Does this apply if we sell direct?

Yes, in a lighter form. Independent reps, group purchasing organizations, and hybrid models all create versions of the same gaps. The readiness gate shrinks, but claims ownership, forecast definitions, and field-feedback format still need owners.

Is asking for named accounts and stage definitions realistic with a strong distributor?

It is a negotiation, and it belongs in the agreement or the operating plan attached to it, not in a side email. A distributor that will not say which accounts it will open first has told you something useful before signing.

What is the single most common failure?

Claims drift and forecast blur are tied, and both come from the same root: the manufacturer stopped treating field information as something it has to engineer. Neither failure announces itself. Both surface late, usually in a quarter-end review.

When should a manufacturer take a market direct?

When the value of account-level truth exceeds the cost of building the team, and not before. CapsoVision's sequence, exclusive U.S. distribution followed by a direct build, then selective direct conversion abroad, is one defensible pattern, not a universal rule.

The Distributor Decision Is a Governance Decision

Short answer

A distribution agreement settles territory and price. It does not settle which accounts open first, who approves claims, how much inventory is in place, what counts as a forecast, or when weak field data becomes an escalation. Until those have owners, the manufacturer has traded reach for visibility and may end up with neither.

The CapsoVision record

In November 2016, CapsoVision announced FDA 510(k) clearance for CapsoCam Plus, a small-bowel capsule endoscope. The capsule stores its images on board rather than transmitting them to a recorder the patient wears, which removes equipment from the exam and from the clinic. It was the company's third-generation capsule and the first it commercialized in the United States.

In June 2018, PENTAX Medical announced an exclusive distribution agreement under which it would serve as the exclusive U.S. sales agent for the CapsoCam Plus system, including the CapsoCloud data platform and CapsoView software. For the largest market, the account path now ran through a partner's sales organization.

That is a legitimate way to enter a market. It is also a transfer of information. Every account conversation, training gap, stalled evaluation, and reorder signal now happened inside someone else's field team.

The rest of the story is visible only at the level public filings allow. In its 2025 public-offering registration statement, CapsoVision reported that it began U.S. direct sales in 2020. It described a U.S. sales group built primarily on a direct model, using an in-house team, integrated group purchasing organizations, and, to a limited extent, independent sales representatives. Internationally, it reported approximately 50 exclusive distributors across non-U.S. regions as of December 31, 2024, with the top ten distributors accounting for roughly 67% of non-U.S. sales revenue in 2024. In 2023 it established a direct sales team in Germany, and it said it was transitioning to a direct model in some pivotal countries, including the G7. Revenue grew from approximately $9.8 million in 2023 to approximately $11.8 million in 2024.

The same filing describes what its distribution agreements do settle: purchase price, payment terms, minimum purchase obligations based on projected volume and growth potential, territory, an initial term of one to five years, renewal, and termination, including termination for failure to meet minimum purchase obligations, with a break-up fee and an obligation to honor the distributor's key customers for at least a year afterward.

Read the two halves together. The contract settles reach, price, and exit. Everything that determines whether the launch is actually working sits outside it.

The agreement settles the territory. It does not settle the launch.

A distribution agreement is a legal document about rights. A launch is an operating system about accounts, people, stock, messages, and evidence. The gap between the two is where channel launches fail quietly.

The contract will not tell you which twenty accounts open first, whether the distributor's reps finished training before or after their first visit, how many demo units are in the territory, whether the forecast means signed orders or hoped-for orders, who approved the claim on the local slide, or when three weeks of silence from the field becomes a problem someone must act on.

None of this argues against distributors. CapsoVision used exclusive distribution in the United States and still uses it across most of the world. A good distributor carries relationships, regulatory knowledge, procurement access, and local language that a small manufacturer cannot build alone. The argument is narrower: the day the agreement is signed, the manufacturer knows less about its own launch than it did the day before, and it has to build the mechanisms that buy that information back.

Run a readiness gate before the first account

Do not let a signed agreement start the clock by itself. Before the distributor's first selling visit, settle ten items, each with a named owner on both sides:

Territory and account ownership. Named accounts, not a map. Who owns each account, and what happens to house accounts and inbound leads.
Approved claims. The deck, the claims matrix, and the rule that no local claim ships without manufacturer sign-off.
Training completion. Who must be trained, to what competency standard, and who certifies it before the first visit.
Inventory and demo units. What stock, demo equipment, and consumables must be in the territory before selling starts.
First-account list. The named first five to ten accounts and why each was chosen.
Forecast definition. What a forecast means: stage definitions, evidence required, and update cadence.
Field-feedback format. The template and rhythm for what the distributor reports back, and to whom.
Escalation rules. The conditions that force a joint review, and who can call one.
Evidence-request routing. Where a clinical or economic question from an account goes, and the response time promised.
First-90-day review criteria. The specific results that keep the arrangement green, and the ones that reopen it.

