ERP Too Big a Leap? Start with Workflow Alone, Prove ROI, Then Expand

The boardroom conversation almost always goes the same way. The president of a mid-size manufacturer reads about digital transformation, hears that a competitor digitized its approvals, and asks the IT lead to look at an ERP. Two months later the IT lead comes back with a quote, a migration plan that touches sales, purchasing, inventory, and accounting all at once, and a training calendar that would pull every department head off the floor for weeks. The president looks at the number, looks at the disruption, and quietly shelves the idea for another year. Nothing changes. The paper approvals keep circulating, the receipts keep fading, and the gap with the digitized competitor keeps widening.

This is the ERP paradox for small and mid-size companies in Japan. They know they need a core business system. They cannot stomach the cost, risk, or disruption of a big-bang rollout all at once. The honest answer, and the one most successful adopters follow, is to start with one module, prove the return, and expand from a position of evidence rather than faith.

Why the Big-Bang ERP Roll Fails for Small Makers

A full ERP implementation is a bet. You commit money, time, and the attention of your best people on the promise that everything will be better when it is done. For a large enterprise with a project office, that bet can pay off. For a 280-staff maker with no one to spare, it is the wrong shape.

The cost is the obvious problem, but the deeper problem is disruption. A full rollout asks every department to change how it works in the same quarter. Sales learns new order entry while purchasing learns new vendor registration while accounting learns new closing procedures. When something breaks, no one is sure whether the fault is in the system, the data, or the process. The team loses confidence, and the project loses momentum.

There is also a timing problem baked into the Japanese calendar. Companies avoid changing systems near the month-end close, near the quarterly consumption-tax filing, and absolutely near the year-end close. That leaves narrow windows, and a full ERP rarely fits inside one. The rollout stretches across fiscal periods, which means parallel running, double entry, and a team that starts to resent the new system before it is even live.

The 2025 edition of the Small and Medium Enterprise Agency White Paper makes the stakes plain. The share of small and mid-size companies with zero digitalization dropped from 30.8 percent in 2023 to 12.5 percent in 2024, but 31.5 percent still have no digitalization plans at all. The cliff is real, and the companies standing still are the most exposed. The question is not whether to move, but how to move without betting the company on a single throw.

-> Related: The Manufacturer Workflow Catalog and Where the ROI Lives

Start with the Workflow Module Alone

The alternative to the big-bang rollout is to start with workflow alone. Workflow, in this context, means the engine that handles every internal request and approval: expense claims, leave applications, purchase requests, capital expenditure approvals, vendor registration, and the dozens of other ringi (internal approval proposals) and application flows a manufacturer runs every week. It is the layer where paper, email, and the hanko stamp still do most of the work today.

The reason workflow is the right place to start is structural. The workflow module can be deployed on its own first, with no dependency on the rest of the ERP. You do not need to migrate your chart of accounts, your item master, or your vendor master before you can file an expense claim or route a purchase request. You stand up the approval engine, move the highest-volume flows onto it, and measure what happens. Everything else in the core business system connects later, once you have proof the approach works.

This matters because it flips the risk. In a big-bang rollout, you spend the money and disruption up front, and you find out whether it worked at the end. When you start with workflow alone, you spend a small fraction of the budget, target the flows where the pain is loudest, and get a measurable result in the first quarter. If the result is good, you have built the internal case for the next module. If not, you have lost very little, and learned exactly where the friction is.

Consider what workflow replaces. A 2024 TOKIUM survey found that about 90 percent of employees still submit expenses by physically carrying paper receipts to the office, and that only 7 percent of companies have fully digitized expense processing. An atled survey found that 31.4 percent of companies still ran internal approvals on paper even during the pandemic. A SIOS survey from 2025 found that roughly 40 percent of companies that did adopt a workflow tool were stuck at merely electronifying the paper form, with no deeper benefit. These are not edge cases. They are the median, and the first-mover advantage available to a company that starts now is large. It does not require a full ERP to capture.

Prove the ROI on the Three Loudest Flows

The mistake teams make when they start small is choosing the wrong flows. They pick something rare or something only one department cares about, and the result is too thin to convince anyone. The right move is to target flows where the volume is high, the pain is shared across departments, and the savings show up fast.

The first flow is expense reimbursement. Every traveling sales engineer, every site-visit manager, every executive at a client dinner generates receipts. A conservative model puts processing at roughly 20 minutes per paper claim, dropping to about 5 minutes once the claim, the receipt image, and the approval live in one place. For a 280-person maker running a few thousand claims a year, that is hundreds of hours and several million yen, plus the input consumption tax that stops leaking when receipts are attached before submission instead of hunted for at month-end. The benchmark is steeper still: providers report expense processing dropping from 3 hours to 30 minutes end to end.

The second flow is internal approvals, the ringi that drive purchasing, pricing, and capital decisions. When these circulate on paper with the hanko stamp, a single approver on a business trip freezes the whole chain. One widely cited case saw a company cut approval decision time by 7 days and eliminate 4,000 hours of handling cost after digitizing ringi. A conservative move from 60 minutes to 10 minutes per approval yields thousands of hours and real cycle-time improvement on orders that previously sat waiting for a signature.

The third flow is purchase requests and approvals. Email and spreadsheet ordering quietly breaks the authority rules in the approval policy, the thresholds that say who can commit what amount. Month-end budget overruns get discovered after the fact, not at the point of request. Routing purchase requests through the workflow engine, with amount thresholds that escalate to the right role, restores the rule without forcing anyone to memorize it.

These three flows share the same shape: high volume, every department, a number you can compare before and after, and none require the rest of the ERP to be live. You prove the ROI on workflow alone in one quarter, with a budget that fits inside a normal operating expense.

