3 years ago a smart boss of mine made an offhand comment that “no one understands what the clouds are doing.” it was true then, and even more true now.
so, a thread on cloud platform incentives, and how to survive (and thrive!) as an ISV in the cloud era... contrary to popular belief, the big clouds aren’t taking over the market. they are expanding it.
First: you have to understand what the large clouds care about. Incentives drive behavior. Luckily, this is actually not difficult. The big clouds care about selling you compute and storage.
What does this mean to you, as an ISV? It means that if you drive compute and storage, clouds are incented to work with you - because they make money when you make money. Your success is their success.
Second: cloud is, simply speaking, a deployment model. A hybrid deployment model is table stakes going forward. Some customers may care where it’s deployed, others may not.
But your value as an ISV is in delivering a product experience that sits on top of clouds. You should be, religiously, data plane agnostic.
The value in your product has to be in your control plane, not the data plane. Why? Because the control plane is where you manage the customer relationship.
your control plane delivers the unique experience and differentiation that the cloud providers don’t. It’s where you own the customer relationship.
So, whether your data plane is deployed at AWS or Softlayer, you shouldn’t care. If you’re finding yourself worrying about this, take that as a sign.
Third: customers are centralizing their data to be managed in clouds, which means that it’s going to be in databases and data warehouses controlled by clouds. Assume a world where this is true of all data.
What does this centralization of data mean for you? Remember that clouds care about compute and storage - the more data is stored, and the more customers do with it, the happier the clouds are.
Their single strongest incentive is for customers to move data to the cloud, and will look to leverage both 1st and 3rd party products to make it happen.
Because of this, products should be designed such that they operate on a variety of data stores, rather than requiring centralization and management in your own product.
While the data may live elsewhere, your app has the opportunity to own the control of that data. the data model, the analytics — the value.
storage and execution remains a relative commodity, and the main driver of the broader cloud business, while you deliver value that sits on top, independent of cloud choices.
There are many examples of this in the data warehouse market already — both BigQuery and Snowflake have acquired new “apps'' that leverage their position as the data store. (Looker and Stride, respectively)
They both also maintain marketplaces of data sets and data apps, that give customers new reasons to move more data into the cloud, and to run more queries on it. Expect to see more of this.
In this way, the cloud is a market expansion story. The clouds capture enormous amounts of value in replacing the legacy data center with modern cloud infrastructure and tooling. ISVs can ride atop that wave in a symbiotic, not adversarial, manner.
BTW, the size of that wave? $120B in 2021, growing at 35%, according to Forrester.
If you’re looking for examples, Looker and Databricks are great ones. Snowflake is another, though they’ve done so well that they’re now a cloud platform themselves (built on the existing cloud platforms!).