Hello! Welcome to your sixth lesson in the Architect Track.
In our last session, we focused on two key metrics for evaluating a product's economic viability: Customer Acquisition Cost (CAC) and Customer Lifetime Value (LTV). You learned how to calculate these from raw business data and use the LTV/CAC ratio to assess the health of a business model.
Today, we're going to dissect the "cost" side of the equation even further. We will move beyond just the cost to acquire a customer and explore the fundamental cost structure of a software business. Our learning outcome is to explain the difference between marginal cost and operational cost in a software context.
Understanding this distinction is critical for a systems architect. It explains the unique economic properties that make software businesses so scalable, but it also reveals the hidden complexities, particularly in the modern SaaS world, that can impact long-term profitability. This knowledge is fundamental to designing systems that are not just technically sound, but also economically sustainable.
1. The Building Blocks: Fixed, Variable, and Marginal Costs
Before we can compare marginal and operational costs, we need to be clear on the basic definitions that economists use. The best way to grasp these is to see them in action.
Fixed, Variable, and Marginal Cost.
The Khan Academy video, 'Fixed, Variable, and Marginal Cost,' provides a great practical introduction to these concepts using a very relevant example: hiring software programmers to write code.
Please watch the following segments of the video: Fixed Costs (00:41 - 01:39): Notice how costs like office space and a product manager's salary don't change with the number of programmers. Variable Costs (01:39 - 03:00): See how the total cost of programmers' compensation scales directly with the number of programmers hired. Total Costs (03:00 - 04:28): Understand that total cost is simply the sum of fixed and variable costs. Marginal Cost (09:17 - 10:54): This is the most important part. Pay close attention to how marginal cost is defined as the incremental cost to get the next unit of output (in this case, more lines of code).
To summarize the key definitions from the video:
- Fixed Costs: Expenses that do not change regardless of the level of production. (e.g., rent, manager's salary).
- Variable Costs: Expenses that change in direct proportion to the level of production. (e.g., raw materials, programmer salaries in the video's example).
- Marginal Cost: The additional cost incurred to produce one more unit of a good or service.
This image provides a simplified, static view of a marginal cost calculation, reinforcing how fixed costs stay constant while variable costs increase with production.
2. Marginal Cost in the Unique World of Software
Now, let's apply this concept to software. This is where things get interesting and depart from the physical world you might be familiar with from your mechanical engineering background.
The Marginal Cost of Software Approaches Zero
To frame this, let's start with the article 'The Marginal Cost of Software Approaches Zero' by Chris Kluis. It offers a concise definition and a clear contrast between software and physical goods.
Please read the first section, up to 'So let’s dive into the contentious areas'. Focus on the definitions and the comparison between producing Microsoft Office and producing a pair of jeans.
As the article highlights, the cost to design software is very high (a fixed cost), but the cost to replicate it is extremely low. This leads to the famous "zero marginal cost" characteristic of software.
The next article illustrates this with a perfect example.
Do software companies actually have good margins?
The article 'Do software companies actually have good margins?' by Benn Stancil provides an excellent, modern take on software economics. We will use it throughout this lesson.
Read the first two sections of the article, starting from 'Software is expensive to design...' and ending just before 'In practice, the math isn’t quite that good.' The Microsoft Teams example powerfully illustrates the concept of near-zero marginal cost.
As Stancil explains, adding you as a new user to Microsoft Teams costs Microsoft a few cents in server load but generates dollars in revenue. This huge gap between marginal cost and price is the fundamental economic engine of a scalable software business.
3. Operational Costs: Keeping the Business Running
If the cost to serve one more customer is near-zero, where does all the money go? The answer lies in Operational Costs (OpEx).
Operational Costs are the expenses a business incurs to stay in business, which are not directly tied to the production of one more unit. Think of them as the cost of running the entire operation.
Let's return to the benn.substack article to see this distinction made clear.
Do software companies actually have good margins?
This next section of the article introduces operating margins and uses a great analogy of a coffee shop to differentiate them from gross margins, thereby defining operational costs.
Read the section that begins 'On the other hand, something else quite clearly doesn’t work.' and ends just before 'I mean, what? Where does all of it go?'. Focus on the difference between the coffee shop's gross margin (profit on the ingredients) and its operating margin (the actual profit after paying rent, scientists, etc.).
In a software context, operational costs typically include:
- Research & Development (R&D): Salaries for engineers, designers (like you!), and product managers who build and improve the product.
- Sales & Marketing (S&M): Salaries for sales teams, marketing campaigns, branding efforts.
- General & Administrative (G&A): Rent, utilities, legal, HR, and finance department costs.
The key takeaway is that while marginal costs scale with each new user, operational costs are about sustaining the organization that supports all users.
To solidify this, let's watch one more short clip.
Consumer Startup Metrics | Startup School
This clip from a Y Combinator lecture on startup metrics clearly separates the costs tied to a single customer (variable costs) from the overall business costs (fixed costs).
Watch the segment from 10:59 to 13:25. The speaker discusses 'unit economics' for the online bank Monzo. Notice how he explicitly separates 'variable costs' (plastic cards, customer service, fraud) from 'fixed costs' like engineering salaries and office rent. These 'fixed costs' are the operational costs we're discussing.
4. The Architect's View: Where the Lines Blur in SaaS
Here is a summary of the classical view:
| Cost Type | Definition | Software Examples |
|---|---|---|
| Marginal Cost | Cost to produce/serve one additional user. | Incremental server/API usage, data storage for one user, a third-party license fee per seat. |
| Operational Cost | Ongoing costs to run the business. | R&D salaries, office rent, marketing budgets, core infrastructure maintenance. |
However, as a system architect, you need to appreciate a critical nuance of the modern Software-as-a-Service (SaaS) model. The line between these two cost types is not as sharp as it seems.
The final section of the benn.substack article makes a provocative and important argument.
Do software companies actually have good margins?
This is the most crucial insight of the lesson. The author argues that in a subscription world, the 'operational cost' of continuous development is effectively a required cost to keep your customers, blurring the line with marginal costs or cost of goods sold.
Please read the section that starts with 'The uncomfortable answer, I think, is that software margins...' through to the end of the article. Focus on the core argument: because customers expect constant updates, R&D spending isn't just a 'voluntary reinvestment for growth'—it's a mandatory cost to prevent churn.
This is a powerful concept. In the SaaS world, you aren't just selling a static product; you are selling a promise of future improvement. The operational cost of your design and engineering teams is, in a functional sense, part of the cost of delivering the product your customers are paying for today. This challenges the simple "zero marginal cost" narrative and forces a more realistic view of long-term profitability.
Conclusion
In this lesson, we dissected the cost structure of software businesses, moving from the simple idea of "costs" to a more nuanced understanding of marginal versus operational expenses.
Key Takeaways:
- Marginal Cost in software is the cost to produce and serve one additional user, which is famously close to zero. This is the basis of software's scalability.
- Operational Costs are the significant, ongoing expenses required to run the business, such as R&D, S&M, and G&A, which are largely independent of adding one more user.
- The Architect's Nuance: In modern SaaS, the line blurs. Continuous R&D (an operational cost) is often a requirement for customer retention, making it behave like a cost of revenue. Ignoring this reality can lead to flawed economic models.
This understanding of a system's underlying economic structure is a vital complement to your design skills.
In our next lesson, we will shift from economic architecture to data architecture. We'll explore a topic that has huge implications for both user experience and system cost: distinguishing between real-time, near-real-time, and delayed data systems.