Modeling lifetime value across devices, disposables and service

Ask what a da Vinci account is worth and the answer changes depending on which year you stop counting. Intuitive Surgical, which sells both the systems and the instruments that run on them, has reported for years that recurring revenue (instruments, accessories and service) is around 80 percent or more of total revenue. The capital line is the smallest part of the account, and it is usually the only part the marketing plan models. Stretch the view to a decade and the mix drifts again, because leases replace outright sales, software gets attached, and per-test prices fall while volumes rise.

The revenue layers of a placed device

A placement creates streams that behave differently and should never be averaged together:

1. Capital. Sold outright, leased, or placed under a usage-based agreement. Sometimes near zero upfront to win the site (the reagent rental structure common in diagnostics). Intuitive has moved a growing share of placements to leases rather than sales.
2. Consumables. Cartridges, reagents, instruments, masks and cushions. Consumed per procedure, per test or per week of therapy.
3. Service. Maintenance, calibration, warranty extensions, uptime guarantees.
4. Software. Per-test cloud fees, planning and analytics subscriptions. ResMed runs this as a reported segment of its own, on the order of hundreds of millions of dollars a year from its out-of-hospital software businesses.

Illumina, which sells the sequencers and the chemistry that runs on them, states the logic openly: instrument placements are judged on expected consumable pull-through per instrument per year, roughly a million dollars annually for a high-throughput sequencer in its stronger years. The box is a distribution channel for the chemistry.

Why account-level LTVLTVLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → beats per-transaction

A per-transaction view asks whether a reagent order carried margin. An account-level view asks a harder question: over the installed life of this placement, what is the total contribution margin from all four streams, net of what it cost to win and hold the site?

The gap between those two questions decides budget. Optimise per-order margin and you underprice nothing and win nothing. Optimise account value and you will discount the capital line, or give it away, to own the stream behind it.

Building the LTV model

We use contribution margin (revenue minus variable cost), not net profit, because this is a marketing decision model, not a P&L.

Account LTV = (capital margin) + (annual consumables margin x years x retention) + (annual service and software margin x years x retention)

Then subtract the cost of acquiring the account: the committee-scale, 18-month figure the foundations lesson on hospital buying units builds. Assume that number here rather than rebuild it.

A worked example (illustrative numbers)

One hospital account for a mid-range surgical device. These figures are illustrative for teaching, not market data.

• Capital sale: $400,000 at 40 percent contribution margin = $160,000
• Consumables: 500 procedures/year x $300 x 50 percent margin = $75,000/year
• Service contract: $40,000/year x 60 percent margin = $24,000/year
• Useful life: 7 years
• Annual retention of consumables and service: 90 percent

Recurring margin per year = $99,000, decayed by the survival factor:

Year 1 recurring: 99,000 x 0.90^0 = 99,000
Year 2:           99,000 x 0.90^1 = 89,100
Year 3:           99,000 x 0.90^2 = 80,190
Year 4:           99,000 x 0.90^3 = 72,171
Year 5:           99,000 x 0.90^4 = 64,954
Year 6:           99,000 x 0.90^5 = 58,458
Year 7:           99,000 x 0.90^6 = 52,613
                  ----------------------------
Sum recurring   = 516,486
Plus capital    = 160,000
Gross account LTV = 676,486

With a $120,000 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 →, net account LTV is $556,486, and the ratio to acquisition cost is 5.6:1. Capital contributed $160,000; the recurring layers contributed $516,486, which is 76 percent of gross value.

Now discount it, because a decade model that ignores the 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 → flatters itself. At 10 percent, the recurring stack falls from $516,486 to about $411,000, a fifth off, and the ratio drops to 4.8:1. A dollar of year-ten margin is worth under 40 cents today. This matters most for the businesses that place equipment on lease: the margin arrives later, the asset sits on your balance sheet, and the account that looks richest on an undiscounted sum can be the one that starves cash. Whether 4.8:1 clears the bar is the benchmarks lesson's argument, not this one's; what belongs here is that the same account can present two ratios a full point apart depending on one assumption you chose in a spreadsheet (First Round Review on SaaS metrics shows how software firms handle the same choice).

