Commercial discipline and architecture
A while back, I delivered a lunch-and-learn on commercial discipline for architects. As usual, I took a different angle. I also had a banking CEO introduce it and lend it credibility. I recently found the notes and thought it was worth sharing as a thought provoker, and not necessarily the answer.
Who this is for: architects at any grade, and the people who lead them, who want their work to keep making a commercial difference after it goes live.
BLUF: Commercial discipline is the habit of keeping architecture connected to business value, from the first conversation to the years a capability spends in service. It depends less on finance than on relationships, early engagement, tested decisions and an honest view of scope and lifetime cost. We neglect what happens after launch, when the business and technical designs start to drift apart.
What I mean by commercial discipline (in this article)
Commercial discipline is consistently keeping your focus on practice, performance and outcome when you approach commercial challenges, so that what you build plays its part in securing the future of the business. The word that matters is consistently. Most architects have the competence; what we lack is the habit of applying it every time, noticing when we have drifted off it, and coming back.
It is not mainly about finance. Commercial acumen matters and architects should take a bigger part in the numbers, but treating cost as the whole of it is too simple a view of how value gets created. Ask a business leader what commercial leadership involves, and you'll hear the same things.
- Staying connected to the context for change
- Knowing the design levers that can be pulled
- Understanding the value they can deliver
- Keeping all of that aligned to strategy and coherent.
That list reads a lot like a description of architecture.
Our pull often goes toward technical excellence, and most writing about architecture decisions reflects that; it focuses on making the best technology choices against principles that describe technical discipline, but very little covers what happens after launch.
When we put an architecture into service, we are in effect saying it will create a certain kind of sustained value for the business. What happens to that promise afterwards is the part we rarely talk about.
Why it matters now
Two things have raised the stakes. First, architecture work now involves more partners, a wider spread of views, and an opportunity to get involved earlier. Reaching a coherent conclusion across that group is harder, and architects get a seat at that table only because of commercial credibility.
Second, most organisations are now digital businesses. In the old supply model, organisations judged a technology decision on whether it delivered a good service within budget. In a digital business, a poor architecture decision hits the business directly, quickly and widely. A sequence is at work: the quality of the commercial context shapes the quality of the thinking, and the thinking shapes the quality of the decisions.
Good decisions let a business pivot, scale, partner, acquire, divest, and compete.
Getting into the room
Guiding the business means going back to the start of the process and keeping one question in view: is what we are doing about business success, and what are we doing to keep that focus?
This is not a job reserved for senior architects. A junior architect can have an "I don't know how this works" conversation with someone in the business, and most business people are glad to help with a learning conversation like that. The harder question for all of us is how often we speak to the real business customer rather than a proxy.
I sense that many architects feel unable to have these conversations. You cannot do this role properly from behind a project team; on a building site, the architect and the engineers are the first ones in.
What a CEO told us
For the session, I interviewed the chief executive of a banking business about architecture and commercial thinking.
He started with who the client is. People in technology roles tend to call the business their client, because that is who they deal with. He thinks that misleads them. Work back from the end client, the retail or corporate customer, and tie what you do to that relationship. Even people who say "I don't generate P&L" find themselves close to the organisation's most commercial conversation. His advice to every team was to use each activity to ask what it does towards that end outcome for the client.
On what commercial means, he was clear that it is not just revenue. In his words, strategic commercial value is "finding the solutions that persist over time through the volatility we see in the market environment". He used payments as the example. Ten years ago, the obvious market reading was that more people would pay by phone. The bigger change was platforms launching in a new market and wanting to make hundreds of thousands of payments a second. An architecture built on the first reading would have looked sensible and still left the bank unprepared for the second; the difference came down to how well the architects understood where the market was going and planned for that scale before it arrived. For an architect, he said, a commercial outcome is future-proofing. He does not claim to know how to set that architecture up; that is where our expertise comes in, and it only works if we are in the conversation early enough to understand what the business is trying to do.
He also flipped the usual view of seniority. On digital client experience, he thought the best-informed people in the bank were the most junior, those who had just joined from university and knew from their own lives what good looks like. He joked that innovation needs the pyramid turned upside down. He added that corporate treasury expectations lag consumer ones by a generation, as today's consumers become tomorrow's treasurers, so that junior insight matters for corporate design too.
His passion was data, and his advice applies to architecture more broadly. Don't wait for a perfect data lake before you try to get value from it; that journey never ends. Pick a use case where an insight might be worth having, catalogue only the data it needs, test it quickly, then build on it. Over time the good data platform arrives as a by-product of ideas you have already commercialised. His example was the early months of Covid, when nobody knew what was happening to consumers. A group of people asked whether card payment flows could show the effect of different state lockdowns on spending, and then whether that could indicate the direction of GDP ahead of the official figures. Nobody could have planned that use case. It came from people with different views, architects and business people together, willing to play with an idea and challenge each other.
