ERP
Outgrown Spreadsheets? A Decision Framework for SMBs
May 20269 min read
Photo by Lukas Blazek
Photo by Lukas Blazek
Every Indonesian SMB I have consulted for runs on spreadsheets at some point — and most of them should. Excel and Google Sheets are genuinely excellent tools: zero training cost, infinite flexibility, and everyone already has them. The interesting question is never whether spreadsheets are bad. It is whether your business has crossed the specific lines where their strengths flip into liabilities. Most owners feel that moment as vague pain; this post turns it into a checklist.
To be clear about scope: this is a decision framework, not a migration manual. The how of moving Excel data into an ERP is its own topic. What follows are the six signals I look for in the first hour of talking to a client, a scorecard you can run on your own business this week, and — just as important — the cases where my honest advice is to stay on spreadsheets longer.
I build ERPs for a living, and I will still say it: for a business under roughly ten people with one location, one person per function, and a few hundred transactions a month, a disciplined spreadsheet setup beats a badly chosen ERP. The spreadsheet does not need a server, never has a login outage, and bends instantly to whatever the business needs tomorrow. Flexibility is a real feature while the business is still discovering its own processes.
The trap is that the same flexibility silently accumulates risk as the team grows. Nothing announces the transition. The file that worked perfectly at five users develops version conflicts at eight, formula breakage at twelve, and a full-blown trust crisis at twenty. The framework below exists because the right moment to move is visible in operations long before it is visible in the accounting.
These come from real engagements — each one is a pattern I have personally watched cost a business money. One signal is a conversation; three or more is a decision:
1. The reconciliation ritual
Someone spends hours every week making two files agree — sales versus stock, bank versus ledger. That person is a human foreign key, and their vacation is your data outage.
2. Concurrent editing collisions
Two people need the same file at the same time, so you have FINAL_v3_revB filenames or an always-locked shared workbook. Multi-user writes are precisely what databases do and files do not.
3. Questions take minutes, not seconds
How much stock do we have right now, who owes us money beyond 60 days — if answers require opening three workbooks and running a mental join, decision speed is already throttled.
4. Anyone can change anything
No roles, no approvals, no history of who edited which cell when. The first serious dispute — with an employee, a vendor, or the tax office — turns absence of an audit trail into a real liability.
5. Formula fear
There is one workbook only one person dares to touch, because a broken VLOOKUP once silently corrupted a month of margins. Critical logic that nobody can review or test is unmanaged operational risk.
6. Volume is meeting the math
Excel caps at 1,048,576 rows per sheet, but practical pain arrives far earlier — multi-megabyte files that take a minute to open, recalculate on every edit, and crash on save. If you are archiving by year just to keep files usable, the data outgrew the container.
Feelings argue; scores decide. Answer these five questions honestly, sum the points, and the band tells you where you stand:
| Question | 0 points | 1 point | 2 points |
|---|---|---|---|
| People entering data weekly | 1-3 | 4-9 | 10 or more |
| Hours per week spent reconciling files | Under 1 | 1-5 | More than 5 |
| Decisions delayed waiting for a number | Rarely | Monthly | Weekly or worse |
| Errors found in customer-facing documents | Almost never | A few per quarter | Monthly or more |
| Processes needing approval before money moves | None | 1-2 | 3 or more |
Zero to three points: stay on spreadsheets, but add discipline — one master file per domain, protected sheets, a weekly backup. Four to six: you are in the transition zone; start scoping, budgeting, and cleaning your data now, because you will move within a year or two. Seven to ten: every month of delay is actively costing you money in labor, errors, and missed decisions — the move is overdue.
Pro tip: run the scorecard with three people separately — the owner, the person who maintains the spreadsheets, and one person who only consumes the reports. The spread between their scores is itself diagnostic: owners consistently score two points lower than the operators who live in the files.
An honest framework needs a negative branch. I have told prospective clients not to hire me yet in each of these situations:
The worst possible timing is moving to an ERP during a cash crisis because leadership hopes software will impose control that management has not. An implementation consumes attention and money for months before it pays back. Microsoft's own ERP guidance frames adoption around growth limitations and visibility — it is an investment made from stability, not a rescue performed mid-fall.
Spreadsheets feel free because their costs hide inside salaries: the admin spending Friday afternoons reconciling, the owner re-checking quotations because one went out 10 percent low, the stock bought twice because two sheets disagreed. When I help a client run this math, the recurring number that surprises them is labor — reconciliation hours alone often exceed what a modest system would cost monthly, before counting a single prevented error.
Against that: an SMB-grade ERP path in Indonesia — whether a hosted open-source system like Odoo Community, a SaaS package, or a focused custom build like the AP/AR module I built on NestJS and PostgreSQL for ANCoraPRO — typically lands in the tens to low hundreds of millions of rupiah for year one, depending on scope. The point of the scorecard is to know whether the hidden spreadsheet costs already exceed that figure. Frequently they do, and nobody had ever added them up.
The decision to move is not a purchase order for software. The right sequence starts before any vendor is contacted:
Spreadsheets do not fail loudly; they fail one reconciliation hour at a time. Run the scorecard, count the hidden hours, and make the call on numbers instead of frustration. And if the score says stay — stay proudly, tighten the discipline, and re-run the test in six months. The businesses that time this transition well are the ones that treated it as a decision, not a destiny.