Being Honest About What Connects When

A start-small approach only works if the vendor is honest about what is built today and what connects later. The workflow engine handles every request type with its own configurable form, so no developer is needed to add a new application. Approvals route to the role, position, or department, so they survive reorganizations. Large spend can require committee sign-off with a quorum, or any-of-N approvers for speed. Routing can follow the requester's own manager dynamically. You can see exactly where every request is stuck. Approvals do not freeze when a manager travels, because safe delegation is built in, with mandatory re-approval for high-risk items. Business-day deadlines, watchers who can follow a request without being approvers, and a frozen snapshot of what was approved all come standard.

The part that matters for a phased rollout is the writeback boundary. When an expense reimbursement is approved, the accounting record is created automatically, because that writeback is built. When a leave application is approved, the leave record is created automatically, because that writeback is also built. These two close the loop inside the workflow module itself.

For the wider core business system, the picture is on the roadmap, not built today. Automatic creation of journal entries, purchase orders, vendor master records, bills, invoices, fixed-asset records, and lot records straight from an approval is designed for but not yet implemented. This is exactly why starting with workflow alone makes sense. You capture the time and risk savings now, with the two writebacks that are built, and sequence the rest of the ERP connection once the first wave has proven itself. The honesty here is not a limitation. It is the roadmap.

A Scenario: The Precision Parts Maker in Shizuoka

Picture a precision parts maker in Shizuoka, about 280 staff, supplying automotive and industrial machinery OEMs from a head office, a factory, and a sales branch. The president has watched two competitors digitize. The IT lead has quoted a full ERP twice, and twice the president has paused it. The risk feels too large for a company where the month-end close already runs on overtime.

They start with workflow alone. The first quarter targets expense reimbursement, paid leave, and purchase requests. Sales engineers file claims from their phones the evening of a trip, with receipts attached to lines and consumption tax captured by rate. Managers approve during the commute. Paid leave requests route to the right team lead based on the requester, not on whoever happens to be in the office. Purchase requests above the threshold escalate to the department head, and above the higher threshold to an officer, exactly as the approval policy always said but never enforced.

The result in the first quarter is visible. Month-end expense reconciliation, which used to consume the finance team for days, drops to hours. The input consumption tax that used to vanish with faded receipts is captured in full because the receipt is attached before submission. Purchase requests that used to slip past the authority limits now stop at the threshold. The IT lead takes the before-and-after numbers to the president, and the conversation about the next module starts from evidence instead of hope. From there, the company connects purchasing, then inventory with lot tracking, then manufacturing and BOM, then the full accounting close, each phase justified by the one before.

The 2025 Pressure and the Labor Shortage

Two forces make the start-small path urgent right now. The first is the 2025 legacy cliff. METI has warned for years that companies running on unsupported legacy systems face a growing maintenance and security risk, and that small and mid-size manufacturers are the most exposed. Starting with workflow alone is a concrete first step off the cliff that does not require a full system replacement in one move.

The second is the labor shortage. The 2026 edition of the Information and Communications White Paper found that 48.7 percent of Japanese companies cite the talent shortage as the top barrier to digitalization. That figure cuts both ways. The shortage is a reason companies feel they cannot spare the people for a big project, and it is also the reason they must automate the paper work that those same scarce people are stuck doing. A phased rollout respects the first reading and answers the second, freeing back-office hours now with a small first step instead of a full migration on top of the day job.

A start-small path also lets you ask the right questions of any vendor. Can the workflow module run on its own without the rest of the ERP? Which approvals create downstream records automatically today, and which are on the roadmap? A vendor that distinguishes built from roadmap is one you can plan a sequence around.

Key Takeaway

A full ERP is the right destination for a growing manufacturer, but it is rarely the right first step. Start with the workflow module alone, target the loudest flows, prove the return in one quarter, and expand from evidence. You move off the 2025 cliff, answer the labor shortage, and build the internal case for the rest of the core business system without betting the company on a single throw.

Frequently Asked Questions

Can the workflow module really run on its own without the rest of the ERP?

Yes. The approval workflow engine deploys standalone first, with no dependency on the chart of accounts, item master, or vendor master. You file expense claims, leave applications, and purchase requests through it and measure the result before anything else connects. That independence is exactly what makes a phased rollout possible.

Which approvals write back to records automatically today?

Kikan System writes the accounting record automatically when an expense reimbursement is approved, and creates the leave record automatically when a leave application is approved. Those two writebacks are built and close the loop inside the workflow module. Automatic creation of journal entries, purchase orders, vendor records, and invoices from an approval is on the roadmap, not built today.

How does a phased rollout survive a manager traveling or on leave?

The workflow engine supports safe delegation, so an approver hands their queue to a deputy while they are away. For high-risk items the system can require the original approver to re-confirm after they return, so the safeguard holds. Business-day deadlines and watchers who follow a request without being approvers keep the flow visible even when key people are out.

What ROI should we expect in the first quarter?

Teams that start with workflow alone typically target expense reimbursement, internal ringi approvals, and purchase requests because those flows are high volume and cut across every department. Processing time per expense claim and per approval drops sharply, month-end reconciliation speeds up, and input consumption tax stops leaking when receipts are attached before submission. You can start on the free plan with up to 2 users and no credit card required, then take the before-and-after numbers to leadership to justify the next module.

Get Started With Kikan System

If a full rollout feels too large, start where the ROI is loudest. Kikan System ships the approval workflow module on its own, with built-in writeback for expense reimbursement and leave applications, and a clear roadmap for connecting purchasing, inventory, and accounting as you grow. Try the free plan, which covers up to 2 users with no credit card required. Begin at → Start free, then expand when the numbers say so.