Building a Business Model: How Your Company Creates and Captures Value

Startups rarely fail because the product is bad. They fail because the model doesn't close: the value exists but the customer won't pay for it, or pays less than it costs to reach them, or pays only once.

1. Five questions that form the model

Who exactly is the customer? Not "small businesses" but "a retail manager running two to five branches, currently using a point-of-sale system that isn't connected to accounting." The narrower the description, the easier every subsequent decision about product, marketing, and pricing becomes.

What value do you create? It falls into one of three buckets: time saved, money saved or earned, or risk reduced. If you can't place your value in one of these with an approximate figure attached, you're probably selling a cosmetic improvement rather than a need.

How does the product reach the customer? The channel is part of the model, not an accessory. A low-priced product needs a self-serve channel; a high-priced one needs direct sales. Conflict between price and channel is one of the most common causes of model collapse.

How do you earn? Covered in the next section.

What is the cost structure? Which costs do you bear regardless of volume, and which rise with each additional customer? In successful models, revenue grows faster than variable costs.

2. Common revenue models and when they fit

• Subscription: recurring, predictable revenue; suits products used continuously. Requires renewed value or churn rises.
• Commission or transaction fees: suits platforms connecting two sides. Attractive because the customer pays when value is actually realised, but needs volume to become viable.
• Metered usage: ties price to consumption. Fair and flexible, but makes revenue harder to forecast.
• Enterprise licensing: large annual contracts with a limited number of customers. Long sales cycles and high concentration risk.
• Freemium: widens the funnel quickly, but only works when there's a clear natural limit that pushes users to upgrade.

The rule: choose the model that makes the customer pay at the moment they feel the value. Distance between the moment of payment and the moment of value is the source of most collection and churn problems.

3. Unit economics: the decisive test

A model is judged not by its totals but by its single unit. Four measures determine whether it's viable:

• Customer acquisition cost (CAC): total marketing and sales spend in a period divided by new customers in that period. Make sure it includes salaries, not just advertising.
• Contribution margin: revenue from a customer less the variable costs of serving them. This is what actually remains to cover fixed costs.
• Lifetime value (LTV): monthly contribution margin multiplied by the average number of months a customer stays.
• Payback period: how many months a customer needs to return what you spent acquiring them. This is the measure that determines your growth rate, because it determines when capital returns to be reinvested in acquiring the next customer.

The commonly cited healthy indicator is that lifetime value exceeds acquisition cost by a clear multiple, with payback occurring within a period appropriate to your sector's customer cycle. More important than the absolute number is the direction: is it improving over time, or deteriorating?

4. Testing the model before scaling

Test your weakest assumption first, not your easiest. Write down your three riskiest assumptions in falsifiable terms — "customers will pay X per month," "customers can be acquired through this channel for under Y," "customers stay longer than Z months" — then design the cheapest test that proves or disproves each.

Test pricing early, and don't assume a low price makes selling easier. A very low price weakens the signal of value and attracts a segment that churns more and demands more support.

5. When to change the model

Change is warranted when specific signals repeat: customers using the product for a purpose you didn't design it for; high churn despite stated satisfaction; acquisition cost that won't fall however much execution improves; or a sales cycle longer than your cash can absorb.

Change is not warranted when the reason is boredom with execution, a desire to chase a new trend, or a reaction to two weak months. The practical rule: change the smallest variable first — segment, channel, or pricing — before changing the product itself.

Common mistakes

• Confusing the business model with the revenue model. Revenue is one part of five.
• Calculating CAC without salaries. Produces a falsely optimistic picture of the economics.
• Scaling before unit economics close. Doubling spend on a losing model only accelerates the loss.
• Cost-based rather than value-based pricing. Leaves money on the table and ties your price to your internal efficiency rather than customer benefit.
• Serving multiple segments at once, too early. Fragments the product and the message and prevents genuine repeatability.

Checklist

• Target customer described in a specific, targetable sentence
• Value expressed with an approximate figure (time / money / risk)
• Revenue model aligned to the moment value is felt
• Channel consistent with the price point
• Acquisition cost calculated inclusive of salaries
• Positive contribution margin at the level of a single customer
• Payback period calculated and trending toward improvement
• Three riskiest assumptions tested

FAQ

How do I know my model has "closed"?
When you can acquire customers repeatably through a known channel, at a cost below the value they generate, without founder involvement in every deal.

Should I start with a free model to attract users?
Only if you have a clear upgrade path and a natural reason for users to cross it. Free without an upgrade path builds a user base, not a business.

How do I price at the beginning?
Start from the estimated value to the customer rather than your cost, and test three levels on real customers. Price rejection is as informative as acceptance.