How to Win in Cloud
Let’s pretend for one terrifying minute that I found myself working for a cloud provider. Setting the problems with that aside, let’s further fantasize that my remit were to be overly broad: to grow the company’s cloud business. Lastly, let’s go fully divorced from reality and say that I would be allowed to operate on achieving this on a 5-10 year time horizon.
I know exactly what I would do; let’s get to work!
An Overnight Success in Ten Years
On a longer planning horizon than many companies seem to operate, it would appear that the approach that most of the existing cloud providers are taking today is suboptimal. Their messaging, their product offerings, their enhancements around existing services, and their hiring all swirls around a central theme: migrating enterprise workloads into cloud.
I get why this is true. I’m unsure that I’ve seen much discussion about why this isn’t a good long term play.
Today, big enterprises moving to cloud are where the money is. They spend huge amounts on their IT estates, and they don’t migrate lightly. There will inevitably come a day when most of them have either migrated, or decided firmly that they will not migrate.
Therefore, that market has an inherent shelf life, and is ultimately one in decline well within most of our lifetimes.
Consider also that the average tenure of a company on the S&P 500 index is just over 21 years. Bearing that in mind, it therefore tracks that a decent portion of 2050’s S&P 500 have not been founded yet.
And lastly consider that the things startups are doing today is what large enterprises will be doing in a few years. We’ve seen that trend accelerate; easy examples range from “developing on Macs” to “CI/CD” to “using cloud” to “launching ridiculous NFT scams.”
If you’re interested in being where the market is going rather than in charging to where it currently is, it therefore tracks to make your cloud provider appealing to startups.
My job becomes clear. Whether or not I believe that existing “legacy” businesses are a market in decline or not, it’s apparent that the way to win the future is to capture the startup market. The best thing about a new company is that there’s no technical debt; ergo it’s not a migration, every application is all purely net new.
Let’s Start a Company
Okay. I have registered TwitterForPets.com, told some good stories to well-monied friends, and hired a small team. Today, which cloud provider should I pick?
Everyone, and I do mean EVERYONE has an opinion as to what the right answer should be for this, but the correct answer is incredibly simple: I talk to that small team that I just hired, ask them what they have the most experience with, and then I pick that one. Therefore, it logically follows that if I as the cloud provider want to capture the startup market, I have to be the one that their team is the most experienced with.
Engineers don’t spring fully formed from the forehead of some ancient Greek god like Kubernetes (the god of spending money on cloud services); they come from somewhere. Sometimes that’s the traditional education system, sometimes it’s from being self-taught, sometimes it’s from bootcamps. My goal as a cloud provider is to therefore make myself incredibly compelling as a platform to people operating within those three environments.
The Plan’s Outline
The first item on my list is to fix the platform’s free tier, because it kinda cuts against being “the place to learn” if there’s a steady drumbeat of people just learning your platform getting soaked by surprise bills. Since I wrote that article, the bill surprises have dramatically increased; $70K bill shocks are dangerous to the learner (someone will eventually do something tragic and rash), but also dangerous in a safer way to the cloud provider’s reputation in this market. If someone gets a terrifying bill, they scream for help. People are watching and listening! Their takeaway will be “well I won’t try THAT platform–holy hell!” That has to stop immediately.
Next, I’m going to take a page from Apple’s playbook. In the late 1970s Apple started doing some very strange things. They donated a pile of computers to schools, they provided free hands-on training for teachers, and within a few years most computers in schools were made by Apple. I learned to type in a computer lab full of ancient (even for the time period!) Apple II machines.
I would thus build entire school and bootcamp curricula that teach my cloud’s platform in an ostensibly agnostic way, offer completely free sandboxed environments to people going through those educational pathways, and then make the exact same thing free to anyone who dropped by our marketing website and wanted to learn. Making this stuff accessible to anyone who wants to learn is key; you can’t gatekeep it by restricting the educational tools to registered institutions and bootcamps if you want to dominate the market that’s to come.
I talked about this some in my article The Dumbest Dollars a Cloud Provider Can Make; charging for training and then certification is a mistake.
While it’s tempting to say that the final step is “wait a few years,” there’s a lot to do in the intervening time. I’ll want to continually send my own staff through these programs not only to learn the platform themselves, but also to assess the general feeling throughout the industry. You can’t just launch an initiative like this and not keep in intimate touch with how it’s going over in the market. This won’t work at all from an ivory tower perspective; it’s going to need constant adjustment and work.
Aren’t Cloud Providers Doing This Now?
It’s a good question. You can certainly make the argument that every one of the major providers has at least made motions in this direction. The problem is that those motions are drowned out by the Big Enterprise Migration Story messaging that increasingly is presenting to tomorrow’s builders as squabbling over the companies that they’re planning to disrupt rather than conspiring with them to help build that disruption.
There’s a brighter future here that will look to all the world (and specifically the markets) as if any cloud provider that pursues it is retreating from the field that represent today’s “cloud wars.” Done right, this strategically cedes the battle to absolutely win the war.
It’s going to be a fun decade.