Should Cloud Providers Stop Building Features and Start Building for Interoperability?

I wouldn't call it an either / or proposition, but from a balance perspective I would put the ratio of feature development to interoperability development at about 85 / 15. The thinking behind this is twofold - on the one hand, the majority of your users will be leveraging the native UI for accessing the service and will not (immediately) require integration capabilities; on the other hand, the move towards standardization on RESTful or SOAP-based web service API's means that the lion's share of the interoperability architecture is already defined...the service provider just needs to design and implement well-described web services that can be consumed easily and securely (think Salesforce, Freshbooks, Basecamp).

The cloud has enabled long-tail services that provide niche capabilities that were never going to be realistic within the constraints of monolithic, tightly coupled software packages; these services were once only available to companies with budgets for custom development, and now cloud service providers are opening up deep functionality for a much larger audience, and the value being created by this paradigm is exponentially greater than the value of cross-service interoperability.

This is a big quandary for providers. Features allow them to charge money and make a profit. Interoperability allows them to reach the widest possible market.
The more interoperability and the less features you have, the more you're offering a commodity that won't make money.
The more features and the less interoperability you have, the more you narrow your appeal to a narrow, specialist market.
Who'd be a cloud provider?

Interestingly, providers that don't think about interoperability actually get commoditized the fastest. If we look at the cloud infrastructure market, what differentiates Amazon, GoGrid, RackSpace and others? Is it price? Probably not - it tends to be the management control panels. So, it would seem that this extra "feature" becomes the differentiating value add to attract clients. But if these providers focus on delivering better management consoles only for their own services, they will end up getting commoditized even more quickly. There is a whole new genre of uber consoles that are appearing in the market (or are about to appear) that bridge providers seemlessly. These "cross-cloud" consoles will turn many lower-end providers almost instantly into commodity cloud providers. Sound familiar? Just look at the ISP market in the late '90s.

We think it is too early for cloud providers to take this step to stop building features and work on interoperability. At this stage, the platform vendor has very little to gain by providing interoperability. What the platform vendors need to focus on is providing the right set of tools and abstractions and services across the board to attract more customers / data and create a bigger marketplace for all the stakeholders