What Iterative, Agile and Composable Approaches Really Mean in Practice

There’s been growing interest in modular and composable approaches to supply chain innovation but much of the language still feels abstract. In conversations with operators, the same questions keep coming up: What does this actually look like in practice? Where do you start? And how do you avoid ending up with more complexity rather than less?

This article is an attempt to bring the conversation down to earth. It draws on what we've been hearing, reading and exploring through SCL-X and aims to offer a practical way of thinking about what composability means on the ground - especially for those working inside large, complex supply chains where change is hard-won.

From abstract to applied

When people talk about modular or composable approaches, they often mean breaking large systems into smaller, replaceable parts, blocks or Lego bricks. In supply chain terms, that might look like:

• Swapping a forecasting model without changing your whole planning platform

• Plugging in a third-party supplier risk tool without a full integration project

• Using a lightweight microservice to handle approvals or onboarding

Iterative and agile approaches just mean starting small, testing fast and embedding what works rather than committing upfront to big programmes. This doesn’t mean avoiding strategic thinking. It means designing innovation in ways that allow course correction and reduce the cost of being wrong.

Together, these ideas are gaining traction as a response to a familiar set of challenges: slow projects, vendor lock-in, over-customised systems and the difficulty of making any change without breaking something else.

Where it often starts

Most organisations don’t get to start from scratch. That’s why the best use cases for modular approaches tend to be focused and incremental. Some areas where we've seen this show up include:

• Forecasting and demand planning - piloting new models for specific segments

• Inventory optimisation - layering in simulation tools or digital twins

• Supplier onboarding - using simple front-end tools that feed core systems

• Process visibility - adding process mining or shared dashboards

• Workflow automation - creating small services to speed up approvals or alerts

In these examples, the point isn’t to replace everything. It’s to introduce capabilities that are easier to test, evolve and, if needed, swap out. Over time, this creates a kind of capability stack that better reflects your actual operating model.

A practical roadmap

There’s no single path here but a rough pattern does emerge. We’ve been using a simple crawl-walk-run model that aligns with three core goals supply chain leaders keep returning to:

Goals:

1. Improving resilience and continuity
2. Driving cost and asset efficiency
3. Strengthening customer and service performance

But just as important as the pace or sequence of change is having a clear ‘north star’ - a shared view of the operating model you're building toward. Without this, efforts can easily fragment into disconnected pilots or vanity projects that never scale or stick.

It’s also worth noting that foundational enablers - especially a shared data orchestration layer and a degree of digital and data fluency within teams - make this kind of roadmap far more achievable.

Crawl: De-risk and stabilise today

• Focus on where change feels hardest often due to complexity, integration issues or fear of disruption

• Start cleaning and connecting key data sets

• Try simple tools that add value without heavy integration

• Begin building digital literacy within teams

Walk: Optimise and modularise tomorrow

• Introduce composable capabilities where change needs to move faster

• Run parallel pilots - forecasting, analytics, planning - without touching the core

• Put in place a data orchestration layer to help tools talk to each other

• Prioritise interoperability and reusability in how solutions are assessed and deployed

Run: Redesign for advantage

• Re-architect capabilities around your desired operating model

• Build internal capacity to create, integrate or evolve tools

• Establish innovation governance and risk-sharing partnerships

• Shift from fixed workflows to more dynamic, event-driven operations

We’ve mapped some common innovation topics across these stages and goals in our Innovation Grid below. It’s not exhaustive but it’s helping us organise discussion and spot where different vendors or case studies fit.

What about MACH?

Some of this thinking overlaps with MACH (Microservices, API-first, Cloud-native, Headless). While the terminology comes from IT and digital commerce, the principles are increasingly relevant in supply chain. You don’t have to go “full MACH” to benefit. But the emphasis on interoperability, modularity and pace of change aligns closely with what many operators are looking for - especially those feeling boxed in by rigid platforms.

Cyber risk: both sides of the ledger

Modular systems sometimes raise questions about cyber risk. It's a fair challenge. More moving parts can create more entry points. But the flip side is also worth noting: greater modularity allows faster patching, more targeted isolation of breaches and better ability to evolve security policies in step with capabilities. A tightly coupled monolith can create the illusion of control while limiting your ability to respond. As with most things in composability, it’s less about inherent good or bad and more about how it’s approached.

Signs you're heading in this direction

Based on what we’ve seen and heard, you’re probably already on the path if:

• You’re trialling solutions in focused areas rather than rolling out everything at once

• You’re avoiding full replatforming and looking for ways to layer on or plug in

• You’re prioritising capabilities that reflect your actual decision model and data structure

• You’re empowering teams to experiment within guardrails

None of this requires a full architectural redesign. Often, the real shift is in how change is approached: moving from linear projects to reusable components and from fixed end states to iterative learning.

Innovation GRID: by Goal and Stage

Cross-Cutting Enablers (All Goals)

JP Doggett