ROI Case File No.484 'A Company Split into Five'
A Company Split into Five
Chapter One: Same Company, Numbers That Don't Match
"The same customer's information exists in five places."
Mayumi Takahashi, Management Planning Director at TechWave, lined up multiple systems on the screen. Inventory management, production management, sales management, customer management, accounting—each running on a separate system and Excel.
"When customer A places an order, we enter it into sales," Takahashi continued. "Check inventory on a different system, production on yet another, post to accounting after shipment processing, manage customer information on another Excel sheet. A transcription error anywhere, and month-end stops matching overall."
"How often do transcription errors happen?" I asked.
"Two or three cases are caught per week," Takahashi answered. "Sometimes they go uncaught until month-end. Uncaught errors only surface when the monthly report is compiled. Accounting cross-references with other systems and notices the numbers don't reconcile somewhere."
"How long does cross-referencing take?" Claude confirmed.
"From month-end to close, two accounting staff spend almost three full days on nothing else," Takahashi answered. "Cross-reference, investigate discrepancies, confirm with each owner, request corrections, re-aggregate. The monthly report, which would come out the next business day if systems were integrated, now takes four days."
"Why are you considering ERP introduction?" Gemini asked.
"Unified information management," Takahashi answered. "But a major ERP quote came back over ¥20 million in initial cost. We're a forty-person startup. That number we can't decide on. Yet inexpensive systems often can't handle Japanese business practices—multiple tax rates, bills of exchange, month-end invoicing."
"I heard you're considering open-source ERP too," Claude confirmed.
"We're looking at Odoo," Takahashi answered. "But no one in-house can customize it. We've been looking for an outsourcing partner with open-source ERP experience and Japanese business-practice support, but there are few. Selection has taken three months already."
"During the three months selection has stalled, the monthly work continues," I said quietly. Takahashi nodded slightly.
Chapter Two: The Five Domains 5F Demands
"This case needs 5F."
Claude wrote five words on the whiteboard. Strategy, Organization, Process, Technology, Finance.
"5F is a framework that analyzes corporate activity across five domains—strategy, organization, process, technology, and finance," I explained. "Company-wide transformations like ERP introduction fail if you only look at technology. Introducing technology doesn't mean it gets used if the organization doesn't move. If process isn't organized, what goes onto the system stays ambiguous. Designing five domains simultaneously determines ERP introduction success."
"Let's measure current cost first," Gemini said, opening ROI Polygraph. Takahashi's operational logs went in.
"Monthly fragmentation cost is out," Gemini read. "Double- and triple-entry labor across sales, production, inventory, and accounting staff averages 140 hours monthly. At ¥3,000/hour, ¥420,000 per month. Month-end reconciliation and adjustment by two accounting staff add 48 hours at ¥144,000. Customer-response delay from transcription errors: four cases monthly at three hours each, ¥36,000. Decision-making opportunity loss from monthly-report delay, estimated at ¥150,000. Total: ¥750,000 monthly arising from fragmentation. Annualized: ¥9 million."
Takahashi checked the figures. "Compared to the major ERP's ¥20 million, ¥9 million annually isn't a number to ignore."
"So, let's design with 5F," I continued.
[Domain 1—Strategy: Why Introduce ERP]
"First, articulate the purpose of ERP introduction," Claude said. "'Unified information management' is a means. The purpose is beyond. For TechWave, the purpose decomposes into three: faster monthly closing, zero transcription errors, scalability for the growth phase. These three become the basis for determining introduction scope and customization policy."
"Without a defined purpose, customization swells," I added. "Trying to fully localize every function to Japanese specs pushes initial cost toward major-ERP levels. Functions not directly linked to purpose should stay as standard features initially."
[Domain 2—Organization: Who Uses It, Who Grows It]
"Design the post-introduction operating structure," Gemini continued. "Currently, each department has independently selected systems. ERP is a cross-organizational system, so set up an operating committee with representatives from each department. Meet monthly to gather improvement proposals."
"What about employee resistance?" Takahashi confirmed.
"Staff attached to existing systems always emerge," Claude answered. "Put those staff on the operating committee. Pull the resisters into the design side. That's the standard play in the organization domain."
