8 Unit Economics Mistakes Founders Make

The most common unit economics mistakes are not complicated errors of math, they are quiet choices that make a business look better than it is: using revenue where contribution belongs, leaving shipping and returns out of the model, trusting blended acquisition cost, and projecting customer lifetimes that nobody has observed. Each one is easy to make, easy to miss in a board deck and expensive to discover late. This article lists eight of them, with a worked example for each so you can check your own model against the arithmetic. It belongs to unit economics and budgeting and forecasting, builds on what is unit economics and unit economics formulas, and is worth reading alongside unit economics for marketplaces.

1. Calculating lifetime value on revenue, not contribution

A customer pays $50 a month, the gross margin on the service is 40 percent, and the average customer stays 20 months. CAC is $300.

• Revenue-based LTV = $50 × 20 = $1,000, so LTV : CAC = 1,000 / 300 = 3.3
• Contribution-based LTV = $50 × 40% × 20 = $400, so LTV : CAC = 400 / 300 = 1.3

The business looked comfortably healthy and is in fact marginal. Revenue is not what you keep. Use contribution, or at least gross margin, in every LTV calculation. See customer lifetime value.

2. Leaving costs out of the per-unit model

A founder builds a per-order model: price $40, product cost $14, fulfillment $4 and shipping $5. Contribution looks like $17 (42.5 percent). The model omits payment fees ($1.20) and a returns allowance ($1.60).

• True contribution = 40 - 14 - 4 - 5 - 1.20 - 1.60 = $14.20 (35.5 percent)
• Overstatement = 17.00 - 14.20 = $2.80, which is 2.80 / 14.20 = 19.7% more than the real number

On 50,000 orders a year, the missing costs are worth $140,000 of profit that was never there. Build the cost list from invoices and statements, not from memory. See contribution margin and the true cost of returns.

3. Trusting blended CAC

A company spends $200,000 on marketing in a quarter and gets 2,000 new customers, so blended CAC looks like $100. But 800 of those customers came from organic search, referrals and word of mouth. Only 1,200 were driven by the paid spend.

• Paid CAC = 200,000 / 1,200 = $166.67
• Twelve-month contribution per customer = $150

On the blended view, each customer returns a $50 profit (150 - 100). On the paid view, each paid customer loses $16.67 (150 - 166.67). Scaling the spend, which is what a founder wants to do with a "profitable" channel, would buy customers at a loss, because new paid customers cost the marginal rate, not the average. Track both numbers, and base scaling decisions on the paid or marginal figure. See blended CAC vs. paid CAC.

4. Projecting lifetimes you have not observed

A subscription business has 12 months of data and monthly churn of 3 percent. The lifetime formula gives 1 / 0.03 = 33.3 months, and with $20 of monthly contribution, LTV = 20 × 33.3 = $667.

But the data cover only a year. In a 12-month window, expected months of subscription for a customer with 3 percent monthly churn are the sum of 0.97 raised to the powers 0 through 11, which is (1 - 0.97^12) / 0.03 = (1 - 0.694) / 0.03 = 10.2 months, or about $204 of contribution actually earned. The projection of $667 assumes that churn stays at 3 percent for nearly three years, that the product does not change, that competitors do not appear and that prices hold. Treat anything beyond the observed window as a forecast, cap the lifetime (many teams use 24 to 36 months), and discount future contribution. Show the observed twelve-month figure next to the projection. See how to calculate customer lifetime value.

5. Ignoring payback and cash

A business acquires 500 customers a month at CAC of $120 and earns $15 of contribution per customer per month. Payback is 120 / 15 = 8 months, and LTV looks great. But cash goes out up front.

Each month's cohort costs $60,000 (500 × 120). Each cohort then returns $7,500 a month (500 × 15). Ignoring churn for simplicity, cumulative spend after 8 months is $480,000, and cumulative contribution is 7,500 × (1 + 2 + ... + 8) = 7,500 × 36 = $270,000. The business is $210,000 in the hole, even though every cohort is profitable. That is the peak funding need (it is the same at month 7: 420,000 - 7,500 × 28 = 210,000), and after that the gap begins to close.

