The Half-Life of a Launch Assumption

The TransMedics case

In late 2021, TransMedics launched the National OCS Program, or NOP. The program gave transplant centers an outsourced service for donor-organ retrieval and management using the company's Organ Care System. It addressed an operating problem that sat around the device: a transplant center could use the technology, but it still needed trained people to retrieve the organ, manage it on the OCS, and get it back to the recipient center.

The next commercial dependency was transportation.

TransMedics still depended on third parties to transport donor organs and medical personnel. That dependence was stated as a risk in the company's 2022 annual report. A delayed or unavailable flight was not simply a logistics issue. It could affect whether the NOP service could be delivered at all, which meant it sat upstream of OCS utilization, transplant-center experience, service revenue, and the company's ability to expand the program nationally.

By August 2023, TransMedics had changed the model. It acquired Summit Aviation, a charter operator based in Bozeman, Montana, and set up TransMedics Aviation. By the end of 2023, it had bought 11 fixed-wing aircraft for a total of $141.9 million. It added two more in early 2024. The company said the acquisition let it add aircraft transportation to NOP and become a national provider of donor-organ retrieval and delivery.

This was more than a make-or-buy decision about flights. The launch model had changed from providing a device and clinical service around organ retrieval to controlling another part of the delivery system.

The public filings do not show the internal decision process. They do show the sequence. NOP launched in late 2021. Dependence on third-party transport became a stated operating risk. In 2023, TransMedics bought an aviation business and aircraft to reduce that dependence.

Revenue also grew during this period, from $30.3 million in 2021 to $93.5 million in 2022 and $241.6 million in 2023. TransMedics said the 2023 increase came mainly from higher OCS utilization through NOP and added logistics services. This does not prove that owning aircraft caused the growth. It does show that logistics had become part of the commercial system and a reported source of revenue, not a support activity outside the launch plan.

The lesson is useful for smaller medtech teams because most launch plans contain assumptions that look like facts after they have been repeated enough times.

A company may assume that a hospital will complete contracting in 90 days, a physician champion can produce 12 cases a week, a distributor will hold enough inventory, one clinical specialist can support four sites, or a reimbursement decision will arrive before the sales team expands. Each assumption can be reasonable when the plan is approved. None stays true for the life of the plan.

A launch assumption has a useful life

Companies usually record decisions and tasks. They are less disciplined about recording the assumptions that made those decisions reasonable.

A hiring plan may say: add four territory managers in the third quarter. A site plan may say: open five ASCs before the end of the year. An inventory plan may say: build 60 procedure kits. The assumptions under those decisions are often spread across a spreadsheet, a board deck, an email, and the memory of the person who built the model.

This creates a problem when new evidence arrives. The team can update the evidence without reopening the decisions that depended on it.

Suppose an ASC pilot was expected to produce 12 procedures a week. That volume supported an annual contract price, the number of kits placed at the site, the amount of clinical support assigned, the timing of the next site, and the revenue forecast. After launch, the site produces five procedures a week.

Changing 12 to five in the model is easy. The actual work is identifying which decisions are no longer supported.

The price may still work. The contract may need a different structure. Inventory may need to move to another site. A planned hire may need to wait. The next site may need a different case-volume threshold. The board forecast may need to change. None of those outcomes should be automatic, but each decision should reopen for review.

This is the half-life of an assumption: the period during which the evidence under it is current enough to support the decisions attached to it.

The half-life is not the same for every assumption. A product label changes through a formal regulatory process and may stay stable for years. A hospital contracting estimate can change after the first value-analysis review. Site utilization can change every week. A distributor forecast may be useful for one planning cycle and weak after the first 30 days of orders. A reimbursement assumption may remain stable until a specific CMS or payer event.

For this reason, reviewing every assumption once a quarter is not enough. Each important assumption needs its own condition for review.

The five fields an assumption needs

An assumption register does not need to be large. It needs to answer five questions.

Confidence can also be recorded, but confidence should not replace a trigger. A team can be highly confident and still be wrong. A low-confidence assumption can be acceptable if the company limits the commitment attached to it.

The reopen trigger is the most important field because it converts a statement into an operating control.

"We expect 12 procedures a week" is a planning input.

"If the four-week average falls below nine, reopen price, inventory, coverage, and forecast" is a launch control.

The trigger should be specific enough that the team can tell when it has been crossed. "If utilization is weak" leaves room to postpone the discussion. "If clearance is late" has the same problem. A better trigger names the date or threshold: if FDA clearance is not received by September 30, reopen the launch date, field hiring, inventory release, and cash plan.

The assumption and the decision are different records

When evidence changes, teams often overwrite the old number and move on. That removes the history needed to understand why a decision was made.

The original assumption should remain visible with its source and date. The new observation should be added beside it. The linked decisions should then be marked for review.

This matters because a changed assumption does not always produce a changed decision.

A hospital may take six months longer to contract, but the company may keep the site in the first launch group because it provides evidence that no other site can produce. A distributor forecast may fall, but the inventory decision may stay in place because manufacturing lead time makes a reduction more costly than holding extra units. A utilization result may miss plan, but the contract price may remain reasonable after the team finds that setup time, rather than demand, is limiting cases.

The team should be able to record one of four outcomes for each reopened decision:

• keep the decision because the new evidence does not change it;
• revise the decision and approve a new version;
• retire the decision because it no longer applies; or
• mark it not affected, with a reason.

This preserves judgment. Software can show the path from a changed assumption to a decision. It should not decide the commercial response.

Reopen commitments before they become sunk costs

The value of the register increases when decisions are linked to commitments.

A commitment is where an assumption begins to cost money or becomes difficult to reverse. Examples include hiring a rep, ordering inventory, signing a distributor agreement, giving the board a revenue forecast, scheduling training, or representing a launch date to a customer.

In the TransMedics case, the third-party transportation model sat beneath the delivery of NOP. Once TransMedics decided to address that dependence, the possible responses carried very different commitments: accept the risk, add more vendors, build an internal logistics team, acquire an operator, or buy aircraft. The company chose a path that added an aviation business, aircraft, crews, maintenance, regulatory duties, and capital requirements.

A smaller company will rarely face that scale of decision. The control is the same. When the assumption under a commitment changes, the team should see the commitment before it becomes harder to reverse.

If expected clearance moves by six months, a launch decision review should not stop at the regulatory date. It should show the reps whose start dates were based on it, the finished goods due to arrive, the training dates promised to pilot sites, the cash forecast, and the distributor plan. The regulatory update belongs to one function. The response belongs to the launch team.

What this does not prove

The public record does not show that TransMedics used an assumption register or a formal decision graph. It also does not prove that dependence on third-party aviation was a mistake. Using outside providers may have been the right way to launch NOP before demand and operating requirements were clear.

The case shows that a dependency stated in one year's filing became part of a major operating and capital decision in the next. It is an example of a launch model changing as the company learned what national delivery required.

The same discipline applies before the change becomes large. Write down the assumption, attach the evidence, name the owner, link the decisions and commitments, and set the condition that reopens them.

A launch plan is current only while the assumptions under it are current.

Sources

• TransMedics Group, 2022 Form 10-K, filed February 27, 2023: SEC filing
• TransMedics Group, 2023 Form 10-K, filed February 26, 2024: SEC filing
• TransMedics, "TransMedics Completes Acquisition of Summit Aviation," August 16, 2023: Press release
• TransMedics, "TransMedics Reports Fourth Quarter and Full Year 2023 Financial Results," February 26, 2024: Press release

Related reading: The Graveyard Manifesto · Finding Your MedTech Beachhead. This is a thought-leadership brief based on public filings, not investment or operational advice.