- Why do renewal numbers fall in year two when the launch went well?
- Launch is carried by the landing page and a founder or GM paying close attention. Renewal is carried by whatever customer success has been quietly sending since onboarding ended — renewal notices, invoices, support replies, in-app messages. That material is usually built once, at launch, and never revisited.
- Is this a product problem?
- Usually not. Product fit and pricing are the first things teams check, and they often check out fine. What's easy to miss is the lifecycle communication layer between onboarding and renewal — and whether it still reads like an afterthought.
TL;DR
Year one in Japan tends to go fine: a localized launch page, a first cohort of logos, maybe a press mention or two. Year two, renewal and upsell numbers come in soft, and the usual suspects — product fit, pricing, competition — check out clean. What we keep finding instead is a gap in the parts of the customer lifecycle that nobody re-localizes after launch: the renewal notice, the invoice, the macro a support agent pastes into a ticket, the in-app message that ships six months after onboarding ends. None of it was ever treated as "the product." It was treated as operational output, often written once in English and translated — or left in English outright — and then never touched again while the launch page kept getting attention. Japanese customers who are quietly unhappy with this rarely escalate; they let the relationship lapse instead, which means the absence of complaints tells you nothing. The fix isn't a bigger localization budget. It's treating customer success communication as a living surface that needs the same ongoing care as the site that won the account in the first place.
Key Takeaways
- Launch and renewal run on different material. A landing page wins the deal. Renewal notices, invoices, and support replies keep it — and those are usually built once and forgotten.
- Product and pricing checks often come back clean. When renewal slips, the first instinct is to audit the product. The actual gap tends to sit one layer down, in day-to-day account communication.
- Silence isn't the same as satisfaction. Customers who find lifecycle communication cold or awkward in Japanese usually don't complain about it — they just quietly decline to renew.
- The un-owned surfaces are predictable. Renewal notices, invoice line items, support macros, in-app messages, and win-back emails are the ones that ship after launch and rarely get reviewed again.
- This is a lifecycle problem, not a one-time project. A page localized once for launch day doesn't stay accurate for the messages a customer sees eighteen months later.
Year one looks like it worked, because it did
We don't want to undersell what a good Japan launch actually accomplishes. The site reads naturally, the pricing page doesn't trip anyone up, the sales deck holds up in a room full of skeptical stakeholders. A first cohort of accounts signs. Somebody writes a case study. For a foreign SaaS or fintech company, that's a real milestone, and it deserves to be treated as one.
The trouble starts because that milestone becomes the story everyone tells about the Japan operation — including, often, the GM running it. Launch was the hard part; everything after should be maintenance.
Then the second-year renewal number comes in soft
Renewal targets miss. Upsell conversations stall or don't happen at all. Net revenue retention for the Japan book lags behind other regions, and it's not a one-quarter blip — it holds for the cohort that's now in its second contract year. The GM pulls the usual levers: checks whether the product still fits, whether a competitor undercut on price, whether the account owner dropped the ball on a specific renewal call. Individually, those checks often come back clean. The product works. The price is fair. The account team did their job on the calls they had.
What doesn't get checked, because nobody thinks to
What rarely gets pulled up and read end-to-end is the material a customer actually receives between onboarding and the renewal date — most of which was never part of the original launch scope, because it didn't exist yet when the launch page was being built.
The renewal notice itself is a common one: often a template written in English first and translated once, sometimes by whoever on the team happened to have the strongest Japanese, sometimes by a translation tool with no review pass after. The invoice is another — line items, tax display conventions, payment terms that read as slightly off to someone used to Japanese business paperwork, even when the numbers are correct. Support macros are a third: canned replies a support team pastes into tickets, written for a different market and adapted just enough to pass a first glance, not enough to sound like they were written for the person reading them. In-app messages and changelog notes that ship long after onboarding ends are a fourth — and by the time a customer is in their second year, they've seen a lot of these.
None of this was ever treated as "the product." It was treated as operational output — necessary, but not something anyone put through the same scrutiny as the site that won the account.
The pattern in one line: localization gets budgeted and reviewed hardest at the moment a customer is being won, and least at every moment after — which is exactly backwards from where a renewal decision actually gets made.
