It's been a couple of decades since Salesforce first brought the idea of enterprise cloud services into the mainstream. Since then, firms have been closing their datacenters, moving away from solutions like Sharepoint, and putting their trust in subscription services that they access through a web browser. Even venerable on-premise tools like Microsoft SharePoint have made their way online.

But to move everything into the cloud is to make the mistake of putting technology before customers. You can't build a product that solves deep organizational problems by simply declaring that the cloud is the future and walking away. You need to actually understand the user holistically, test hypotheses around what you think will solve their problem, and iterate towards a real solution. That includes every aspect of the product, including its technology stack. Our role as technologists is to fight for the user, not to advocate for particular technologies we happen to like. (If those technologies are the best way to solve the user's problem, then great.)

There are cases where the cloud is not the right solution; particularly when security is a consideration, or when users are particularly concerned about their own privacy.

For consumers, we're beginning to see products like the Helm server, which allows you to host your email inside your home. While traditional home servers require a great deal of technical knowledge, you can buy a Helm, plug it in, and get going quickly. You don't need to be a Linux server admin (or pay one). It just works. I'm excited for future iterations of the idea, which I hope will allow you to host your own access-controlled social spaces from your home.

I believe there is a similar need for modular self-hosted software and hardware for businesses that makes it easy to run on-premise applications.

These software applications will be split: the portion that handles user data at rest is hosted on-premise. Meanwhile, a companion API piece sits in the cloud. This way the user receives the best of both worlds: their data is kept safe, and the services surrounding their data are continuously updated and managed for them. This split means that the on-premise software remains relatively thin, keeping updates simple.

Vendors can charge an up-front licensing fee for the on-premise product as well as a recurring subscription for the hosted service. Not every business requires that level of security for their data at rest, or even cares about it; for them, a fully-hosted service is available.

I've got skin in this game already: I designed a product called Hub for Latakoo, a service that allows journalists to quickly send video from the field using commodity internet connections. Hub sits in newsrooms and allows video to be automatically integrated into their content management systems, in the correct format. It's useful for the newsrooms, and lucrative for the company. And I think the model has broad applications elsewhere.

Of course, I didn't invent the hybrid cloud. It's in wide use in larger enterprises, and services like AWS have existing solutions. But these applications are often bespoke. I think there's room to bring it to privacy-minded startups and SMEs - and build a whole new era of privacy-aware business applications in the process.