Before incorporating in Japan, test the approval process

Published Sep 16, 2026 Updated Oct 17, 2026

Japan usually reaches the board as an entity question. Do we set up a KK, hire a country manager, and fund a year of runway?

So the validation work goes where validation work goes: market size, competitive landscape, price acceptance, a few analyst calls. Six months later the answer is yes.

You incorporate, you hire, and four quarters after go-live the Japan pipeline looks nothing like the model.

The research was not weak. It answered a question that does not decide Japan entry.

Demand is something you can establish from your desk

Market size, who else is selling, what Japanese companies already pay: all of that is desk-researchable. So is the fact that more foreign companies are arriving.

JETRO’s 2025 report on inward investment recorded greenfield investment, the kind that creates a new legal entity, at 4.8 trillion yen, up 15.4% year on year and the highest on record. That is not evidence of demand for your product.

What it tells you is that the shortlist your Japanese buyer looks at is getting longer.

None of this requires an entity. Which means demand is not what the entity buys you information about.

What the entity buys you information about is one thing only: whether the person who wants your product can get it approved inside their own company. That is the variable you cannot read off a slide, and it is the one that decides the first two years.

The Japanese approval route stalls in places that have nothing to do with your product

In the survey by IDEATECH and Hiroyasu Kitagawa of Demagen Research (published 16 April 2026, fielded 23 to 26 March 2026, n=307, restricted to people who took part with at least one colleague in a Japanese purchase of business software worth 3 million yen or more, and run as survey-as-PR, so read it for order of magnitude), 60.9% of purchases required two approval stages, more than 80% required two or more, and roughly 20% required three or more.

Evaluation ran 3 to 8 months in 63.6% of cases, with about 40% taking six months or more.

Somewhere inside that route, your champion gets handed a set of questions.

Can the invoice be issued in yen. Can you issue a qualified invoice under Japan’s invoice system. Is there a Japanese-language contract, and what is the governing law. Is there a support window in Japan Standard Time. How does procurement run its credit check and its anti-organised-crime screening on a foreign counterparty.

None of these are product questions.

And they split into two groups that look identical from headquarters: the ones a Japanese entity solves, and the ones it does not. Incorporating before you know the split means the entity gets specified on instinct, and the first budget goes to whichever function your global playbook happens to fund first.

Why this pre-contact stretch has to sit inside the revenue model rather than outside it is covered in Japan Market GTM and Messaging.

Run ten deals with no entity and the map of where they stop is the spec for the entity

You can sell into Japan without an entity.

Contract from the foreign parent, sell through a distributor, or use an invoicing agent for yen billing. All three work commercially, but only after tax and legal review. Before the first deal, take at least these questions to Japanese tax and legal advisers:

whether your activity in Japan could be treated as a permanent establishment
how Japanese consumption tax applies, and whether to register under the invoice system (the buyer’s finance team will ask)
whether any payment is subject to withholding tax
which party contracts with the buyer (the foreign parent, the distributor or the invoicing agent) and under which governing law
what is required to handle Japanese personal data outside Japan

The answers depend on the structure and terms you choose, so this note does not give them. It is not tax or legal advice.

Then take five to ten named prospects and run them to the end. Offer terms you could actually deliver if they sign.

The champion spends their own internal effort routing your deal through approval, and a prospect you ran for research only will not take your call twice. Since the goal is not revenue, do not push hard on negotiation.

Record two things: the point at which the deal stopped, and the job title of the person who stopped it.

Keep the record somewhere it can be counted, not in each rep’s memory. In the CRM, add a picklist on the opportunity for the stage that stopped it (legal, finance, IT, credit check, final approval) and a field for the stopper’s job title. For five to ten deals, a shared sheet is enough.

The distribution tells you what to build.

Everything stops at legal review: what you need first is a Japanese-language contract template and a decided position on governing law and liability caps, not sales headcount.
Everything stops at finance: you need yen billing and Japanese payment terms, which is a treasury and billing-vendor decision more than an entity decision.
Everything stops at “we like it, but nobody in Japan has done this yet”: an entity does not fix that, and no amount of local hiring will.

Two alternatives were on the table and are worth naming. Commission a local research firm for a market study. Or publish a Japanese landing page and watch the response.

Both measure demand. Neither measures approvability, and both spend two quarters improving the precision of something you already knew.

”Company size mismatch” does not get fixed by incorporating

In the same survey, the top reasons a vendor was cut from the shortlist were insufficient functionality at 39.1%, company size mismatch at 34.5%, and no case study from a similar company at 26.4% (multiple response). The top reason a vendor was selected was having a case study from a company of similar size, at 44.3% (that question alone, n=298).

Registering a company in Tokyo does not add a single Japanese reference. The count stays at zero on the day you open and on the day your country manager starts.

So the objective of the validation period is not revenue. It is producing one or two customers whose names you are allowed to use.

Decide that before you start, or you will pick the first few accounts on likelihood of closing.

The selection criterion is not who will buy. It is who you can name in the next meeting: right industry, right headcount band, using the product the way your target segment would.

Do not validate Japanese demand before you incorporate. Validate whether a Japanese buyer can get you approved internally, by running live deals with no entity and recording where each one stops and who stopped it.

Common mistakes

Treating the entity as the thing that unlocks selling. It unlocks yen billing, local employment and a domestic counterparty. It does not unlock approval, references or precedent, and those are what the shortlist turns on.
Measuring the validation period on bookings. Bookings during validation are noise. The two useful outputs are a map of stall points and one or two nameable customers.
Choosing the first customers for close probability. The easiest logo to win is often the one you cannot cite, either because it is too small to matter to your target segment or because it will not approve the reference.
Reading a long Japanese evaluation cycle as hesitation. With two or more approval stages in more than 80% of purchases, three to eight months is the normal path, not a stalled one. Building a Japan forecast on the global cycle length produces a miss in the first year regardless of how the deals go.

Once you know where deals stop, the entry-route question changes shape, because a distributor sits between you and exactly that record. Distributor or direct: choosing your Japan entry model covers what has to be true for the reason a deal stalled to reach your own CRM.

Other notes on the same problem are collected under Japan market entry.