Why PMs Must Understand Business Models
Many PMs treat the business model as somebody else's problem. Sales and finance figure out how the company makes money; the PM just builds product. This is wrong, and it is one of the most career-limiting blind spots in product management. The business model dictates what features matter, who the user actually is, what success looks like, and which trade-offs make sense. A PM who does not understand the model their product runs on is making decisions in the dark.
We have watched PMs ship features that delighted users but destroyed unit economics. We have watched PMs reject features that customers were willing to pay extra for, because they did not realise the company's pricing structure rewarded those features. The connection between business model and product decision is direct and constant. This article describes the major models, the trade-offs each forces, and the implications for daily product work.
The Major Software Business Models
Software companies make money through a relatively small set of models, each with distinct economics, customer dynamics, and product implications. Most companies use one as a primary model and may layer a secondary one on top.
1. Subscription
The customer pays a recurring fee, typically monthly or annually, for ongoing access to the product. This is the default model for modern B2B software (often called SaaS) and increasingly for consumer products too. Revenue is predictable, customer relationships are long, and the focus is on retention and expansion.
Product implications: every feature that increases the chance a customer renews next year is valuable. Every feature that risks churn is dangerous, even if individual users would enjoy it. Onboarding matters enormously because the first month of usage predicts long-term retention. Product-led growth motions tend to fit naturally with subscription, because users can self-serve and expand over time.
2. Transactional / Pay-per-Use
The customer pays each time they use the product, or per unit of consumption. Common in payment processors, cloud infrastructure, AI APIs, and parts of the marketplace economy. Revenue scales directly with usage.
Product implications: features that increase frequency of use are revenue events, not just engagement events. Friction in any transaction directly costs revenue. Pricing is often complex (multi-dimensional, tier-based) and explaining it clearly is a product feature. Low-usage customers may be unprofitable to serve; segmentation matters.
3. Marketplace / Platform
The product connects buyers and sellers, taking a cut of transactions between them. Examples include ride-sharing, freelancer platforms, e-commerce marketplaces, and app stores. Revenue depends on the volume of transactions and the take rate.
Product implications: liquidity is everything. Below a certain level of supply and demand on each side, the product does not work, regardless of feature quality. Cold-start problems are severe. The PM must think about both sides of the marketplace simultaneously, often making one side dramatically better off to attract them, and balancing trust and safety to prevent the marketplace from collapsing into low-quality interactions.
4. Advertising
The product is provided to users free or at low cost, and revenue comes from advertisers who pay for access to those users. Common in consumer media, social platforms, and search. The user and the customer are different people.
Product implications: features that increase user time on the product or increase user data create more advertising inventory or better targeting. Features that delight users at the expense of attention may be commercially harmful. The fundamental tension between user and advertiser interests has to be managed constantly, and many famous product controversies arise from this tension.
5. Freemium
Not strictly its own model, but a hybrid where the basic product is free and revenue comes from upgrades to paid tiers. Common in consumer apps and product-led B2B software. Revenue depends on free-to-paid conversion rate.
Product implications: the free tier must be valuable enough to drive adoption but limited enough to motivate upgrades. Designing this boundary is one of the hardest product problems in software. The features that distinguish free from paid are deeply strategic and shape adoption, conversion, and word-of-mouth dynamics.
6. Licence / Perpetual
The customer pays once for permanent rights to use the software. This was the dominant model in the 1990s and early 2000s and is now largely replaced by subscription, except in some enterprise and gaming contexts. Revenue is large per sale but does not recur reliably.
Product implications: post-sale features and improvements are less directly tied to revenue. The economic incentive to invest in existing customers is weaker, which often produces abandonment of older versions. Most modern PMs will not work in this model directly but should understand it because some enterprise customers still expect it.
7. Hardware-Plus-Software
Revenue comes from selling physical devices, often with software either bundled (sold once with the hardware) or recurring (subscriptions or services tied to the device). Common in consumer electronics, IoT, and certain enterprise infrastructure.
Product implications: hardware constraints shape software decisions. Iteration cycles are slower because hardware does not iterate weekly. The product must work for years, often without continuous updates. Software designed for hardware products is more conservative and more durable than typical web software.
Unit Economics: The Number That Underpins Everything
Whatever the model, every business eventually depends on its unit economics: the relationship between what it costs to acquire and serve a customer and what that customer is worth over time. PMs who understand unit economics make better decisions than PMs who do not, because most product trade-offs have implications for at least one component of the unit economics equation.
Customer Acquisition Cost (CAC)
What does it cost to acquire one new customer, on average? Includes marketing spend, sales effort, content, partnerships, and any other investment that drives acquisition. Lower is generally better. CAC is the denominator of efficiency.
Customer Lifetime Value (LTV)
What is one customer worth in revenue, less the cost to serve them, over the entire time they use the product? In a subscription model, a customer who pays one hundred dollars per month and stays for thirty months has gross lifetime revenue of three thousand dollars. Subtract the cost to serve them, and you have LTV.
The LTV-to-CAC Ratio
If you spend more on acquiring a customer than you eventually earn from them, you lose money on each one. Most healthy software businesses target an LTV-to-CAC ratio of three or higher, meaning each dollar spent on acquisition produces at least three dollars of customer lifetime value. Below three, the business struggles. Above five, you may be under-investing in growth.
Payback Period
How long does it take to earn back the cost of acquiring a customer? Twelve months or less is healthy; eighteen months is tolerable in most B2B contexts; longer than that strains cash flow and limits growth speed.