Items two, six, and eight fail most often. Claims drift because drift is invisible until it is not. Forecasts blur because both sides want the number to be encouraging. Escalations stall because nobody wants to be the one who called the problem early.

Keep claims under one owner

A distributor's incentive is to sell. The manufacturer's obligation is that every claim in the field matches the cleared labeling and the approved evidence.

This is not a trust question. It is a systems question. The distributor's reps adapt material to their accounts because adaptation is what good reps do. Without a single approved deck, a claims matrix, and a named reviewer with a fast turnaround, the field version of the story will drift from the regulatory version within a quarter.

The practical control is boring and works: one owner, one current deck, a version date, and a stated rule that local material requires sign-off before use. The 90-day review should sample what accounts actually heard, not what the content library says they heard.

Treat the forecast as a defined term

CapsoVision's filing describes minimum purchase obligations based on projected sales volume and growth potential. That is the contract's version of a forecast. It says what the distributor must buy, not what accounts are about to do.

A launch needs the second kind. "The distributor forecasts $2 million" is unusable until the word forecast has a definition: named accounts, stage, the evidence for the stage, and the date the estimate was last touched by the person closest to the account.

Agree on the definitions in writing before the first pipeline review. A short set is enough: qualified, evaluating, trained, first use scheduled, reordering. Then hold the distributor's updates to it. When the manufacturer and the distributor use different definitions, every review meeting becomes a translation exercise and the truth lives in whichever version is presented second.

Demand field truth, not activity

The manufacturer's real risk is not a distributor that fails loudly. It is a distributor that reports motion while the launch stalls: visits made, demos scheduled, interest expressed.

Define the field report around evidence: which named account moved which stage, what the account asked for that the team could not answer, which claims were used, what inventory moved, and what the next dated event is. Activity counts belong in the report, but they are context, not progress.

CapsoVision's disclosed terms hint at the same problem from the contract side: termination for failure to meet minimum purchase obligations, and a duty to keep serving the distributor's key customers for a year after termination. Even the exit has a customer-continuity tail. Governance during the relationship is cheaper than governance at the end of it.

Review the channel at 90 days

Set the first review before the agreement is signed, and put its criteria in the readiness gate. Ninety days is long enough for training, first visits, first evaluations, and the first forecast cycle. It is short enough to correct course before a bad quarter hardens into a bad year.

Decision	Owner	Evidence required	Green	Red
First accounts	Manufacturer + distributor	Named account list	Prioritized by fit	Relationship-based only
Claims	Regulatory + commercial	Approved deck in use	Field message matches library	Local claims drift
Training	Clinical ops	Attendance and competency check	Users ready before visits	Training trailing sales
Inventory	Operations	Units, demos, consumables on hand	In place before visits	Field selling vapor
Data cadence	Commercial lead	Weekly account updates	Evidence-rich, stage-based	Activity counts only
Forecast	Commercial lead	First pipeline cycle	Definitions held	Number without accounts

A red row is not a failure. It is the review working. The failure mode is a review where every row is green because nobody defined what green required.

What this does not prove

The public record does not show why CapsoVision's U.S. arrangement with PENTAX Medical ended, and nothing in the filings says the arrangement failed. Companies take channels direct for many reasons: scale, margin, data, regulatory control, or a partner's shifting priorities. The registration statement's description of typical distributor terms is not a description of the PENTAX agreement, whose specific terms are not public.

The record also does not show how CapsoVision governs its roughly 50 international distributors, or whether the transition to direct sales in Germany and other pivotal countries will produce better results than distribution did.

What the record does show is the shape of the problem. A manufacturer signed over its largest market to an exclusive partner, then built a direct U.S. organization two years later. Its current disclosures describe contracts that settle price, minimums, territory, and exit, and say nothing, because contracts cannot, about account selection, claims discipline, training readiness, forecast definitions, or escalation. Those decisions are the governance layer, and the manufacturer owns it in every channel model, including the ones it does not control.

Choose the distributor for reach. Build the governance for truth. The first gets the product into accounts. The second tells you whether it is working.

Sources

Frequently asked questions

Does this apply if we sell direct?: Yes, in a lighter form. Independent reps, group purchasing organizations, and hybrid models all create versions of the same gaps. The readiness gate shrinks, but claims ownership, forecast definitions, and field-feedback format still need owners.
Is asking for named accounts and stage definitions realistic with a strong distributor?: It is a negotiation, and it belongs in the agreement or the operating plan attached to it, not in a side email. A distributor that will not say which accounts it will open first has told you something useful before signing.
What is the single most common failure?: Claims drift and forecast blur are tied, and both come from the same root: the manufacturer stopped treating field information as something it has to engineer. Neither failure announces itself. Both surface late, usually in a quarter-end review.
When should a manufacturer take a market direct?: When the value of account-level truth exceeds the cost of building the team, and not before. CapsoVision's sequence, exclusive U.S. distribution followed by a direct build, then selective direct conversion abroad, is one defensible pattern, not a universal rule.