Which comes first - Great Products or Great Platforms?

This is a continuation of my other essay on Products and Platform where successful platforms, specifically Microsoft and Apple, found strategic value in not just being a Platform, but a Product company built on their own Platforms.

AWS is an interesting case study because they are a dominant platform whose initial users did NOT come to AWS because of their core Product, which was Amazon.com.

Microsoft's Platform benefitted because two of its flagship Products, Excel and Word, drove adoption of Microsoft the Platform amongst knowledge workers. That User base attracted more developers to build onto that Platform, and one of the greatest platform flywheels kicked into place.

Users, the knowledge workers, did not come for the Platform. They probably didn't know what an Operating System was.

AWS, on the other hand, attracted developers in droves with nominal advertising; Bezos basically made one trip to Silicon Valley and gave a talk, and that unleashed word of mouth.

I remember being at a happy hour shortly after the announcement of S3 and it was all the rage, and already developers were hacking on new things.

AWS may have benefitted from scale and design principles gleaned from running the most demanding application in the world, Amazon.com. But it didn't benefit from the users who came for that Product in the same way Amazon did.

In fact, for context, we need to appreciate who were the legacy incumbents that AWS slaughtered. The hardware providers -- IBM, HP, Dell -- who were the compute and storage giants. VMWare, which had the tools for virtualization (the concepts which AWS applied to massive success). Even the Data Center companies like Equinix and Networking giant Cisco, while not directly competing, missed the opportunity of cloud services, which they would have had deep insight into.

The common response is that because they had APIs, they won. But many companies offer APIs now, and they largely will not ever reach the scale of AWS.

What was important was the context which AWS disrupted. The developer's user journey helps in one part.

In the past, for the developer to get work done, they had to deal with IT; IT held the keys to the kingdom for provisioning or rationing out the resources like compute and storage and everything else.

This was not a great experience in the journey. It often involved a service ticket, waiting, justification to other people....

AWS's actual innovation was that the APIs permitted developers to bypass IT and just start building with a credit card.

That alone unlocked developers, especially among start ups, who were typically fettered by this process.

But why was the process so onerous?

Because the investment impacted the business models for companies, big and small.

Buying a rack of hardware from Dell, HP, IBM, EMC, whatever -- was an up front capital expense that needed to be properly used.

Capital expenditures that have difficult payback periods have a difficult time being approved.

This represented business risk to the companies to do this. Thus they had IT be the watch dog effectively.

Once developers were unfettered, they could begin to build.

When it became time to escalate the bill, it was pay for the drink, not a big CapEx expense the CFO needed to get involved with.

These two forces unlocked demand: developers were unblocked to get what they want; CFOs saw a much smaller initial bill but could still start to realize payback on the digital initiatives.

What a win.

A business which unblocks a hungry audience from something burdensome; and frees up cash-flow while reducing payback risk will have massive tailwinds...and this is largely why developer tools and software-as-a-service in general has had such a run in the past 15 years.

In general, unless one finds something that impacts the business models of other companies in such a dramatic way, I think they will need a lighthouse product to attract the developers.

Does the scale need to be as dramatic as what AWS did?

Some of it is timing. Now the world is used to service models, so there isn't as big of an unlock.

Crypto, however, introduces much lower costs for institutions. This is why we are seeing institutional adoption.

Crypto also made it easier for a segment of the population to get unblocked from accessing financial rails...removing friction for pent-up demand, as AWS did for developers, drove many individual retailers, especially in areas with bad economic infrastructure, to jump in.

So we are beginning to see some potential similarities, but the issue has to do with access to those users.

The Apple Store, for example, attracted users with its native apps, Keynote and Pages (which now we see as low-demand); but it was enough to bootstrap Users to buy the iPhone and go to the Store, which then allowed a huge wave of successful mobile apps to get free distribution and discovery (till it became congested, as it is now).

However, the rise of stablecoins didn't benefit builders who interacted with the stablecoins, because stablecoin usage did not directly translate into user flow to their applications.

What did happen was new products were built on the substrate of blockchain to allow access and trading to stablecoins and other assets, particularly exchanges.

Exchanges now have users (and if they can figure out how to drive those users to developers, such as a marketplace or perhaps something more web3 innovative), developers will come. But the flywheel only really works if those new dapps drive the users to come back to the platform.

Rollups as a Service have this opportunity potentially. The rollups hopefully will be driving users to their respective chains; and surface to the chains the dapps that are running (if they don't do this, then the users don't really drive end-users to the dapps).

Yes, there is some indirect reasons, which is more liquidity on chain; dapps want to be where liquidity already is. But the challenge I find is that the User discovery is what is important.

Whoever owns that User discovery ultimately wins because it is actual Users that typically most applications want.