strategy allows you to deliver initial value, which warms customers to your brand and proposition. Moreover, it may be possible to gather data or integrate further into the main product, ensuring that GPT users have a clear path to upgrade and migrate into full product use.

This allows customers to start experiencing value, potentially sending or saving data to an account.

Create a GPT as a bonus for your existing users

So far, I've mostly discussed creating GPTs as a tactic for new customer acquisition. However, it could also be valuable for customer retention. Given that your competitors might try to lure your customers with a GPT function added to their offerings, it might be beneficial to offer a free GPT accessible only to existing users. This could provide them with additional reasons to remain subscribed to your service and deter them from switching.

Increase productivity of internal functions or billable services

Considering ROI in a broader sense, we can envision a Custom GPT helping by reducing internal function costs or by increasing output. This also means that a service provider, such as an agency, could utilize a GPT for internal use. This would allow them to deliver more results or enhance deliverables for a given client at the same price or potentially less.

Imagine setting up a "social media GPT" with your company's target demographic and overall style/tone, and then handing it to a social media manager who can become 5x more efficient at brainstorming posts. –Charlie Guo

All of the above, blended within the bigger picture of your channel strategy

There's no need to limit yourself by thinking of custom GPT's as being only for acquisition or retention, or any one of these other functions. In fact, you can consider the many roles it may play within your broader ecosystem.

So far, our GPT has been critical for both growth and engagement. Many new users found us through the GPT, and many existing users love the flexibility that the ChatGPT interface provides as a supplement to our flagship search engine product.
–Christian Salem, Consensus

Use the GPT, but charge offline

You don't need to wait for OpenAI to create a charging model that suits you.

Any custom GPT you develop can be published to a private, unlisted URL. This means it won't be discoverable on the GPT store, but you can make the link available privately. One opportunity is to place the link behind a paywall. This means users can only access the GPT once they're part of a paid app or community.

Skip the GPT Store entirely, choose a different builder

A significant limitation of the GPT store that hasn't been discussed yet is that custom GPTs are only available to users with a paid ChatGPT subscription. Naturally, this restricts the potential reach and distribution of your GPT.

However, the ChatGPT store isn't the only builder you can use. Zapier is another option. With Zapier, you can create similar logic, pre-training, fine-tuning of behaviors, and integration with custom data. Plus, it's connected to a front-end that is accessible to any user.

So, consider your potential demographic. If it's probable that your users won't purchase a ChatGPT premium subscription, creating functionality using a different builder like Zapier might be a better choice.

So Will the GPT Store Create Millionaires?

Perhaps, but not in the direct manner that most pre-launch hype suggested.

In the specific case of wanting to be paid directly through or by the GPT Store, except for a few notable successes like Consensus and Grimoire, I suspect the GPT Store may primarily benefit established players rather than newcomers.

However, when viewed as a new tool within your wider offering and ecosystem, Custom GPTs do open up many exciting opportunities. They allow businesses and creators to monetize their knowledge and data through a unique new interface that is already in the pocket of millions of ChatGPT users worldwide.

The prospect of integrating Custom GPTs with no-code/low-code applications, communities, and premium content creates a highly alluring package. This package is accessible to small startups, solo entrepreneurs, and established businesses alike. It's plausible that this could propel many businesses to new revenue heights and create new millionaires.

I hope you will be among them.

What do you think?

Once again, I have indulged in speculation and predicting the future, which means I may be wrong on some points.

However, I am eager to learn through the process of contemplating these potential futures and from any feedback you might have.

So if you have opinions on this, whether you agree or disagree, please don't hesitate to leave a comment and let me know.

Subscribe for more

If you're interested in how evolving traction channels in the post-AI era impact your startup, please consider becoming a free subscriber.

You can also follow and DM me on LinkedIn and Twitter if you want to chat or visit me at TractionDesigner.com