Avoid Hotel California
I find cloud economics intriguing. It’s new, yet old school (time-sharing of yesteryear — https://en.wikipedia.org/wiki/Time-sharing). Today, economics are driving the growth engine of companies, as they can start small with a small budget and build something magical. The average lifespan of a company on the S&P 500 is shrinking. “In 1965, the average tenure of companies on the S&P 500 was 33 years. By 1990, it was 20 years. It’s forecast to shrink to 14 years by 2026,” as an article from Innosight states (https://www.innosight.com/insight/creative-destruction/). We’ve labeled many of these companies “Unicorns,” but the foundation of many of them are built on top of those classic cloud economics. Moving finances from capital expenditures to operating ones.
At the same time, the economics can be so appealing during a companies growth stage that it becomes Hotel California when their growth stagnates, “you can checkout anytime, but you can never leave.” An article published by Andreessen Horowitz (https://a16z.com/2021/05/27/cost-of-cloud-paradox-market-cap-cloud-lifecycle-scale-growth-repatriation-optimization/) documents this idea of cloud repatriation and the unit economics of pursuing such action. They use Dropbox as a prime example of how moving OFF the cloud can save a tremendous amount of overhead/cost of revenue.
The end of the article spells out 5 key characteristics to keep in mind:
- Cloud spend as a KPI
- Incentivize the right behaviors
- Optimization, optimization, optimization
- Think about repatriation upfront
- Incrementally repatriate
These five characteristics are so vital that I believe they should be added as the 13th factor in the 12 factor apps (https://12factor.net/). I’ll start a motion :P Cloud spend as a KPI is imperative and I find the most critical. I find more companies longing for this insight and yes we need to improve upon these within IBM Cloud.
For a long time, we’ve fought Gartner in their definition of “cloud.” Our take is that anything -aaS could be defined as a pseudo cloud, including Bare Metal-aaS. Gartner historically has not defined bare metal as cloud. Well, let’s stop fighting Gartner and move forward with this idea of cloud repatriation. There are few vendors that can meet the scale, network and abilities of IBM Cloud to deliver Bare Metal as a Service. As companies start to realize they are “over their skis” for their extreme Cost of Revenue, a great middle-ground is IBM Cloud. A customer can automate provisioning and deployment of computing resources, while having the balance of “colocation” cost. Although IBM Cloud will not be able to beat the cost of true colocation, IBM Cloud is a happy medium to cross the divide between Cloud-Native and Colocation. Let’s start with incremental repatriating, a POC to understand what works and what doesn’t on a less critical application. If you need funding…ask. Make it happen and best of luck!