Metric Healthy Concerning Crisis
LTV/CAC > 3 1.5 to 3 < 1.5
CAC payback (B2B) < 12 months 12 to 18 months > 18 months
Net revenue retention > 110% 90% to 110% < 90%
Gross margin (SaaS) > 75% 60% to 75% < 60%
These benchmarks vary by category and stage but are useful rough guides. PMs should know their product's actual numbers and how their decisions affect them. A feature that improves retention even slightly may be worth an enormous amount in LTV terms; a feature that increases churn is correspondingly expensive even if it sounds nice in a meeting.
How the Model Shapes the Product
The model affects almost every product decision, often in non-obvious ways. Here are the most important effects.
What "Engagement" Means
In subscription software, engagement matters because it predicts retention. In transactional software, engagement matters because each engagement is a revenue event. In advertising, engagement matters because it produces inventory. The metrics may be similar; the underlying logic is different. PMs should be clear on which logic applies to their product.
Who the Customer Is
In subscription B2B, the customer is often a buyer (someone with purchasing authority) who is different from the user. In advertising, the customer is the advertiser, not the user. In marketplaces, there are two customers (buyers and sellers) who may have conflicting interests. The PM must build for the user while serving the customer, and this navigation is much of the job.
What "Success" Looks Like
A new feature in a subscription product succeeds if it improves retention or expansion. The same feature in a transactional product succeeds if it drives more transactions. The same feature in an advertising product succeeds if it increases inventory or improves targeting. PMs who use one definition across all contexts will misjudge half their bets.
Friction Tolerance
Subscription products can tolerate more onboarding friction because users have committed to a relationship. Transactional products cannot, because friction in a single transaction loses the revenue. Advertising products cannot tolerate friction in the consumption flow but can tolerate it in side activities. These differences should shape design decisions.
Choosing the Right Model
Most products do not get to choose their business model in the abstract; the model is implied by the category, the customer, and the value delivered. But there are still strategic choices, especially for new products. Some questions that shape the right choice.
How Often Will Customers Use It?
High-frequency products tend to fit subscription or advertising. Low-frequency, high-stakes products often fit transactional (pay when you need it) or licence. A product used once a year is hard to subscribe to; a product used daily is hard to charge for transactionally without nickel-and-diming.
How Much Are Customers Willing to Pay?
Different models support different price points. Enterprise subscriptions can support thousands per seat per year. Consumer subscriptions usually max out around twenty dollars per month per user before resistance. Advertising allows the product to be free at the surface, which lowers the user's barrier but constrains average revenue per user.
Who Has the Budget Authority?
If the user is the buyer, simple subscription works. If the user is different from the buyer, you have a B2B sale and the model gets more complex (usually annual or multi-year contracts, negotiated terms, procurement processes). If neither the user nor an obvious buyer wants to pay, advertising or marketplace models may be the only option.
How Sticky Is the Solution?
Products that integrate deeply into a workflow have higher switching costs and can support subscription. Products that are easily replaced are vulnerable to subscription churn and may be better suited to transactional pricing. Marketplaces depend on stickiness through liquidity, which takes time to build.
Switching Models
Companies sometimes change their business model. The transition is hard. Customers acquired under one model may not accept another. The internal organisation built for one model may not function under a different one. We have seen companies pivot from licence to subscription, from advertising to subscription, and from product-led to sales-led. Each transition takes years and produces disruption.
If considering a model change, key questions: which customers will accept the new model and which will not? What is the transition path for existing customers? What does the organisation need to look like to support the new model? Most model changes that fail do so because the company underestimated the organisational change required.
Common Mistakes Around Business Models
Mistake One: Ignoring the Model When Designing Features
Designing features without thinking about how they affect revenue, churn, or expansion is the most common mistake. The fix is to make the connection explicit in PRDs: this feature improves the model by X. If you cannot complete that sentence, the feature may not be worth building.
Mistake Two: Choosing the Trendy Model Over the Right One
Subscription has been fashionable for over a decade. Many companies have adopted it because it is fashionable, not because it fits their product or customers. The fit matters more than the trend.
Mistake Three: Confusing Revenue With Profit
A common mistake especially in early-stage companies is celebrating revenue while ignoring whether the unit economics are healthy. A million dollars in revenue with negative gross margin is worse than a hundred thousand dollars with healthy margin. Revenue without profitability is a treadmill.
Mistake Four: Misaligned Pricing and Value
If your pricing is per seat but the value scales with workflows (or vice versa), customers will engineer around the mismatch in ways that hurt your revenue. Pricing should track the dimension along which value is delivered. Misalignment is a constant source of leakage.
Mistake Five: Ignoring Lifetime When Optimising for Acquisition
Many marketing-led organisations optimise CAC down without watching LTV. The customers acquired cheaply often have low LTV, leaving the LTV/CAC ratio unchanged or worse. PMs should watch both numbers and the ratio between them, not just the more visible CAC.
A Final Word
PMs who treat the business model as fundamental to their work have a structural advantage. They make decisions that strengthen the model rather than fight it. Their roadmaps explain how each feature contributes to the underlying economics. Their stakeholder conversations are richer because they speak the language of the company's business, not just the language of features.
If you take one suggestion from this article, take this: spend an hour with your finance team this month understanding your company's unit economics in detail. Understand CAC, LTV, retention, expansion, and gross margin specifically. Then spend one hour per quarter revisiting these numbers. The investment is small. The compound effect on your decision quality is large.
Key Takeaways
- The business model determines what features matter, who the customer really is, and what success looks like. PMs ignore it at their peril.
- The major models are subscription, transactional, marketplace, advertising, freemium, licence, and hardware-plus-software. Each has distinct economics and product implications.
- Unit economics (CAC, LTV, payback period, retention) underpin every business model. PMs should know their product's actual numbers and how decisions affect them.
- Different models reward different product behaviours. The same feature can be valuable in one model and harmful in another.
- Switching models is hard and usually takes years. Be cautious about layered models; pure models are easier to operate.