A growth plan that is "profitable on LTV" can still run out of cash. Always pair LTV : CAC with payback and a cash forecast. See CAC payback period and cash flow management.

6. Averaging across segments that behave differently

Two channels (illustrative):

Blended contribution = 0.6 × 90 + 0.4 × 70 = 54 + 28 = $82. Blended CAC = 0.6 × 60 + 0.4 × 110 = 36 + 44 = $80. The blend shows +$2 per customer, close to break-even. Per 1,000 customers, Channel A makes 600 × 30 = $18,000 and Channel B loses 400 × 40 = $16,000. Reallocating the budget from B to A would improve the result substantially, a decision the average conceals. Always cut unit economics by channel, product, customer segment and cohort. See cohort analysis.

7. Allocating fixed costs as if they were variable

A product sells at $50, with variable cost of $30 and $15 of allocated overhead per unit (based on 100,000 units, so $1.5 million of overhead in total). On a fully loaded basis, the margin is 50 - 30 - 15 = $5, or 10 percent.

A customer offers to buy 10,000 additional units at $42, using spare capacity.

• Fully loaded cost is $45, so the order looks like a loss of $3 per unit. Reject?
• Contribution is 42 - 30 = $12 per unit, so the order adds 10,000 × 12 = $120,000 of profit, because the overhead is already paid.

The correct decision depends on whether capacity is truly spare and whether the lower price will leak to other customers. But the fully loaded view cannot even see the opportunity. The opposite error also occurs: treating all costs as fixed so that every order looks profitable. See variable vs. fixed costs.

8. Building the model once and never stress-testing it

A model gives twelve-month contribution per customer of $150 against CAC of $100, a $50 profit. It was built in a good quarter and never revisited. In the next year, CAC rises 20 percent to $120 and contribution falls 25 percent because repeat purchase weakens, to 150 × 0.75 = $112.50.

• New result = 112.50 - 120 = -$7.50 per customer

A swing from +$50 to -$7.50 came from two ordinary moves. Models should have ranges, not single points: a base case, a downside and an upside, with the assumptions that drive each. Update the model monthly or quarterly, and run a sensitivity table on the three inputs that matter most. See sensitivity analysis on unit economics.

A checklist to avoid all eight

1. Use contribution or gross margin in LTV, never revenue.
2. List every variable cost from source invoices and statements.
3. Track blended and paid CAC, and decide on the marginal one.
4. Show observed value next to projected value, and cap lifetimes.
5. Pair LTV : CAC with payback and cash needs.
6. Break results out by channel, product, segment and cohort.
7. Separate variable from fixed costs before judging incremental decisions.
8. Stress-test and refresh the model on a schedule.

Why operations belong in the model

Several of these errors have an operational root. Returns, expedited freight, markdowns on excess stock and lost sales from stockouts are consequences of the way a company plans demand and manages inventory, and they show up in unit economics as lower contribution than the model assumed. For consumer product companies, the cheapest improvement is often better planning, not better marketing. See demand planning and inventory optimization.

FAQs

What are the most common unit economics mistakes?

Using revenue instead of contribution, omitting shipping, returns and fees, trusting blended CAC, projecting unobserved lifetimes, and ignoring payback and cash.

How do I know if my unit economics model is wrong?

Reconcile it to the ledger. If per-unit contribution times units, less fixed costs, does not equal operating profit, something is missing or misclassified.

Is a high LTV to CAC ratio enough?

No. A high ratio can still hide long payback, heavy cash needs, one weak segment, or lifetimes that have not been observed.

How often should I review unit economics?

Monthly for the main inputs, and whenever prices, costs, channels or customer behavior change materially.

What is the fastest mistake to fix?

Adding the missing variable costs. It requires only invoices and a few hours, and it often changes decisions on its own.