🔮 The Post "App Store" Monetization

We are in the early days of new emerging monetization models that should over time overlay the current ones.

🔮 The Post "App Store" Monetization

I was lately thinking about the latest news around application stores ranging from the pressure over 3rd-party payments on Apple/Google to the “Open App Store Principles” proposed by Microsoft. And it seems to me like we are in the early days of new emerging monetization models that should over time overlay the current ones.

The Drivers

Before we dig into details, there are three critical reasons for new models to appear at all:

  1. Few companies almost monopolized revenue streams in apps and ads. So it is good timing for regulators and challengers of the status quo to play their historical role and push evolution and competition forward.
  2. The dominant models are now with us for more than 15 years — a long enough period for optimization, tuning, etc. It means they are becoming slow in changing and adapting, they can’t generate exponentially more revenue, and in the long-term, they create existential risk for their benefactors (because of p1.)
  3. We have a generation of users who have grown along with the growth of these models. They didn’t experience the world before but would like to see it after. They will define new “post-app-store” rules, policies, and technologies, making them different.

The Status Quo

The existing and future monetization models form an onion-like structure. The new ones will add new dimensions, layers, and options and not substitute the established ones.

But let me briefly remind you, what do we deal with today?

  1. The first fundamental layer is the monetization of the socket and its function. Remember the time when you had to pay for a Windows license? Or when you had to buy an application on a CD? Maybe it changed the form but didn’t go far away. When you buy a new phone, the cost of the operating system and critical apps is an integral part of the device cost. In some cases, it is subsidized by the second layer if it is under the control of the same company.
  2. The second level is the monetization of the socket access. If you control devices, you build an app store. If you own a search engine or popular website (e.g., Facebook), you create an advertisement platform. And then you attempt to secure your sockets access: so you might kind of open-source your “fundamental layer” in exchange for default apps or build a browser that will default to your search engine. The basic monetization rule is straightforward: you take a tax. The more data you collect, the better you optimize the route towards getting “more taxable” streams.

The Emerging Models

Now we are ready for a quick overview of the emerging models. They are not radically new, but they are not mainstream, not yet. I define three models, ranged on their readiness:

The Portfolio Subscription (also known as super-apps and ecosystems)

(Don’t confuse it with the app subscription model, even if it looks like a similar regular payment in your bank account.)

It is the strategy behind the emerging Xbox Games Pass and all the Xbox Game Studios. It is the strategy behind Netflix and all the studios providing exclusive and licensed content. You pay not for an app but for access to a portfolio of content, services, apps, games, etc. (I will generalize all of these as just “content”).

The company managing and building such a portfolio tries to maximize its assortment to fulfill the diversity of customer preferences while reducing licensing payments. It is almost inevitable that such a company will eventually produce first-party content, buy the best content producers, and request the rest for exclusives.

With the growth of the audience, the app itself becomes an access socket (or channel) for those who produce content. But compared with the current two models, there are differences:

  1. We count sockets in user hours (or minutes) of watching, using, or playing and do not care much about the devices. So each minute customers are on your channel, they don’t consume your competitor’s content. That means you should provide not just the main show, but also all the time-fillers.
  2. We share revenue with content producers based on user engagement. E.g., most-watched shows should get more money, but we didn’t reach such an equilibrium yet. It also requires transparency and built-in trust engines to ensure that every creator receives what they deserve.
  3. You begin looking into complementary use cases trying to lock in the customer on your app as much as possible. So you start building a multiformat super-app or ecosystem of services — an envelope covering user life. (E.g., no surprise Netflix is extending into games.)
  4. Eventually, that lock-in feature will mutate into internal “in-app” currency that you will use to credit future development, return customers, and manipulate decisions.

If Portfolio Subscription is something we observe already in wild nature, the following model has just begun rollout.

Federated Recommendations

Google is building it with FLoC (discontinued) and its Topics API (a new incarnation) over Chrome. Android and iOS should include such features based on federated machine learning in the not-so-distant future. I suppose Microsoft is building something similar over Windows 11 and Edge. The list continues but is short. Why so?

The key driver for such a shift from the current ad models is the fight over privacy. So the ghost of Big Brother would be substituted with the host of Little Sister. But, critically, you have to pass through the “Big Brother” phase to collect the data and train the model to build that Little Sisters Swarm.

“Litter Sister” is an advisory agent attached to the customer experience. It combines local knowledge specific to a particular user and global models trained in the cloud on the shared and anonymized data. That is the federated model. Little Sister might act in the background through APIs and as an intelligent assistant interface in the foreground.

The key feature of such an agent is to explain to other services (apps, ads engines, etc.) what is the most preferable or suitable choice at the moment for its client. Like what they need, might need, or will dislike.

