The Three Stages of Cloud Maturity

What your value stream says about you

A friend once told me, “Enterprises only get cloud right on the third attempt”, which has proved to be something of a truism in the consulting world. But why is this?

In the early days of the cloud, an oft-spoken anti-pattern dominated enterprise cloud adoption: “data centre in the cloud.” Nowadays, most businesses have realised this folly, and while they are not making this mistake, the promise of the cloud as an innovation engine doesn’t seem to hold.

You can better understand an organisation’s cloud maturity by examining its value streams rather than its estate and seeing what percentage falls into the three following distinct categories.

Cloud-Naive Value Streams

In the first instance, enterprises lift and shift not only their applications but also their value streams from on-premise to the cloud. The way they go from idea to development to value doesn’t change.

This approach is the inherent issue with data centre in the cloud. Change throughput hasn’t increased, lead time hasn’t decreased, and the amount of work in progress has gone and stayed up.

This style of value stream is the most dominant when an enterprise makes its first foray into the cloud, often accounting for upwards of 90% of the estate. When this is the case, the estate acts as an anchor that prevents you from realising the potential benefits of the cloud.

This state of play is most commonly seen when the compelling moment to move to the cloud is due to a data centre contract coming up for renewal. A hallmark of the project-centric mentality of technology as a cost centre, not a core competency.

Cloud-Accelerated Value Streams

The next step in the journey is optimising the value streams to take advantage of core cloud properties, such as elasticity and potential for automation. Local optimisations target pain points; as such, businesses see an increase in throughput and a decrease in lead times.

Two classic examples are the automated set-up and tear-down of integrated test environments and test data management. However, the application architecture doesn’t move from its on-premise roots.

While well-intentioned, this leaves homegrown applications in a state of platform incoherence. Although the principles on which they were designed and created have changed, no effort has been made to challenge old assumptions in the new world.

This incoherence leads to negative consequences such as operational fragility, high total cost of ownership, and large amounts of inertia to change. The second-order effect that stems from this is an overestimation of the cost of building new or rearchitecting old applications, which cascades into continual poorly made trade-offs across the business.

It’s worth noting that cloud-accelerated value streams and cloud-naive value streams look the same from an application architecture perspective. Still, the value streams speak to a difference in intention and level of maturity.

For COTS applications, this is the optimal state. However, evolving to the third stage of maturity has severely diminishing returns. Therefore, consistently offloading the applications to managed service providers is a more practical option.

Cloud-Empowered Value Streams

The current zenith is the cloud-empowered value stream, which has been designed using the principles and inherent capabilities of the cloud. At this point, we are no longer locally optimising; we are globally optimising.

An example is a cloud-centric testing strategy, where we eschew mockful local development cycles and leverage cloud services for maximum fidelity. Below is a talk I gave in 2022 about serverless testing in which I suggested the future of testing was in the cloud, not with local mocks or even against the cloud.

In the intervening time, products such as Ampt have advanced further, showing how to build for the cloud, not just on top of it.

Cloud-empowered value streams unlock the engineering throughput needed to leverage the cloud as the innovation engine it’s purported to be.

Starting The Conversation

Often, teams end up stuck at maturity level 1 or 2, and the wider business cries foul when the overly large AWS bill drops every month. Hopefully, this lexicon around value streams enables you to have more productive conversations about how software is delivered and why that needs to change in a cloud-centric world rather than debating the merits of serverless and Kubernetes.