When to Stop DIY-ing Your Business Infrastructure (A Japan SME Checklist)
Most small businesses start by figuring out their tools themselves. At some point, that approach stops working. Here are the signs it's time to bring someone in.
Every business starts the same way: the founder or a senior team member figures out the tools. They pick Slack because they’ve used it before. They set up a Google Workspace. They build a spreadsheet that tracks clients. They connect a few things together with whatever automation tool they find first.
This works. For a while.
At some point — usually somewhere between 8 and 20 people — the DIY approach starts generating costs that are hard to see until they’re very large. The spreadsheet breaks. The automation fails silently. The team is using five different places to track work and nobody knows which is current. A key person leaves and takes the institutional knowledge of how the stack works with them.
This is the point where most businesses should stop DIY-ing their infrastructure. Most of them don’t, because the costs are diffuse and invisible and there’s always something more urgent.
This checklist is designed to make the invisible visible.
The checklist
Go through these one by one. Be honest.
1. You don’t know exactly what you’re paying for software each month.
Not “roughly” — exactly. If you can’t produce a complete list of every SaaS subscription your business is paying for, with monthly cost and owner, in under 10 minutes, your stack is unmanaged.
2. At least one person has left your company and taken system knowledge with them.
The integration that person built that “just works.” The automation nobody else can explain. The CRM that was set up to their preferences. This is technical debt in human form. It hasn’t broken yet, but it will.
3. You’ve had at least one incident in the last year where a broken tool cost real time.
Not an inconvenience — real, measurable time. A workflow that stopped running. An integration that produced bad data. A vendor account that expired without warning. One incident is a warning sign. Two is a pattern.
4. Your team uses different tools to do the same thing depending on who you ask.
Some people log client notes in the CRM. Some send an email to themselves. Some keep a personal spreadsheet. This isn’t a behavior problem — it’s a system problem. The system doesn’t make the right behavior easy enough, so people optimize individually.
5. New team members take more than two weeks to understand how the tools work.
Onboarding time to software is a direct measure of system quality. If a new hire needs two weeks just to understand what tools exist and how they’re used, your stack isn’t documented and isn’t intuitive. This cost compounds with every hire.
6. Someone senior is spending more than two hours per week on IT tasks.
“IT tasks” means: managing vendor accounts, troubleshooting integrations, fielding tool questions from the team, evaluating new software, dealing with billing issues. If that person is your operations manager, your office manager, or you — that’s two hours per week of senior-level time going to problems that shouldn’t require senior-level attention.
At ¥5,000/hour (a conservative senior staff rate), two hours/week is ¥520,000/year. This is often more than a fractional management engagement costs.
7. You don’t know which automations you’re running or whether they’re working.
This is the scariest one. Most businesses that have been around for 3+ years have automations that were set up and forgotten. Zapier workflows from 2021. Make scenarios that a former contractor built. Integrations that sync data between two tools in ways nobody fully understands.
If you can’t name every automated process in your business, describe what it does, and verify it’s currently running correctly — your stack is running on faith.
8. You’ve bought a tool to solve a problem and the problem still exists.
This happens when a tool is purchased but never properly implemented. The CRM that nobody logs into. The project management platform that sits alongside the spreadsheet everyone actually uses. The documentation tool where nothing is documented.
Buying tools doesn’t fix problems. Implementing them does. If you have tools that aren’t solving the problem they were bought for, the issue is implementation, not the tool.
9. Vendors pitch you directly and you have no good way to evaluate them.
Every SaaS vendor with a Japan presence is pitching Japan SMEs. If you have no framework for evaluating these pitches — no clear picture of what you already have, what you need, and what good integration looks like — you will keep buying things that don’t fit.
10. Your stack would be hard to explain to someone who needed to run it.
Imagine explaining your complete software stack to a new senior hire who needs to manage it. Can you do it in under an hour? Is there documentation that would let them get up to speed without you? If the answer is no, the stack exists only in the heads of the people who’ve been there longest — and that’s a fragile single point of failure.
What the score means
0–2 yes answers: You’re in reasonable shape. Keep the inventory current and check back in 6 months.
3–5 yes answers: Some meaningful risk in your stack. Worth doing a basic audit yourself — inventory, check the automations, document what you find.
6+ yes answers: Your infrastructure is unmanaged. This is costing you time and money right now, probably more than you realize. It’s also accumulating risk that will surface as an incident at an inconvenient time.
What to do next
If you scored 6 or more, the most useful next step is a full picture of what you’re running. That’s what a Stack Audit provides — inventory, cost analysis, workflow mapping, and a prioritized list of what to fix.
If you scored 3-5 and want to stay ahead of it without a full engagement, a conversation is a reasonable starting point. Sometimes 45 minutes is enough to know whether you need professional help or can handle it internally.
If you scored under 3 — genuinely, you’re doing fine. Come back when something in this list changes.