Compared with existing models:

  1. We monetize the function of choice (ultimately — of any choice). Agent influences the behavior of other apps, engines, and services based on local awareness and global models. Modern app platforms already provide a robust and extensive framework of “user preferences,” but it is just the beginning of the story, with recent additions like the focus mode.
  2. The revenue share will depend on the effectiveness of the recommendation, but that “optimization” function depends on the app or service goals, not just the user context. So the mathematical and economic model for such a shift is not developed yet. And probably it will require some cloud backend and frameworks as well — federated learning as a service or FLaaS as we will name it.
  3. Another aspect is that we have to return users their feeling of having the situation under control. Based on a potential benefit, they should decide what data is shared and the frequency of updates. That is a tricky moment because we should explain how much we know about them. It will surprise and horrify. So building transparency and trust are the keys.
  4. Based on point 2, we should expect a monetization mix like the app “buys” some credits of user awareness on the local side and pays for computing resources and data storage for the federated model on the backend side.

The primary driver for this evolution will be a strong desire to move out of the regulator’s sight. And the direct commercial interest will be only in the second place.

Power Consumption

The third model is in its incubation stage. Today, we might observe future building blocks, experiments, and fundamental infrastructure components.

First of all, it sounds pretty old-fashioned because we know that customers already pay for the power consumption in their electricity/power bills, as well as cloud providers. And power consumption of cloud apps is recalculated through other metrics like computation minutes, storage gigabytes, etc.

Secondly, we have known that operating systems can track power consumption up to every app they run for over a decade. As consumers, we observe how browser vendors, processor/chipset vendors, and many others try to sell us low power consumption as an advantage of their offers. We also have some backbone feeling that buying a more powerful device will impact our bills long-term and that playing an AAA game at 4K@120Hz+ boosts heat a lot.

So how the “power consumption” will become a new monetization model?

It is a four-phases trick:

  1. Make consumers aware of power consumption and its cost. We can examine as an example the crypto-miners communities. These guys perfected converting power consumption into computational operations and digital money. They measure the effectiveness of every chipset and app version they use and adjust it based on the electric power cost in a particular region. Imagine that a future update of your mobile OS will also add a feature that informs you on the total cost of the app or the whole device, notifying you how much you paid for it in your power bill. As if every time you make a call, use a browser, watch a movie, or play a game, your phone notifies: hey, it cost you 5 cents, or 1 cent, or 0.03 cents. And next time you open the app store, it informs you that the game you are going to buy for some $0.99 will cost you $5.99 with all the in-apps and, based on the power rates in your country, the total cost for your family wallet will be $9.83 if you and your partner play it through the end. What if clarifying such costs is mandatory from the regulator’s side?
  2. Make edge and cloud computing measured in the carbon footprint. We already see such early signals in some cloud platforms, but as of now, it is just data. The goal is to build a holistic framework to measure the ecological impact of any app or service ranging from client to network to backend to development and model training. Eventually, we should convince consumers that apps and platforms with less carbon footprint and better energy conversion are better. Such an intervention should bring into life a new dimension for the computational infrastructure: federated computations and, in general, global optimization for computing from cloud to edge, not just networking. It is where and when the ideas of the internet of energy and the global internet should merge.
  3. We should establish a transparent and trusted financial framework and engine to redistribute the power/carbon costs of “digital goods.” To explain this idea, I will start with an already present model for some cloud providers. If you watch regular updates of major companies like Google, Amazon, or Microsoft, they all promise to switch to clean, renewable energy by some year in the future. What does it mean? Are they all going to build their power generation stations? E.g., a hydroelectric power station or a tidal power plant? I don’t think so. They exchange: 1/ they consume power from the local grid, 2/ the network redistributes the energy, and 3/ the payment ultimately goes to the green energy provider. In the future, platforms will promote it as follows: if you build your backend using such a cloud, it is more nature-friendly and greener by default. But the cloud provider can’t yet influence how you charge the device at your home or office. Instead, they offer you an exchange: they run the apps for you in the cloud and stream them over 5G/6G to your lightweight and low-energy-consuming device. But you pay them for the “green” energy consumption. Is it the future? No, it is already present: try the game streaming services. The providers don’t label them as a greener alternative, but believe me: they will.
  4. Finally and this is where the new monetization model comes into play. If your home or office becomes smart enough, it might “sign” the same type of contract as modern cloud vendors do with their power providers. The operating system of your device should be able to bill you for each “watt” it consumed, reducing the corresponding value from your home bill, and “convert” it into a payment for green renewable energy. Now, when they “own” this money stream, they get the ability to optimize it using a mix of algorithms, distributed computing, cloud computing, etc.

I believe, it will be a game-changer for a long period of time, but also will require either cloud providers to take a step into the power generation business, or the power companies to step into the chipset business. Or maybe we should wait here for brand new players?