[Domain 3—Process: Organize the Business First]
"What fits to the ERP is the business process side," I continued. "Redesign business to fit the system, not the other way around. Cover 80% with Odoo standard functions, customize only the remaining 20%. Maintaining this ratio keeps both initial cost and maintenance cost down."
"When someone says 'this part must be kept' about current business, really verify whether it's needed," Gemini continued. "For example, unique approval flows, unique form layouts—many remain simply because they were born that way from historical context. Organizing reduces customization scope."
[Domain 4—Technology: How to Select Odoo]
"Technology selection is three stages," Claude continued. "First, Odoo base introduction and initial setup. Second, Japanese business-practice modules—existing open-source modules often work. Third, customization for unique operations. Checking completion at each stage makes it clear where costs are going."
"What about partner selection criteria?" Takahashi confirmed.
"Odoo Japanese-localization track record, accounting-integration track record, and post-customization code ownership belonging to you," I answered. "Unless open-source freedom is secured by contract, it's meaningless."
[Domain 5—Finance: Estimate the Payback]
"Let's simulate with ROI Proposal Generator," Gemini suggested.
- Initial cost: Odoo introduction, Japanese business-practice modules, customization, data migration, training totaling ¥6.8 million
- Monthly cost: Hosting and maintenance support ¥120,000/month
- Monthly savings: Double-entry elimination ¥420,000, reconciliation elimination ¥144,000, transcription-error response elimination ¥36,000, decision-delay relief ¥150,000, totaling ¥750,000
- Net monthly savings: ¥750,000 − ¥120,000 = ¥630,000
- Payback period: ¥6.8M ÷ ¥630,000 ≈ 10.8 months
"Within one year," Gemini summarized. "At major-ERP prices of ¥20 million initial, payback would take over two and a half years. Choosing Odoo reduces initial cost to one-third and payback period to less than half."
Takahashi reviewed the figures. "I thought it was a two-way choice between high-function ERP or cheap but un-localized. I didn't realize open source—the third choice—would change the numbers this much."
Chapter Three: The Courage to Fit Business to the System
"Let me organize the approach," I said at the whiteboard.
"Month one—launch the operating committee and conduct business-process inventory. Draw out each department's flow and clarify gaps with Odoo standard functions. Month two—partner selection and contracting. Months three through five—Odoo introduction and customization. Month six—data migration and parallel operation. Month seven—production cutover. Month eight onward—monthly improvement through the operating committee."
"Business-process inventory seems like it will take the most time," Takahashi said.
"That's the process with the highest value," Claude answered. "Introducing without organizing processes puts today's chaos directly onto Odoo. Even just for the sake of organizing, it's work that deserves to be done once, as a company with five systems."
Takahashi closed her notebook. "ERP introduction is finally the occasion for business review."
Chapter Four: The Day Numbers Became One
Nine months later, a report from Takahashi arrived.
Two months after Odoo production launch, monthly closing shortened from four days post-close to the next business day. "The state of the previous month-end is in executives' hands the next day," Takahashi wrote. Decision speed changed, and the agenda of the monthly meeting shifted from "reviewing last month" to "moves for this month."
Transcription errors had occurred twice weekly during parallel operation but held at zero after production launch. "Once entered, it reflects everywhere. No opportunity to transcribe, so no transcription errors," the report said.
During business-process inventory, seven of eleven unique approval flows were retired. "Flows with no reason to remain finally got cleaned up on the occasion of ERP introduction," Takahashi wrote.
The operating committee runs monthly. Members propose an average of eight improvements per month, half of which are implemented within the following month. "The field feels like they're evolving it themselves," the report noted.
Customization scope stayed within 20% of the original plan, with no additional cost. The richness of Odoo's standard functions was the precondition for reducing customization.
A day when a company split into five began running on one set of numbers.
"When information is split into five, the company is split into five. The purpose of ERP isn't information integration but the decision speed beyond it. 5F asks five domains—strategy, organization, process, technology, finance—and teaches that system introduction isn't 'a technology-only story.' Strategy decides the purpose, organization builds the operating structure, process organizes the business, technology selects, finance confirms the payback. When all five align, ERP finally takes root in the organization. The courage to fit business to the system made a company split into five into one."
Related Files
Tools Used
- ROI Polygraph — Visualizing fragmentation cost, transcription errors, and monthly delay cost
- ROI Proposal Generator — Open-source ERP payback simulation
