LTV models that survive board scrutiny

Slide four says LTVLTVLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition →/CACCACCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → is 5:1. The three questions that follow are always the same: what margin sits behind it, whose retention curve, and how many months did you count. The same cohort data produces a number twice the size of the one your CFO would sign, depending only on those answers. This lesson builds the version that holds.

Why the naive formula is popular, and wrong

The textbook shortcut:

LTV = ARPU / Churn RateChurn RateChurn rate is the percentage of customers or revenue lost over a period. It measures how fast a business loses its existing customer base.View full definition →

ARPU (average revenue per user, usually monthly) divided by monthly churn gives an average lifetime in months, multiplied by monthly revenue. If ARPU is $100/month and monthly churn is 2%, naive LTV = $100 / 0.02 = $5,000.

Three problems:

1. It uses revenue, not profit. A customer paying $100/month at 40% gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.View full definition → contributes $40.
2. It ignores the time value of money. A dollar collected in month 36 is worth less today than a dollar collected in month 1, because capital has a cost and risk compounds.
3. It assumes flat revenue per customer. Usage-based products (Twilio bills per message and per APIAPIApplication Programming Interface: a standardised interface that lets applications communicate and exchange data without knowing each other's internal workings.View full definition → call) expand and contract with consumption, so a static ARPU misrepresents the cash flow curve.

There is a quieter fourth. A single blended churn rate applied to a mixed book does not give you the average of the segment LTVs. Two equal segmentssegmentsDividing a market into distinct groups of customers who share similar needs, characteristics or behaviours, so each group can be served with a tailored approach.View full definition → churning at 1% and 5% a month blend to 3%, worth 33 months of ARPU; modelled separately and averaged they are worth 60. The blend hides the enterprise cohort carrying the business and the SMB cohort that never pays back.

The defensible formula

LTV = Σ [ (ARPU_t × Gross Margin) / (1 + d)^t ] summed over each period t, up to a stated horizon.

Where:

• ARPU_t: average revenue per user in period t, allowed to vary by period
• Gross Margin: revenue minus cost of goods sold (hosting, support, payment processing). SaaS (Software as a Service) companies typically report 70 to 85% as of 2024 estimates (source: KeyBanc SaaS Survey, figures vary by year); usage-heavy infrastructure runs lower, with Twilio's reported gross margin sitting around half of revenue because messaging carrier fees scale with volume
• d: discount rate per period. A common SaaS convention is 10 to 15% annualised, converted to a monthly rate
• t: the period, month by month
• the horizon: the month you stop. Not optional, and not infinity

Worked example: what each correction costs

Flat subscription product, $100 ARPU, 80% gross margin, 2% monthly churn, 12% annual discount rate (about 1% monthly):

• Naive ARPU / churn: $5,000
• Margin applied, no discounting, infinite tail: $4,000
• Margin and discounting, infinite tail: about $2,650
• Margin and discounting, capped at 60 months: about $2,230
• Same, capped at 36 months: about $1,760

The shortcut overstates the 60-month figure by more than double. Note which correction bites hardest: discounting removes a third, the margin haircut a fifth, the horizon cap a sixth. Marketers argue about margin and skip the discount rate, which is the larger error.

Case B: usage-based pricing. A consumption product where accounts ramp: $100 in month one, $130, $160, plateauing near $200 by month 12, at a 50 to 65% margin because compute or carrier costs scale with each incremental dollar. Here the naive formula is worse than wrong, it is directionally confused. Applying a snapshot ARPU to a churn-based annuity assumes a steady state that usage products never reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → in year one, and it ignores months when a single customer runs a large batch job. Forecast a revenue curve per cohort, apply the lower margin, discount it. In practice this lands 20 to 35% below the naive number even though gross usage grows.

Horizon caps: the month you stop counting

An infinite sum assumes the product, the company and the customer all exist in 2045. Boards do not credit that, and the tail is where fabricated value hides. Two working rules.

Cap at the oldest cohort you can actually observe. A four-year-old company presenting an 84-month LTV is extrapolating three years past its own evidence. Say so on the slide before someone else does. Adobe can model long horizons credibly because it has had subscription cohorts running since it retired perpetual Creative Suite licences in 2013; a Series B company has 30 months of data and a curve fitted to hope.

Then fix the cap and keep it constant. Many finance teams settle on 36 months for monthly SMB motions and 60 for annual enterprise contracts. The number matters less than the consistency: quietly stretching the horizon from 36 to 60 months between two board meetings raises LTV by a quarter with no change in the business, and it is the first thing a diligence team spots.