His closing message to the architects was to be curious and reach out: "If you have that courage to engage, play a little bit with ideas, and really try and understand, I think it will help at every level." What stayed with me was that he valued our expertise and wanted to engage with it as a peer. The question is how we build more relationships like that one.
Strategic thinking as part of commercial discipline
Commercial discipline needs something to be disciplined about: a clear view of what success looks like for the business, what it depends on and what could undermine it. Strategic thinking is how we build that view and keep testing it, which is why the session paired the two. Architects have a strategic role, and part of it is taking people with them as strategy forms. That is not writing a document and presenting it. It is an iterative process that lets ideas be scrutinised and moves people from intent to inception. Following Roger Martin, I find it useful to treat strategy as a hypothesis about the future, developed by prototyping and by asking the same questions repeatedly:
- What is our aspiration, why does it matter, and why does it represent success
- What must be true for it to be a sound ambition?
- What capabilities and supporting systems does it need, and why?
- What might be true that would make it vulnerable, or put sustainable value at risk?
- How do we tell a credible story that builds consensus?
The process does not stop. The aspiration, the capabilities that enable it, and the things that could derail it all keep moving.
Strategic thinking also happens at every level of decision-making, not only when someone announces a big idea. It happens when we optimise and sustain a capability and look after the levers of its success; when we react to a threat or opportunity; and when we are challenged during delivery and need to create good options quickly. That last one is a small strategy process in its own right, and many architects already do it every day without calling it strategy. Each of these is a commercial moment too; the choice made either protects the value the business was promised or slowly erodes it.
The common objections are that it is someone else's problem, that it only applies to big change, and that it is above our pay grade. None of them stands up to challenge. The real gap is that we do not connect the decisions we make with their effect on the ambition. Owning that connection, and asking how a given decision affects the viability of the outcome we want, is where strategic thinking becomes commercial discipline.
Decisions are the substance of architecture.
As we think strategically, we make decisions, and as we act on them, we make more. Architecture is that sequence of decisions, to do, not to do or to defer, which assembles first the picture and then the solution. The core skill is working out that we need to get from A to B and why, and then, often with engineers and other specialists, working out the best decisions in the best order.
So we need to be on top of our decisions, not only the exceptions. We should capture them and test them continuously: are they feasible, viable, deliverable and inclusive? Feasible covers our capability to build; deliverable, our capability to change; viable, our capability to get value from it.
Inclusive is the one we tend to forget. A decision that leaves a group of stakeholders feeling they got nothing will be undone later.
The most useful sentence I know for this is "if we do these things, in this way, this is the business value".
It forces three things:
- We have to be in the discussion in the first place.
- Our decisions need a line of sight to a clearly expressed business need, which comes from good engagement with the people who can articulate it and a relationship strong enough to allow challenge.
- We need options that bring stakeholders together around a shared view of value, which means translating technical value into business value rather than deciding on technical value alone.
When business colleagues arrive with the solution already formed, dropping in late and telling them it is wrong will not help, and they are unlikely to want to reopen their thinking at that stage. Being curious and asking what is important to them still works. In the longer term, the answer is to be there earlier and work more like a good consultant. Some of the best consultants I have worked with rarely offered me a solution; they asked questions until I could see the problem clearly for myself. A service blueprint is one tool I have found useful here, because it quickly shows what a service is trying to achieve in value and experience terms, and the conversation looks like a solution discussion while staying firmly about the business.
One more point on options. A binary choice is usually a lazy framing; either option will disappoint someone, and the easiest one tends to win. Blending the best of several options is something architects are good at, provided we have principles that say what good looks like. It also helps to seek out the people with the most extreme views and understand them. Those relationships are worth having, and we need to hear their perspectives.
Owning the commercial story
By this point we have made a set of decisions, tested them and built consensus. The next step is the lifetime case for cost and value, and this is often left to finance. Finance can help with the numbers, but they cannot be expected to understand the nuances of risk, value, scope and the robustness of the thinking behind the design. Architects need to own that story and provide the context that holds it together.
Scope is where this goes wrong most often. A change to one capability affects others, and those affect others again; cost and risk accumulate along the way, along with opportunity. A capability model helps here. For each change you can ask which capabilities are being improved directly; which are not changing but are consumed by the primary service and will need more capacity or new procedures; which are enabling services further back; which might be hit in ways nobody expects; and where the change creates an opportunity elsewhere. We are generally good at the first of these and much weaker at the rest.
One example. A programme to replace a policy administration system had already budgeted £175M. We assessed its impact against the business capability model through workshops with the business, looking at how a change in one capability would ripple across the rest. That depended on an architecture and business partnership already being in place. It found considerable scope gaps because the project team had focused on the technology they wanted to implement, not on how other parts of the business would need to come along. It was a significant shift, but it broadened thinking, challenged assumptions, and gave the business more confidence in its scope and in how the investment case addressed the wider impact. If there is no time to do this, the answer is to start earlier, before the wave hits.