Why nobody tells you this is happening
A customer who finds a renewal notice stiff, or notices that a support reply clearly wasn't written with them in mind, is unlikely to file a complaint about it. In our experience, the more common pattern is quieter: the account doesn't escalate, doesn't push back, doesn't ask for a call to "discuss the relationship." It just doesn't feel like a priority account to the person on the other end, and when the renewal date arrives, that feeling is part of what they're weighing — alongside the product, the price, and everything else. The absence of complaints tells a GM nothing about whether this is happening. It's one of the reasons the second-year cliff is so easy to miss until the number is already in: the leading indicator isn't a support ticket. It's an account that's gone quiet.
What a real audit of this looks like
If you want to check whether this is happening in your own Japan operation, the useful exercise isn't a translation quality review of your website. It's pulling the actual files a second-year customer receives and reading them as that customer, not as a proofreader.
The last renewal or contract-change notice you sent
Read it as the recipient, not the sender. Does it read like something written for a Japanese customer, or like something translated for one?
The current invoice template
Line items, tax display, payment terms — do they follow conventions a Japanese finance department expects, or the conventions of wherever the template originated?
The last three support replies sent to a Japanese account
Were they written for this customer, or adapted from a macro built for a different market? Would the customer be able to tell the difference?
In-app messages and changelog notes from the last quarter
These usually ship without the review the launch page got. Over eighteen months, a customer sees a lot of them.
Who owns this material now — and who reviewed it last
If the honest answer is "whoever wrote it at launch, and nobody since," that's the gap, independent of how any single message reads.
What actually closes the gap
The fix isn't a bigger localization budget or a full-site retranslation. It's narrower and less glamorous than that: treating customer success communication as a surface that needs the same ongoing ownership as the site that won the account, instead of a one-time deliverable from the launch project. That means someone specific is responsible for reviewing renewal notices, invoice templates, and support macros on a schedule — not only when a customer complains, since by the time that happens the quiet cases have often already declined to renew. It also means writing new lifecycle material for a Japanese customer from the start, rather than writing it in English and translating it once a decision-maker is annoyed enough to ask.
None of this is dramatic work, which is part of why it gets skipped. But it's the work that sits directly between a launch that looked like a win and a renewal number that quietly doesn't confirm it eighteen months later.
Launch was the beginning of the relationship, not proof of it
A strong year-one launch tells you the market entry worked. It doesn't tell you the relationship is durable — that gets tested every time a renewal notice, an invoice, or a support reply lands in a customer's inbox, long after the launch page has moved on to being old news internally. Watching what happens to those touchpoints as the account ages is, in our experience, a better predictor of the second-year number than anything the original launch metrics can tell you.
Frequently Asked Questions
Why do renewal and upsell numbers fall in year two even when the Japan launch went well?
Because launch and renewal draw on different material. Launch is carried by the landing page, the sales deck, and a founder or GM who's paying close attention. Renewal is carried by whatever customer success has been sending in the background since onboarding ended: renewal notices, invoices, support replies, in-app messages, contract paperwork. That material is usually built once, during launch, and then left alone. If it was ever machine-translated or written for an English-speaking template and never revisited, it quietly ages out while the customer relationship keeps running.
Is this a product problem or a churn problem?
Usually neither, at least not first. Product fit and pricing are the first things teams check when renewal slips, and they often check out fine — the product still works, the price hasn't moved. What's easy to miss is the lifecycle communication layer: the emails, notices, and support interactions a customer has with you between onboarding and the renewal date. If that layer reads like an afterthought, customers experience it as one, regardless of how good the product is.
Why don't unhappy customers just say something before they leave?
Some do, but plenty don't escalate at all — they simply let the relationship lapse. A customer who's quietly annoyed by a renewal notice that reads like a direct translation, or a support reply that clearly wasn't written for them, doesn't usually file a complaint about it. They just don't feel like a priority account, and when the renewal date comes, that feeling is what they're weighing. The absence of complaints is not the same as the absence of a problem.
What parts of the customer lifecycle are most often left un-localized after launch?
The pattern we see most often: renewal and contract-change notices, invoices and billing line items, the macros a support team pastes into tickets, in-app messages and changelog notes that ship after the onboarding sequence ends, and win-back or downgrade-prevention emails. These are usually not part of the launch checklist because they don't exist yet at launch — they get built later, under time pressure, by whoever is closest to the task, and rarely get the same scrutiny the original launch page did.
How do you check whether this is happening in your own Japan operation?
Pull the actual files a second-year customer receives — the last renewal notice, the last three support replies, the current invoice template, the in-app messages sent in the last quarter — and read them as if you were that customer, not as if you were checking a translation for accuracy. Ask whether it reads like something written for a Japanese customer, or like something translated for one. A short, structured review of exactly this material is what a Japan Renewal Readiness Check is built to do.