Modeling lifetime value across equipment, parts and service contracts

Otis prints the argument in its own segment reporting. New equipment, roughly $6 billion a year, earns single digit operating margins. Service, roughly $8 billion, earns margins in the low-to-mid twenties and supplies most of the company's operating profit. The elevator wins a maintenance route that can run twenty years or more, and the route is the business. Otis sells service contracts, so read its enthusiasm accordingly; the segment split is audited either way.

Most industrial LTVLTVLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → models never get that far because they stop at the first invoice. This lesson builds the multi-stream version: equipment, parts and consumables, service agreements, financing, and the residual value of the asset when it changes hands.

The four streams and why the machine is the smallest

Each stream has its own margin, cadence and failure mode.

• Equipment: high ticket, thin margin (often 10 to 25%, varies widely by category), one sale every 7 to 15 years per unit, and the only stream a competitor can take from you with a discount.
• Parts and consumables: recurring, 30 to 60% margin (estimate), driven by machine hours rather than calendar time.
• Service agreements: recurring, 40 to 70% margin (estimate), renewed annually or on three year terms.
• Financing: interest spread plus fee income. Caterpillar and John Deere both run captive finance arms as reported 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 →, several billion dollars of annual revenue at Cat Financial and hundreds of millions in net income at John Deere Financial. Financing also holds the asset inside the OEM channel during the years when parts attach is easiest to defend.

Then the residual. A machine traded in at year 8 keeps consuming parts under a second owner, so lifetime value attaches to the serial number as much as to the buyer. That is why Caterpillar's public services goal, set in 2019, to roughly double machine and engine services revenue to $28 billion by 2026, is written against the installed fleet rather than against annual unit sales.

Building the multi-stream model

LTV(asset) = Equipment gross profit
           + Σ_t (Parts revenue_t × parts margin × S_t)
           + Σ_t (Service revenue_t × service margin × R_t)
           + Σ_t (Finance spread_t)
           + Remarketing gross profit at trade-in
           with each year t discounted at your cost of capital

S_t and R_t are cumulative survival curves for parts share and contract renewal, compounded year over year, not flat multipliers. Feed them with the renewal and attach measures the installed-base lesson defines; do not mint new ones here.

Worked example

A machining center sells for $150,000 at 12% 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 →.

• Equipment gross profit: $18,000
• Parts: $12,000 a year (tooling, coolant, filters) at 45% margin = $5,400
• Service: $8,000 a year contract at 55% margin = $4,400
• Recurring gross profit: $9,800 a year, at 90% annual retention

Undiscounted across ten years the recurring stream sums to about $57,000, so total LTV lands near $75,000, four times the profit visible on the machine. Against equipment profit alone, a $10,000 fully loaded acquisition costacquisition costCustomer 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 → looks unaffordable at 1.8:1; against multi-stream LTV the same spend clears the 3:1 rule of thumb imported from SaaS (see OpenView's benchmark work on SaaS metrics for where that ratio comes from) with room to spare. How the $10,000 gets loaded belongs to the acquisition cost lesson.

The twenty year tail is worth less than you think

Now discount it. At an 8% 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 →, retention and discounting compound together (0.90 / 1.08 = 0.833 per year) and the ten year recurring stream drops from about $57,000 to about $41,000. Total LTV: $59,000 rather than $75,000. A fifth of the model gone, purely from arithmetic that marketing decks omit.

Extend the horizon to twenty years, matching a real asset life, and years 11 through 20 add only about $6,600 of present value, some 14% on top. The first five years carry roughly 61% of the whole discounted stream. Two consequences: stop negotiating terminal assumptions for year 18, and move retention spend into years one to three, where the curve is still worth something.

Failure modes: double counting and the uptime paradox

The most common modeling error is counting one machine twice. When a dealer takes a trade-in, refurbishes it and resells it, the parts and service streams of that serial number get booked once against the original account and again against the second owner. Certified used equipment programmes exist because that second life is profitable; the accounting has to follow the asset, then roll up to the account.