The same applies across the lifecycle. A capability has a cost to create and integrate, transition, operate, sustain, extend, and exit, and managing that lifetime means looking in two directions. Looking forward, nobody can know all of those costs at the start of a product that may run for 25 years. We can, though, identify the major inflexion points, such as a move to scale, a major upgrade, a new use for the data or an eventual exit, and ask which forces could change the assumptions we have made. Handled early, those events can sustain value rather than erode it. If we choose an approach that works today but has a clear future of high costs for scaling, upgrade or replacement, it is hard to call that the best architecture.
Looking back matters as much. Every capability was approved on a commercial case: this design, at this cost, would deliver this value. That case is the original commercial architecture, and it is the reference point for every later change. When someone proposes extending the scope, adding data or reusing the service elsewhere, the question is whether the change keeps faith with that case, improves on it or slowly undermines it. Without the backward look, the forward one has nothing to measure against.
Keeping the discipline downstream
This is the part we talk about least.
Once something is built, architecture and engineering don't stop; otherwise, there was little point in the effort. We may have started small and strategic, and now it is time to scale, whether that means breadth across the enterprise, wider integration or more data.
The first problem is goals: people find growth objectives more compelling than goals to sustain what exists, so attention moves to the next future and what we built drifts. Setting good sustain goals, and making them compelling, is part of the discipline. Everything that got us here still applies; stopping it dilutes the value of having got here at all.
The second problem is drift. The business, the data, the software and the technology each move on their own, a little like plate tectonics; they pull apart or collide. Each move tends to sound reasonable on its own: let's make it a group-wide service; I want to add data so another division can use it; I'll use this external service for my data; deploy it on the capacity we already have; I'll put this personal data in the payload. None of that is necessarily bad, but together it can erode the value the design was meant to deliver. The moment to watch for is "that's not what we designed it to do", and the response is to govern capabilities against the original commercial commitment, the principles and the purpose of the design, rather than only governing individual designs.
This is where governance has to change. Once a capability is live, most governance focuses on the technology and good practice: whether the design complies, whether the platform is supported, and whether the standards are met. Very little asks whether the capability is still delivering the value it was approved to deliver, or who is accountable if it is not. Future governance needs to focus on sustained value, which means joining the business and technical views of the same capability. The business knows how its levers are being pulled; architects and engineers know where the design is sensitive to them. Neither view alone can say whether value is being delivered, is still deliverable, or is being impaired.
Technology sits inside a larger system. Change in the market and business model, in the business design and in the operating model is constant and necessary. Those levers drive the measures that matter to the business; in an insurer, that might be combined ratio, expense ratio, premium written, new business and solvency capital. Our job is not to stop the levers moving. It is to understand where our design is sensitive to them, so we can say: if this changes, the commercial architecture changes in this way.
Stewart Brand's How Buildings Learn is the best metaphor I know. Software is not a building; it is closer to the services and space plans inside one, which change constantly. A building has layers that change at different rates: site, structure, skin, services, space plan and stuff. The furniture moves all the time at little cost. Changing the room’s purpose needs planning, investment and some architecture. Over a long life, changing the space plan costs far more than the original build. Sustainable architecture is about designing each layer to adapt at the right pace, making change at the right layer, and knowing whether what you are creating is structure or stuff. Invest in structure and innovate around the stuff.
If business, architects and engineers stay connected to the purpose and value of what they built, we may not need to knock it down and start again every ten years.
Demosthenes put it: maintaining possessions is more difficult than acquiring them. Paul Dourish made the modern version of the point: asking what we might build tomorrow blinds us to our responsibilities for what we built yesterday.
What it asks of architects
Commercial discipline is a set of skills and a mindset that make sure good design makes a commercial difference and keeps making it after launch. It rests on the relationships we build, the conversations we have, and how we bring people together to deliver and sustain value. None of it depends on grade.
In practice, that comes down to a few things anyone can start on:
- Understand the business well enough to have learning conversations with it, and with the real customer where you can.
- Build relationships that let you engage early and challenge thinking.
- Have the technical discussion but always ask how it enables business success and what could threaten it.
- Notice your decisions, anticipate their impact, and check they are feasible, viable, deliverable and inclusive.
- Look wide at scope, including the secondary and tertiary impacts.
- Keep doing good architecture once the thing is live; manage the space plan.
This is hard, which is why it needs discipline.
PS: I get Copilot to do the graphics for these; in analysing the text and after a chat, we agreed the centre of the piece was about:
- Holding one's discipline when circumstances tempt compromise.
- The character required to continue practising the craft properly when shortcuts, fashion, pressure, politics, urgency or reward encourage otherwise.
I couldn’t agree more.
A note on how I write. The observations, questions and arguments are mine; so is the experience behind them. I work with Claude, an AI assistant, on research and drafting, and I edit and check until I can stand behind each statement. Any errors are mine too.