“Good enough” systems cost a Japanese SME far more than they appear to, because the cost is paid in staff time, workarounds and key-person dependence (属人化) rather than on an invoice. The spreadsheet, the shared folder and the employee who remembers everything work fine until the business tries something new: hiring quickly, handing over to a successor, selling, automating or using AI. At that point the weakness becomes visible and expensive. The fix is to map how the business really runs, sort what you find into constraints, habits and risks, and fix the risks first.

A process works. A spreadsheet works. A vendor relationship works. A staff member remembers the important details. The company gets through another busy season, and nothing is obviously on fire. That is exactly why the system survives.

Where does the cost of good enough systems hide?

Fragile systems rarely show up as a strategy problem. They show up as small annoyances. Someone cannot find the latest version of a document. A quote has to be rebuilt from an old email. A new employee asks the same question again because the process was never written down. A manager exports data and fixes it by hand before every meeting. A vendor invoice goes to the wrong address. A password reset takes two days because the account belongs to someone who left.

Each incident is small enough to ignore. Together they become an operating tax on the whole company.

Open file cabinet with index cards
Photo by Maksym Kaharlytskyi on Unsplash.

In Japanese SMEs this tax is especially hard to see, because people absorb it through effort. Staff stay late. Someone makes a phone call, someone apologizes, someone remembers the workaround. The business keeps running, which makes the system look sounder than it is.

Why do capable staff hide weak systems?

A capable office manager can make a poor workflow look fine. A senior salesperson keeps customer relationships alive outside the CRM. A bilingual employee quietly repairs every broken handoff between Japanese and English. The founder remembers why a pricing exception exists. A local vendor fixes problems because they know the business personally.

Those people are valuable. They are also carrying risk. If the business depends on their memory, private notes, inboxes or relationships, then part of the company is not really inside the company. It is inside a person.

That is normal in a young business. It should not stay invisible as the business grows. The aim is not to remove human judgment but to stop using one person’s memory as the only database.

When do good enough systems fail?

They usually fail at transition points:

  • Hiring. A new employee cannot learn the workflow without weeks of shadowing.
  • Departure or illness. A senior employee leaves or is off for a month, and the context goes with them.
  • Succession. The owner wants to step back and discovers every decision still runs through them. With so many Japanese SME owners reaching retirement age, this is one of the most common moments of exposure.
  • Sale or investment. A buyer’s due diligence asks for customer, contract and financial records the company cannot produce cleanly.
  • Foreign customers. Overseas clients expect online booking, English documents or self-service that the current setup cannot support.
  • Automation or AI. A project starts and immediately reveals that the input data is inconsistent and there is no single source of truth.

The system did not suddenly become weak. The weakness became visible because the business asked more of it. “We have always done it this way” can be a perfectly rational sentence about the old business and a dangerous one about the next.

Can Japanese politeness keep bad systems alive?

Japan is very good at making friction socially manageable. People apologize, wait, call back and adjust. They tolerate extra steps because the relationship matters. That is a strength, and it can also preserve processes that should have been rebuilt years ago. A broken workflow handled with good manners is still broken.

Foreign managers sometimes misread this as resistance or slowness. Japanese staff sometimes read the foreign side’s requests as impatience rather than a reasonable need for visibility. Both miss the underlying problem: there is no clear way to move work, responsibility and information between people. Better systems do not mean stripping out local etiquette. They mean less of the business depends on everyone politely compensating for a bad process.

Will AI fix a good enough setup?

No. AI is very good at making weak systems produce more output, which is not the same as making them better. If documents are scattered, AI summarizes the wrong version faster. If customer data is dirty, it produces confident nonsense. If the workflow is unclear, it creates drafts nobody knows how to check. Good enough systems depend on hidden context, and AI does not know that context unless the business has captured it somewhere usable.

Video: The Urgency of Digital Transformation in Japan | Historical and Geopolitical Factors of Japanese DX from Build+ (formerly Wahl+Case).

What does strong enough look like?

Small businesses do not need enterprise perfection. They need systems that are strong enough, which means the company:

  • knows which tools it uses, why, and who administers each one;
  • holds its accounts, domains and data in the company’s name, not an employee’s;
  • has one agreed place for customer records and active work;
  • can onboard a new employee without relying entirely on oral tradition;
  • can explain its sales process, vendor dependencies, renewal dates and bilingual handoffs;
  • can keep operating if any one person is unavailable for a week.

Strong enough still leaves plenty of room for judgment. It just stops confusing undocumented heroics with resilience.

How do you fix a good enough system?

Map it honestly. Record how the business really runs, including LINE messages, Excel files, paper forms, personal inboxes, unofficial vendor habits and the person everyone asks when the official process fails. Do not tidy it up for the diagram.

Sort what you find into three groups. Constraints exist because of law, customer expectations, local trust or the real shape of the business, so keep them and build around them. Habits once had a reason that has since disappeared, so remove them carefully. Risks still work but would fail if a person left, a tool changed or volume grew, so fix those first.

Start with key-person risk. For each critical process, write down what only one person knows, then document it or train a second person. The same goes for accounts: move admin access and recovery details to company-controlled addresses. I cover that side in treating domains, accounts and access as business assets.

Fix the source of truth before automating. Agree where customer, order and financial data lives, clean what matters, then consider automation.

This work is not glamorous, but it is what lets every later improvement stick.

Good enough is a phase, not a strategy

Every small business starts with good enough, and early systems should be scrappy. The danger comes when scrappy becomes sacred. A workaround that helped in year one can quietly cap the business in year ten. Many Japanese companies are at that point now: real businesses with real value, running on systems never designed for their next stage.

If you want a quick first read, the technology risk self-check takes a few minutes. For a full picture, a Diagnostics review maps your tools, workflows, records, vendors, costs and hidden dependencies, then sets out what to document, simplify, assign, automate or leave alone.

Further reading: why DX projects in Japan fail before the software arrives · what Japan Post’s lost truck license says about safety records · the Japanese SME stack nobody admits they are running