Retention is a variable, not a constant

Recurring value assumes you keep selling the blade. Two forces attack it and one erodes it quietly.

• Third-party servicing. The FDA has examined how independent servicers and remanufacturers should be regulated. Where hospitals can buy cheaper service, the service layer of your model compresses first, because it is the easiest line for a procurement team to unbundle.
• Compatible consumables. Competitors reverse-engineer cartridges and reagents, and patent litigation over sequencing consumables has been a recurring feature of Illumina's history. Model leakage, not a sealed system. In the EU, MDR (Regulation 2017/745) governs whether a compatible or reprocessed item can be placed on the market at all, which is why leakage risk differs by geography inside the same global model.
• Price per unit falling faster than volume rises. This is the failure mode nobody puts in the spreadsheet. Illumina has spent fifteen years driving the cost of a genome down, announcing a $200 genome with the NovaSeq X in 2022. Volume grew; dollars per test collapsed. A seven-year model with a flat $300 per procedure will overstate a decade of value in any category where the chemistry keeps improving. Run the consumable line with a price curve, not a price.

SegmentingSegmentingDividing 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 → LTV: not all accounts are equal

By utilisation

Account value is skewed, not distributed. A site running 1,000 procedures a year on the same hardware is worth several times a site running 150, and the gap widens every year the survival factor compounds. The failure mode is the dormant placement: a system installed, trained once, and used twice a month by a single surgeon. Under an outright sale you booked the capital margin and lost the stream. Under a lease or usage-based placement you now carry the asset and earn almost nothing, so a portfolio of low-utilisation leases can be worse than no placement at all. Track the utilisation distribution, not the mean, and use the measurement approach the post-adoption lesson sets out rather than inventing a second one here.

By expansion potential

One placement can become three, or spread from cardiology to general surgery. Net revenue retentionNet revenue retentionNet Revenue Retention measures the percentage of recurring revenue retained and grown from existing customers over a period, including upsell and expansion, net of downgrades and churn.View full definition → is the metric: recurring revenue held plus expansion within existing accounts. Above 100 percent, the installed base grows without a single new placement, which changes what a placement is worth before you have sold it.

By stream mix

Two edge cases break the razor-and-blade template.

Coloplast has no razor. Its ostomy and continence business is consumables from the first order, sold to individual chronic users with no capital anchor and gross margins in the high sixties. There is no hardware to discount, so the entire model rests on years of continued use, and churn is patient-level: therapy change, hospitalisation, death, or a competitor's sample pack.

ResMed shows the opposite asymmetry: the consumable stream outlives the device. Masks, cushions and tubing get replaced on a cadence set largely by payer replacement schedules, not by the manufacturer's ambitions, which is why coverage rules (the subject of the payer adoption lesson) sit upstream of the consumable line in the model. A change in resupply entitlement moves LTV across an entire installed base at once, without any competitor doing anything.

FunnelFunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → implications

In the worked model, 76 percent of gross account value arrives after the placement, and none of it is guaranteed by the signature. Most medtech marketing budgets concentrate on the stages the adoption funnel lesson maps up to purchase, and thin out exactly where the value is created. The reallocation is rarely popular internally, because adoption and renewal spend has no deal date attached to it, and the salesperson who won the placement has already moved on.

The second-order effect is organisational. If most value sits in years two to seven, the people who protect it (clinical education, application specialists, service) sit outside marketing's budget and outside its metrics. Building the LTV model is the easy part. Getting the money to follow it is the argument you will have every planning cycle.

Key takeaways

• Model at the account level across four streams, and expect the capital line to be a minority of ten-year value: Intuitive has run at roughly 80 percent or more recurring for years.
• Apply a survival factor across useful life, then discount. Ten percent took a fifth off the recurring stack in the example and moved the ratio from 5.6:1 to 4.8:1.
• Price per unit is an assumption, not a constant. Falling consumable prices, third-party service and MDR-dependent compatible products all erode the blade.
• Utilisation is skewed. Dormant placements under lease or usage terms destroy value rather than merely disappointing.
• Some businesses have no razor (Coloplast) and some have consumables that outlive the device on a payer-set cadence (ResMed). Fit the model to the mix, not the mix to the model.