Wednesday’s news about Google shuttering its much-hyped IoT core service came as a big surprise to many. And while it is surprising that an amazing company like Google hasn’t found the IoT services traction that it expected, it isn’t surprising that services continue to come, go, and change on all public cloud platforms. And it reminds us that, if we assume our cloud platforms are just one-way, perpetual streams of linear expansion and improvement in functionality, cost, performance, or other variables, we’re in for some potentially painful surprises.

Remember, we users think of the cloud platforms as our “production” environment and no doubt, companies like Microsoft, Amazon, and Google provide numerous production-ready, enterprise-grade services. But if you “zoom out” a little from those individual services, one can’t ignore the fact that the major cloud providers are using their platforms for continuous market experimentation and validation.

The Cloud is a Laboratory, and It Should Be

Of course, they’d be stupid not to. In some ways, your organization’s use of services on a platform like GCP amounts to a “digital focus group” where cloud providers can gather enormous data about how services are used and by what kinds of companies. It’s a modern, digitized version of a Coca-Cola brand manager doing consumer taste-tests on a new flavor extension, but at cloud scale with cloud precision and velocity.

That’s really useful, especially in dynamic environments like our evolving global economy where companies’ priorities in August of 2022 are far different than they imagined in August of 2021. And thanks to cloud delivery, Google can put new capabilities in its users’ hands for testing and validation literally “overnight.”

But consider the challenge for a Google Product Manager. Technology is changing quickly along with the business environment. If the Product Manager never introduces a new feature or service until there’s ample evidence of sustainable market acceptance, she’ll be uncompetitive and irrelevant, beaten by other platforms that move faster. At the same time, if she moves quickly to introduce a service that some customers are interested in, without complete market validation, she runs the risk of having to deprecate the service, rearchitect it, or merge it with something else in a non-backward-compatible way, etc.  See Google’s IoT Core…

As a product person, I feel empathy for the product managers who put the “BETA” designation on their services to lower expectations for brand-new services, because they’re suddenly introducing a product to millions that may not be exactly what the market needs. Don’t forget, even gmail was in “beta” for over 5(!) years.

Every major cloud provider has this opportunity and faces this challenge, and Google has done an admirable job being transparent about its approach to things like backward-compatibility in a very customer-centric way.

So let’s be thankful that providers like Google “lean in” on introducing new services for us,  and stay mindful that every new service should be considered an experiment at some level.

So How Do I Future-Proof if Everything Will Keep Changing?

For the foreseeable future, the only way to have a totally future-proofed configuration is to go on-prem, and that’s not viable long-term. So we’re going to run our businesses on these dynamic platforms, and as much as we try to standardize on one, service availability and functionality will likely require most enterprises to have a secondary cloud platform, if not multiple.

Your cloud functionality needs will evolve in line with your business strategy and that may mean that major functionality for analytics, IoT, AI/ML, Blockchain, and other key investment areas will be more important or less important at various phases of your cloud and business journey. Regardless of functional needs, what won’t change is your need to deliver both high performance and manageable costs. Silk can’t fix it if your cloud provider deprecates a service. But what we do for companies like Priceline, eToro and others, is future-proof their ability to manage performance and cost through platform, team, or workload changes.

Grab your lab coat and join the experiment. The evolution of public clouds is going to be a fun ride for years to come and will bring far more innovation and efficiency than surprises like this one. But changes and surprises will happen, and we should acknowledge that that is a “feature” not a “bug” in our cloud providers’ strategies.