Why are some companies able to capture such a chokehold on their market? Love them or hate them, the most successful companies in the world have reached that status because their platforms have evolved into ecosystems.
Let’s deep dive into Meta.
Back when it was Facebook, Meta created a product that would fundamentally change the way we interact with one another: a platform used to keep tabs with one another, a network upon which we could communicate with more than one person. As it has evolved, this platform became a basis for several different innovations as people became reliant on the siloed technology. Having delivered a truly intuitive UX and amassing a critical number of users on the platform, it was naturally difficult for the initial core group to defect to other social platforms like Myspace. As time went on, they learned to adapt to their user base, quickly maneuvering and catering to what their customers wanted. In more recent times, we can see this directly with the cannibalization of Instagram and a build-out of Threads – their Twitter rival. This has enabled Facebook to further build out their platform, creating an ecosystem of social networks and communication platforms that keeps their ecosystem sticky. With the addition of Whatsapp, the company has further entrenched their users, keeping them on platform, enabling them to capture more data (and value) from their customers, long term.
From a UX perspective, platforms like this are incredible for the user. They enable users to navigate throughout their different accounts quickly and efficiently, uniquely tailoring the experience across platforms to what you as the consumer do (or don’t) like. It streamlines the experience – one login, connected accounts, similar feeds…
Let’s take the example of social media algorithms in creating a good user experience. The algorithm is paramount to keeping user engaged. It learns and develops and understanding for what types of videos and content you might like as a user, subsequently showing the content you want to see in your feed. This is also how it know what to advertise to you.
Facebook, because of its breadth of access to your interests through Facebook, Instagram, Threads, etc. can uniquely curate the entirety of the experience on (and off) platform for you.
Did you like a photo of a designer living room on Instagram? Well, now you might find yourself seeing that same aesthetic come up again in your Instagram/Threads/Facebook feed.
This is the beauty of the platform. If you can create an “ecosystem,” whereby the company interacts and learns about the user from multiple contact points, it can adapt and better deliver a top-tier experience.
Obviously, this intuitive experience is exactly what keeps the customer engaged. It’s the reason you can’t stop doom scrolling.
Different companies have different monetization strategies, but for the case of Meta, as users spend more time on platform, the company makes more money. Marketers pay for eyeballs and eyeballs are drawn to what they want to see. Meta has a fundamental understanding of this – it is precisely how they have scaled. More eyeballs = more money.
Ecosystem: Social Platforms, Ads
This is a prime example of a positive feedback loop (I’ll dive deeper into this in an upcoming post).
The ecosystem approach is the one (seemingly) long-term successful and sustainable approach: it is a self-propelling positive feedback loop. The user keeps coming back because it is too intuitive to seek and adopt an alternative. The lifetime value of the customer is virtually infinite when it’s done right, so the company continues to innovate to keep the platform sticky, and the customer keeps coming back - that’s the power of the ecosystem.
Initially, solely a search engine, the company began to add capabilities to keep users engaged. YouTube, Gmail, Drive on the consumer side. Ads, G-suite, Chips on the commercial side. On both sides, users become entrenched – Where else to find all these great videos? Why change email? How to target large audiences?
Ecosystem: Google Search, G-suite, Ads, Google Chips, Google Pay
Apple
If you use Apple devices, think how much would someone have to pay you to depart from the Apple ecosystem? Think about all of the data migration… learning the new interface of Microsoft or Google, etc..
Where Microsoft has captured the corporate audience, Apple has captured the creative audience. This lean into pop culture has helped to cement their position as a “premium” brand.
Ecosystem: Apple Computers, iPhone, Apple Chips, iCloud, Apple Pay, App Store
Microsoft
Microsoft is uber-well positioned across corporations globally. Especially in finance, they are deeply rooted with Excel and PowerPoint. Even Outlook has become a dependency. They announced the in-house development of their own chips in November of 2023. With their increased emphasis on Azure, the company has further positioned itself as a long-term strategic partner for governments and corporations alike.
Ecosystem: Windows Devices, Android, Microsoft 365, Chips, Copilot, Azure
Amazon
Amazon does not have the hardware, but they certainly play a dominant role as a gatekeeper. With the throughput on their main business (marketplace), they have been able to keep customers coming back by their exhaustive marketplace and constantly evolving ancillary offerings. Initially expanding from books into a broader marketplace, then into Kindle, then further content with Prime Video, Amazon has learned how to keep the user entertained.
On the business-front, AWS has become a go-to for scaling businesses that need heavy cloud infrastructure. Obviously, the marketplace itself is a primary avenue for businesses to sell to customers, enabling Amazon to capture value as the marketplace intermediary.
Ecosystem: Amazon Store, Prime Video, Amazon Payments, AWS
NVIDIA
Sustained growth? Maybe? NVIDIA is ahead of the curve because of their work on AI chips. They are a platform that companies have become reliant upon. They are currently the golden standard and are the default for companies trying to build out AI capabilities: Amazon, Google, Meta, Microsoft, and Oracle all use NVIDIA chips.
These guys are the core infrastructure that other companies are dependent upon (for now). They have a platform, but not yet the ecosystem. The platform is the chips, and because they are so far ahead of the curve, companies are reliant on them for now.
This is why many companies have been making the move to “onshore” their chip-making – bringing it in house will enable them to capture value in the long term when they aren’t subject to cost hikes that are at the discretion of another large player in the space.
Once the big boys have caught up and figure out how to develop their chips in house, the NVIDIA party might be over unless they find another business line and way to keep customers chomping at the bit.