The second error is subtler. Predictive maintenance, sold as a customer benefit, suppresses the emergency parts and premium labour that historically funded the aftermarket P&L. John Deere's telematics and dealer alerting push repairs from unplanned to planned; failures get caught as scheduled work at ordinary urgency. If your model calibrates parts consumption per machine hour on an unconnected fleet, a connected fleet will undershoot it, and the shortfall shows up in year four when nobody is looking. Reprice the agreement before the parts line falls; how that price is structured is the servitization lesson's subject.

Retention, and whose retention it is

Move annual retention from 90% to 80% and the discounted recurring stream falls from about $41,000 to about $27,000, roughly a third of it, and a quarter of total LTV. No pricing decision in the model has that kind of leverage.

The harder question is whose retention the model measures. In the Caterpillar and Deere structures the dealer is an independent business that owns the parts counter and the service van. The OEM books parts at transfer pricetransfer priceThe prices charged when units of the same company trade goods, services or IP across borders, set to satisfy tax authorities and reflect market value.View full definition → and captures a fraction of aftermarket margin; the dealer keeps the rest. That is two LTV models over one machine, and marketing spend that grows dealer LTV without growing OEM LTV is easy to approve by accident. Decide which P&L you are modeling, then check whether the dealer agreement actually gives you the utilization data (hours, fault codes, parts orders) the survival curve needs. Without it you are estimating S_t from invoices that arrive two owners late. The warning signs that a fleet is drifting to independent servicers are the installed-base lesson's territory; here they matter as the inputs that keep the curve honest.

Qualifying on lifetime potential, not order size

Score leads on expected order size and pipelinepipelineAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition → tilts toward one-off capital purchases, away from the fleets that pay for a decade. Weight instead by:

• machine hours per year (a three shift automotive supplier burns tooling several times faster than a single shift job shop running the identical machine)
• fleet size and replacement cadence, which tell you how many future serial numbers sit behind this one order
• who holds the service contract today: a prospect shopping away from a rival's agreement carries higher lifetime value than a greenfield buyer with no maintenance habit at all
• financing intent, since a financed asset tends to stay inside your channel longer

Adding two fields to the handoff criteria the funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → foundations lesson sets out, estimated annual machine hours and current service contract status, does more for LTV forecasting than another revision of scoring weights.

Attach rates: read them as directional

Service contract attach at point of sale is commonly cited around 40 to 60% in North America and higher in parts of Western Europe, Germany in particular, where planned maintenance and uptime obligations weigh more. Both figures are mid-2020s estimates and vary heavily by machine class; OEMs rarely publish methodology behind them. Build your own from the last 200 deliveries: attach at delivery, attach within twelve months, and parts share of wallet at year three. Three internal numbers beat any published range, and unlike the range they move when you act.

🎬 [VIDEO: "Understanding Customer Lifetime Value (CLV)" - youtube.com/results?search_query=understanding+customer+lifetime+value - search and select a marketing-fundamentals explainer from a reputable business school or analytics channel to see the general CLV math before applying the manufacturing-specific layers in this lesson]

Key takeaways

• Multi-stream LTV (equipment, parts, service, financing, residual) runs 3 to 5 times equipment-only gross profit, and that is the base a CAC ceiling should sit against.
• Discount it. At 8% cost of capital, a ten year recurring stream loses about a fifth of the total model, and years 11 to 20 of a twenty year asset add only some 14% more.
• Retention is the most sensitive input: 90% to 80% erases about a third of the discounted recurring profit.
• Model per serial number, then roll up to account, or the second owner's parts revenue gets counted twice.
• With independent dealers, most aftermarket margin sits in the channel. Decide whose P&L the model serves and secure a data clause, or you cannot see the utilization behind your own curve.
• Predictive maintenance suppresses the emergency parts that funded the aftermarket, so a connected fleet will undershoot any model calibrated on unconnected history.