Cloud Deliberate · Infrastructure strategy

Move beyond the autopilot of cloud-first.

Being cloud deliberate isn't pro- or anti-cloud. It's workload-by-workload decisions that keep the benefits of the cloud and give you back control.

04
Guiding principles
05
Tools in the toolbox
05
Decision verbs
07
Control challenges mapped
The shift

From cloud-first to cloud-deliberate.

For more than a decade, cloud-first has been the default choice for speed, flexibility and scale. Today, public cloud is mature, and the landscape keeps moving.

The truth has emerged that maintaining control comes with trade-offs. Misconfigured resources and poor pricing decisions push costs higher than expected. The pace of change demands constant attention to make the right security and financial calls.

Repatriation, a trend since 2018, isn't a retreat. It's a rebalance, part of a multi- or hybrid-cloud strategy. We saw the need to sharpen the focus onto what matters: how to be deliberate in your cloud management.

Principles · How to think

Reference points for a shifting landscape.

Four principles that reduce noise, align teams and keep cloud decisions tied to organisational priorities rather than short-term trends.

01

Cloud-native everywhere

Automation, infrastructure as code, self-service platforms and loosely coupled architectures, wherever the workload runs.

02

Clarity before action

Decisions grounded in an explicit reading of your landscape, constraints and outcomes, rather than habit or hype.

03

Robust, ready to recalibrate

Architectures that tolerate change. Decisions that can be revisited when costs, regulations or requirements shift.

04

Make the right choice easy

Reduce cognitive load. Clear defaults and paved roads point teams toward the preferred path without removing autonomy.

Toolbox · How to act

Turn principles into everyday practice.

  • 01Principle 1

    Automation over manual operations

    Shift operational work from individual actions to shared definitions. Versioned, reviewed, deployed in a controlled way, so the work stays visible, auditable, and reversible.

  • 02Principle 2

    Clarity on starting point, constraints and desired state

    Audit the systems, skills and debt you have. Make regulatory and contractual boundaries explicit. Define the operating model you want before choosing technology.

  • 03Principle 3

    Architectural patterns for hybrid and multi-cloud

    Choose technologies that run across environments. Favour open standards. Keep data portable. Design for reversibility, not for running everything everywhere.

  • 04Principle 3

    Continuous optimisation and review

    Treat cost, performance, security and compliance as ongoing work. FinOps practices and regular review surface change before constraints become entrenched.

  • 05Principle 4

    Paved roads and clear defaults

    A supported, well-understood way of building and operating systems. Deviations stay possible, but explicit and justified.

Framework · How to decide

Five verbs for distributed decisions.

A balanced approach drawn from Andrew Harmel-Law's work on facilitating software architecture, with light additions. Speed and alignment, together.

Step 01
Empower

Authority sits closest to the point of impact.

Step 02
Consult

Input from those affected and from experts before the choice is locked.

Step 03
Measure

Filter by impact and reversibility. One-way doors get more scrutiny.

Step 04
Guide

Principles, standards and tech radars set the guardrails.

Step 05
Document

ADRs capture context, options and reasoning. Distributed becomes safe.

Background

Why control gets difficult in the public cloud.

The reasons companies report losing control. Each one is solvable once it's named.

Escalating, unpredictable costs

Elastic pricing rewards correct configuration and punishes drift. Misconfigured auto-scaling, idle resources and egress fees accumulate quietly.

Security and shared responsibility

Advanced controls, but a moving boundary. Misunderstood ownership leads to data exposure and IAM gaps.

Compliance and data sovereignty

Pre-audited infrastructure helps. Constant platform evolution and overly permissive defaults create regulatory risk.

Performance and latency

Multi-tenancy means noisy neighbours. Hybrid splits add network hops. Data gravity quietly pulls workloads where the data already lives.

Vendor lock-in

Proprietary tools, formats and tightly integrated services create dependencies, and expose you to the provider's pricing and roadmap.

Strategic and geopolitical exposure

Concentration on a handful of hyperscalers makes legal and political risk an architectural concern, not just a procurement one.

Cognitive load

Hundreds of services, two operating models, constant change. The question is not whether complexity exists, but where it lives and who can manage it.

Your turn

Start where it matters.
Move with intention.

Pick what fits: the white paper for a deep read, a quick note for more information, or a lunch or half-day workshop to map your landscape together.

  • Reply within two working days
  • No newsletter, no third parties
  • One form covers all four requests
Step 01 · What would you like?

Send me the full PDF.

We only use your details to follow up. No newsletter, no third parties.