Ecosystem: Chips? For now?
LLMs + AI Companies
I’ll speak broadly here without being an expert on any one company, so the following is more of an overview on the landscape: OpenAI, Anthropic, Groq, Mistral.
Most people seem to think it is inevitable that one (or more) of the large LLM operators will eventually join the ranks of the Googles, Facebooks, Apples, and Microsofts of the world. At this point, it seems like OpenAI is the shoe-in – between their lead in the retail market and their strategic partnerships with Microsoft (and now Apple), it would appear that they have a leg up.
The retail (consumer-facing) application of these LLMs is chump change. It doesn’t make these companies money. Their money is being generated on the commercial side, and the real money won’t be seen until the companies begin white labeling their LLMs for internal usage at other companies. But that’s exactly the issue: the companies that would pay for and propel these LLMs to the heights of the magnificent seven are exactly that: the magnificent 7, and they are all developing their own models.
Yes, they have strategic partnerships with OpenAI (Microsoft, Apple), Anthropic (Amazon, Google), Groq (Meta), but they are only leveraging those models temporarily. LLMs can only get so smart from public data – there’s a finite amount of it out there, and as we’ve learned, when the model tries learning from itself, it’s not the most accurate source of truth (see flubs with Gemini & GPT4).
So, what’s next?
The companies build their own models to learn from their data to heighten the user experience. Yes, the OpenAI, Anthropic, and Groq will be around for a while, but are they going to reach the top of the market? Doubtful. They will likely get gobbled up before they reach that level.
The biggest issue with each of the LLMs out there right now is the broad swath access to the data.. it’s the same reason why so many companies are reluctant to let employees use the chatbots – they don’t want to risk the data getting out there. So rather than relying on external providers of “white label” LLMs, each of these companies will soon launch their own internally, tailored uniquely to individual users (in some cases even localized on device). Yes, the compute will go through the roof, but this is exactly why the companies began to decouple from chip providers years ago. Look at Apple departing from Intel in January of 2023 and Microsoft developing their own chips in starting in November.
OpenAI is undoubtedly here to stick around and will likely be here for some time to come. Solely based on their strategic partnerships with both of the largest hardware distributors (Microsoft & Apple), OpenAI is likely the only operator at the moment that will be able to capture significant value. But how long that will last depends on how quickly Apple can standup its own model.
The reality is that the top performers don’t want to rely on anyone else and they don’t need to. Once they have developed their own models, they can spin out closed circuit models for their users: this is exactly what the top performers are working on: Google – Gemini, Meta – MetaAI, Apple – AppleAI, Microsoft – Copilot.
My thesis?
Apple uses OpenAI until its internal AI is ready to go, then cancels contract with OpenAI. With Apple’s internal AI, they will be able to deliver better UX, maintaining the stickiness of their ecosystem, continuing their dominance in the consumer market.
With the depleted value of OpenAI, Microsoft will be able to buy the company outright and onshore it to fully integrate with Copilot. Copilot’s capabilities will further entrench commercial users of Microsoft.
Google’s Gemini will overtake its own search capability and Apple will continue its long-standing partnership with Google for “search.” This will bolster Google’s bottom line, but they will remain trying to carve out market share from the hardware market.
Amazon will continue improving its AWS business as compute increases in demand with the continued AI adoption & integration.
Meta is the clear winner in entertainment. Strategic partnership with Groq will enable them to leverage cheaper chips than NVIDIA, but reliance on third party will chip away at bottom line unless significant (yet unknown) Zuck equity exists.
NVIDIA is hot, but for how long? Let’s see when the next round of Apple, Microsoft, Amazon chips come out.. Especially with Groq nipping at their heels, it will be interesting to see how long they can extend their lead.
So what?
The money is where the ecosystems are, because that’s where all the personal data is.
This is why Meta is going to make a killing – seeing your screentime, the accounts you interact with, and figuring out how to keep you in the loop for longer so they can sell more ads and generate more revenue, that’s where they want to spend their time and money. Meta AI is going to curate your entertainment experience.
Google is also going to do well, they know your search history, they know what you watch on YouTube, they know how you type your email, and they know what ads to show you. Business users will enjoy the added benefit of Gemini, but lack of market saturation with hardware makes it difficult to capture outsized long-term value unless Gemini is able to surpass the performance of Copilot.
Microsoft is uniquely positioned with their significant stake in OpenAI. With mid-tier customers reliant on OpenAI and their substantial (and well-capitalized) accounts, the simple upsell for Copilot as an addition to Microsoft365 and will prove significant as modalities and use cases for Copilot continue developing. From a commercial standpoint, this addition streamlines your day-to-day, from automating menial tasks to saving time in putting together Excels and PowerPoints, they will be able to curate and simplify your work experience.
Apple – yet another killer. Because they are the hardware interface between you and every digital interaction you make, their hardware knows just about everything about what you do. The applications you use, where you go, what you buy… once they can streamline the entire experience, they will be unstoppable. Their completely verticalized business model enables them to capture the entirety of the value chain – they virtually pay no one. Think about how reliant you are to use your phone/computer for anything: setting alarms, booking trips, making reservations, buying movie tickets, grocery lists. They will be able to curate your living experience.
Returning to the ecosystem note above, once you’ve created an ecosystem that people are reliant upon, it becomes increasingly difficult to leave. The more time you spend in the ecosystem, the more data they collect.
At the end of the day, the winners will be the ones who have the most data – they will be able to better learn you and tailor the entire experience to you. The more time you spend with them, the more valuable they become.
- SL
Comentarios