The second-order issue is that the churn hazard is not constant. Cohorts shed heaviest in the first few months and the survivors leave far more slowly, which is why the early warning signals the engagement lesson instruments are worth most in that window. Fit a constant rate to months one to six and you underprice the tail; fit it to month 24 survivors and you overprice everything.

Why gross margin matters more than people think

Two subscription businesses with identical ARPU and churn can differ threefold in LTV. Spotify's premium tier looks like textbook recurring revenue: monthly price, low ARPU, measurable churn. Its gross margin runs somewhere between a quarter and a third of revenue, because rights holders take most of each euro before Spotify sees it. Adobe, selling software it wrote itself, reports gross margins in the high 80s. Per dollar of subscription revenue, the Adobe customer is worth roughly three times the Spotify one before either side has churned anybody.

If your deck uses revenue instead of margin dollars, you are implicitly claiming a 100% margin business. Vertical SaaS reselling embedded payments or financing often blends down to 50 to 65% once transaction costs land.

Discounting: the piece almost everyone skips

Cost of capitalCost of capitalThe blended rate a company pays to finance itself through debt and equity. It sets the minimum return an investment must clear to create value.View full definition → is a marketing accountability question, not a finance department curiosity. When the fully loaded acquisition cost the CAC lesson builds takes two years to come back, an undiscounted LTV overstates the return on every dollar of that spend; the payback and magic number lesson sets the thresholds, but only a discounted LTV makes the numerator comparable.

Rough monthly factor: a 12% annual rate is about 1% monthly (precisely, 1.12^(1/12) − 1 ≈ 0.95%). Multiply each period's cash flow by 1/(1+monthly rate)^t. Ask finance for the company's actual rate rather than defaulting to 10% because it is round. Venture-stage businesses often carry 15 to 20%, and at 20% the 60-month LTV in the example above falls under $1,900.

Simple spreadsheet logic

Month | Survival % | ARPU  | Margin$  | Discount factor | PV contribution
1     | 100%       | $100  | $80      | 0.990            | $79.20
2     | 98%        | $100  | $78.4    | 0.980            | $76.83
3     | 96.04%     | $100  | $76.8    | 0.971            | $74.58
...
Sum across all months = Defensible LTV

Five minutes in any spreadsheet: one column compounding the survival probability, one for margin dollars, one for the discount factor, then sum the products and stop at your stated horizon.

Benchmarks to sanity-check your number

As of 2024 to 2025 estimates, treat these as directional:

• LTV/CAC ratio: 3:1 or better is the commonly cited target, based on the defensible calculation (Bessemer's State of the Cloud; Bessemer invests in cloud companies and publishes on its own portfolio's category). A naive 5:1 usually resolves to 2 to 3:1.
• Gross revenue retention: 85 to 95% annually for healthy B2B SaaS, and this is the input your survival curve should come from.
• Net revenue retention above 120% is achievable in expansion-heavy usage businesses, as the expansion lesson covers, but it changes the shape of the revenue curve you discount rather than excusing you from discounting it.
• The gap itself: run both numbers. At typical SaaS churn and margin, the defensible LTV lands 40 to 60% below the naive one. If yours is within 20%, you have dropped an input.

🎬 [VIDEO: "SaaS Metrics That Matter" - youtube.com/@SaaStr - a practitioner walkthrough of LTV, CAC, and retention metrics used in real board decks, search SaaStr's channel for the latest metrics session]

The board-ready checklist

Before presenting an LTV number, confirm you have applied:

1. Gross margin, not revenue
2. A discount rate from finance, with the rate itself stated on the slide
3. A survival curve from actual cohort data, not one blended churn rate
4. Segments modelled separately where churn differs materially
5. A stated horizon, and a note on how much of it is observed versus extrapolated
6. For usage-based products, a revenue curve per cohort rather than static ARPU

Key Takeaways

• Naive LTV (ARPU / churn) can more than double the defensible number: margin costs it 20%, discounting a third, the horizon cap another sixth.
• The defensible formula discounts margin-adjusted cash flows period by period, up to a horizon you declare: LTV = Σ (ARPU_t × Gross Margin) / (1 + d)^t.
• Margin is the difference between two identical-looking subscription businesses. Spotify at roughly a third and Adobe in the high 80s are not in the same LTV conversation.
• Extending the horizon is the easiest way to inflate LTV without changing anything real, which is why the cap and its evidence base belong on the slide.
• Blended churn across mismatched segments produces a number that describes no customer you actually have.