Designing Business Architecture: Processes, Measurement, and Realization
This is post 14 of 23 in a series on Systems Thinking: Managing Chaos and Complexity by Jamshid Gharajedaghi (ISBN: 978-0-7506-7973-2).
In the previous post, we looked at the framework for designing business architecture: inputs, outputs, markets, and why you need multiple dimensions working together. Now we’re getting into the mechanics. How do the pieces actually interact? How do you measure whether the design is working? And how do you go from a dream design to something that exists in the real world?
Internal Market Economy: Stop Pretending Your Departments Aren’t Trading
Here’s the problem with most organizations. You have shared services like HR, IT, or manufacturing. Multiple business units need them. But those services run on a fixed budget handed down from above. The people actually using the service have zero influence over it.
Think about what that creates. If the service provider is too agreeable, everyone just piles on requests because nobody’s paying for them directly. Costs explode. If the provider is difficult or slow, the business units start building their own shadow versions of the same service. Now you have duplication everywhere.
Gharajedaghi’s answer is simple and kind of brilliant. Turn the whole organization into an internal market. Every unit becomes a performance center with a variable budget. Input units (like manufacturing or engineering) sell their services to output units (business units). Output units pay for what they use.
This changes the power dynamics completely. When a business unit pays for engineering time, both sides suddenly care about efficiency and quality. The helpless internal user becomes a real customer with purchasing power.
The part that makes it real: units should be able to buy from outside the organization too. If internal engineering is overpriced, the business unit can go to an external vendor. That keeps internal providers honest. If management wants to override that choice, they have to cover the cost difference themselves.
The idea comes from Jay Forrester and Russell Ackoff. It sounds obvious once you hear it but rarely gets implemented because it threatens centralized power.
Processes: Planning That Actually Works
Most planning falls into two categories, and both are broken.
Reactive planning looks backward. Something is wrong, so you fix it. But organizations are systems. Fixing each part separately often makes the whole thing worse.
Proactive planning looks forward. Forecast the future and prepare. Also broken, because the future is shaped by what everyone does. Your plan plus your competitors’ plans plus market changes creates the future. Forecasts are “chronically in error.”
The alternative is interactive planning. Design the future you want and figure out how to make it happen. The planning and control system sits at the top of the architecture, managing interactions between all units and setting the vision.
Two requirements keep the design grounded. Technological feasibility: every technology in your design has to exist right now. Operational viability: the design has to survive in the current environment. To assess viability, you need a measurement system.
Measurement System: You Become What You Measure
This section might be the most important in the entire chapter. Gharajedaghi drops a line that deserves to be framed on every executive’s wall: “Winning is fun. But to win, one has to keep score. And the way one keeps score defines the game.”
The measurement system has two parts: performance criteria (what you measure and why) and performance measures (how you actually calculate it).
Getting the Criteria Right
Most organizations obsess over accuracy of measurement. But Gharajedaghi argues that relevancy matters way more. We struggle to accurately measure what we actually care about. So we end up caring about whatever we can accurately measure. That’s backwards. An approximate measure of the right thing beats a precise measure of the wrong thing every time.
Here’s why this matters. Business viability is an emergent property. You can’t measure it directly. You can only measure its symptoms: growth, ROI, net present value, whatever. But using a single symptom as your scoreboard is dangerous. A company can grow fast through bad acquisitions and still be dying.
You need multiple dimensions measured together. Concern for people combined with concern for production gives you something completely different than either one alone.
The Supermarket Disaster
Gharajedaghi tells a story that perfectly illustrates how bad measurement destroys organizations. A supermarket chain looked at its stores and found that 10 weren’t covering their share of allocated overhead. So they closed those stores. But closing them didn’t reduce the overhead. The remaining stores just had to absorb a bigger share. Now more stores looked unprofitable. So they closed more stores. The company was gradually pulling itself out of the market.
The fix was simple. Each store became responsible only for its own operations. The surplus went to the corporate office as income. Corporate became a profit center responsible for managing overhead within that income.
The lesson applies everywhere. When police departments face cuts, they fire cops. When schools face cuts, they cut teachers. Reducing operating units doesn’t reduce overhead. It crushes the system under a growing burden until it collapses.
The Viability Matrix
Gharajedaghi provides a matrix for picking what to measure. One axis covers structure, function, environment, and process. The other covers throughput (producing output), synergy (managing interactions), and latency (creating the future). Where they intersect, you get specific variables: capacity utilization, profitability, demand reliability, marketplace credibility, and more.
The reward system sits on top. It assigns priorities to specific variables to steer behavior. Want people to launch new products in new markets? Multiply those revenues by a bonus factor. People follow incentives. Design them right and behavior follows.
Realization: Successive Approximation
So you’ve designed your ideal business architecture. Now what? You can’t just flip a switch. The design was created as if the old system was destroyed overnight and you could start fresh. Reality doesn’t work like that.
Realization is about bridging the gap between the ideal and the possible. And it happens through successive approximation. You identify everything that stands between you and the ideal, then work through it in stages.
Gharajedaghi says this is the one moment where nothing should be held back. Every reservation, every concern, every “I have a bad feeling about this” needs to come out.
Three Types of Constraints
Type I constraints can’t be removed right now. The design has to be modified around them, creating a “target design” as the first approximation. You keep monitoring, and as constraints disappear, you update the target to get closer to the ideal.
Type II constraints can be removed but need serious work. Redesigning products, reworking processes, building new measurement systems. Resource-intensive. The measurement system redesign usually falls here and is typically the hardest part.
Type III constraints are behavioral. Fear, mistrust, apathy, self-imposed cultural limitations. These could theoretically be removed right now if people chose to. The work is selling the idea, removing resistance, and building trust.
Dissolving the Second-Order Machine
The “second-order machine” is the set of deep cultural assumptions that keep an organization stuck. The invisible operating system running underneath everything.
Dissolving it means making hidden assumptions explicit, then diagnosing which ones are part of the “mess.”
Gharajedaghi gives a healthcare example. Assumptions like “nurses report only to nurses” and “integration means uniformity” were at the core of the system’s dysfunction. Not written policies. Just things everyone believed. Getting people to openly discuss these assumptions was the first step toward breaking free.
You can’t just tell people their assumptions are wrong. They have to discover it themselves through honest dialogue. When that happens, people see themselves and their organization in a new way and gain the ability to choose a different future.
What Sticks
Three things from this half of the chapter stay with me:
First, the internal market concept. It’s elegant because it uses existing economic logic to solve organizational dysfunction. Stop allocating overhead. Start treating internal services like actual transactions.
Second, the measurement warning. We become what we measure. Get the measures wrong and you’ll run yourself right into the ground. The supermarket story is a cautionary tale everyone should hear.
Third, successive approximation. Perfect is the enemy of done, sure. But Gharajedaghi goes further. He says the ideal design is a compass, not a destination. You move toward it in stages, continuously, and you never stop. That’s not a one-time transformation project. That’s a way of operating.