Modeling lifetime value when contracts, churn and tariffs all move independently

# Modeling lifetime value when contracts, churn and tariffs all move independently

A dual-fuel household on a fixed tariff in Manchester is worth roughly three times more over five years than a single-fuel switcher on a variable rate in the same postcode, even though both pay similar monthly bills today. The gap isn't visible in month-one revenue. It only shows up when you model lifetime value (LTV) properly.

Most energy marketers still calculate LTV as "average revenue times average tenure." That formula quietly assumes every customer behaves like the average customer. In energy retail, that assumption breaks fast, because three variables move independently: contract length, tariff type, and usage band. This lesson shows you how to model LTV when they don't move together.

Why energy LTV is harder than subscription LTV

In a SaaS (Software as a Service) business, LTV usually depends on one dominant lever: churn rate. Energy is messier because value per customer is a moving target, not a fixed subscription fee.

Three forces multiply against each other:

Contract length: fixed-term deals (typically 12 to 24 months in the UK and much of continental Europe) versus rolling/variable contracts common in deregulated US states like Texas.
Tariff type: fixed-rate (price locked for the term) versus standing/variable (price can move with wholesale market changes, subject to regulatory price caps in some markets, such as Ofgem's price cap in Great Britain).
Usage band: low, medium, high consumption households, often segmented by annual kWh (kilowatt-hour) of electricity or therms/kWh of gas.

A customer's LTV is not "revenue x tenure." It's closer to:

LTV = Σ (expected revenue in period t × probability of retention through t × gross margin) − acquisition and service costs

The complexity is that retention probability itself depends on contract length and tariff type, not just on generic "churn rate."

Building the model: three levers, not one

1. Contract length sets the churn clock

Fixed-term contracts create a predictable churn spike at renewal, called the "cliff edge." Retention is high and mechanical during the term (customers usually can't leave without an exit fee), then drops sharply at expiry.

Rolling/variable contracts spread churn more evenly across the year, driven by price comparison sites and switching campaigns rather than contract-end dates.

Modeling implication: don't apply one monthly churn rate to both. Model fixed-term customers with near-zero in-term churn plus a renewal-point churn probability (often 25 to 45% at first renewal in mature European retail markets, as an industry estimate for 2025 to 2026). Model rolling contracts with a constant monthly churn rate instead.

2. Tariff type changes both margin and stickiness

Fixed tariffs lock in a margin for the retailer but expose it to wholesale price risk if energy costs spike (as seen dramatically in Europe's 2022 gas crisis following the invasion of Ukraine, which pushed many suppliers into losses or insolvency).

Standing/variable tariffs pass more price risk to the customer, which is safer for the retailer's margin but tends to correlate with higher switching rates, since customers on variable rates are more price-sensitive and more likely to be actively comparing offers.

Modeling implication: apply a margin adjustment and a churn multiplier by tariff type. A reasonable planning assumption (estimate, not a universal constant): variable-tariff churn runs 1.3 to 1.8x fixed-tariff churn in competitive retail markets.

3. Usage band drives absolute value, not just percentage

A high-usage household (large home, electric heating, EV charging) generates more margin per year even at the same percentage margin, simply because the revenue base is bigger. Usage band also correlates with dual-fuel bundling: high-usage households are more likely to buy both gas and electricity from one supplier, which itself reduces churn (bundled customers switch less because there's more friction in leaving two contracts at once).

Worked example: dual-fuel vs single-fuel switcher

Assume simplified, illustrative figures (clearly estimates, for teaching purposes, not published market data):

Customer A: Dual-fuel, fixed 24-month tariff, medium usage

Annual gross margin per fuel: $180 (estimate)
Two fuels: $360/year combined margin
In-term monthly churn: near 0%
Renewal-point churn probability: 35%
Expected tenure beyond first term: roughly 1.5 additional terms on average given bundled stickiness

Customer B: Single-fuel (electricity only), rolling variable tariff, medium usage

Annual gross margin: $150 (estimate, lower because variable tariffs are typically priced thinner to stay competitive on comparison sites)
Monthly churn rate: 3% (roughly 30% annual attrition, typical of price-comparison-driven markets)

Simple LTV approximation (undiscounted, for clarity):

LTV_A (dual-fuel, fixed) ≈ 360 × 2.5 years expected tenure = $900
LTV_B (single-fuel, variable) ≈ 150 × (1 / 0.30 annual churn) = 150 × 3.3 years = $500

Even with a smaller revenue gap than the opening claim suggested in raw dollar terms, Customer A delivers roughly 1.8x the LTV of Customer B, driven almost entirely by retention structure, not usage or price. Add customer acquisition cost (CAC), and the gap widens further: dual-fuel customers are frequently cheaper to acquire per unit of margin because bundled offers convert better in comparison-site funnels (a commonly cited industry pattern, treat as directional).

LTV:CAC sanity check: A healthy marketing-efficiency benchmark across subscription-like sectors, often cited as a general rule of thumb, is an LTV:CAC ratio of 3:1 or higher (see Bessemer's memo on SaaS metrics for the origin of this heuristic, though it's not energy-specific). If Customer A costs $150 to acquire, LTV: = 6:1. If Customer B costs the same $150, : = 3.3:1, still viable, but far thinner.

Vérification des acquis

1. Why does the simple 'average revenue × average tenure' LTV formula break down in energy retail?

2. What is the key structural difference between SaaS LTV modeling and energy retail LTV modeling?

3. A dual-fuel household on a fixed tariff can be worth roughly three times more in lifetime value than a single-fuel variable-rate customer with similar current bills. What does this gap illustrate?

CHOIX MULTIPLES

4. Select ALL correct answers about the three variables that multiply against each other in energy LTV modeling.

Sélectionnez toutes les réponses correctes.

CHOIX MULTIPLES

5. Select ALL correct answers about why retention probability matters in the energy LTV formula.

Sélectionnez toutes les réponses correctes.

Putting it into a simple cohort model

For a real model, don't collapse everything into one blended number. Build a small matrix: contract length x tariff type x usage band, and calculate LTV per cell. A basic spreadsheet structure:

Segment: [Dual-fuel, Fixed-24mo, High-usage]
  Monthly margin = usage-based revenue × margin %
  Survival curve = 1.0 during term, then apply renewal churn %
  LTV = Σ (month 1..60) margin(t) × survival(t)
  Discount optional: apply monthly discount rate if comparing to CAC payback

Run this for each realistic segment (typically 8 to 16 cells for a mid-size retailer) rather than one company-wide average. This is the same logic used in European Commission energy retail market monitoring reports, which segment household customers by fuel type and contract status precisely because blended averages hide the real economics.

Where this changes marketing decisions

Once LTV is segment-specific, three decisions shift:

Customer acquisition cost in a market where switching is rare and slow

Mapping the funnel from quote to switch to first bill

Modeling lifetime value when contracts, churn and tariffs all move independently

Building the model: three levers, not one

1. Contract length sets the churn clock

2. Tariff type changes both margin and stickiness

3. Usage band drives absolute value, not just percentage

Worked example: dual-fuel vs single-fuel switcher

Putting it into a simple cohort model

Where this changes marketing decisions

Key Takeaways