In this article
Spreadsheets are flexible, familiar and quick to start. They can serve a small, stable process reviewed by a few people well. The limit comes when the spreadsheet has to coordinate several people, control who can change what and show who changed each number. A file was not made for that.
To decide, compare the risk, the manual work and the need to connect routines to each other. There is no number of rows, users or Units that makes migration mandatory. The signs show up in daily work.
When each option makes sense
Direct answer
Use a spreadsheet when the process is small, stable, has few people responsible and can be checked manually. Consider a system when several people, Units, approvals, histories and related events need to work on the same information.
A spreadsheet can be enough for a simple list, an isolated calculation or a routine with a clear owner. It loses strength when the same data has to feed purchasing, stock, production, measurement and closing.
A system does not guarantee correct data. It can apply rules, permissions and validations, but it still depends on master data, checking and human responsibility. The decision should compare the current process with the control that is needed.
Signs that the spreadsheet has become a limit
The first sign is usually parallel versions. Another appears when a change has to be copied into several files or when the closing depends on someone who understands formulas that are not documented.
The literature on spreadsheet errors shows that faults can be hard to detect. This does not allow us to say that every file contains an error, but it reinforces the need for review, testing and controls proportional to the impact of the decision.
- Several people edit the same data
- There are copies with different names and numbers
- The operation serves several Units
- Purchasing, stock, production and billing need to be reconciled
- Users need different permissions
- You need to know who changed each piece of information and when
- Calculations depend on copies and manual formulas
- Errors only show up at closing
- Imports and integrations have become routine
Compare what each option records
A good-looking screen does not solve the problem if the system does not keep history, Unit, permission and the origin of the calculation. Likewise, a simple spreadsheet can be adequate when the decision is local and the person responsible can check everything.
List the critical events and test how each option responds. Does goods receipt need to update stock? Should an approval block a purchase order? Should changing a recipe sheet preserve earlier menus? The answers help define the real requirement.
Swipe the table to see all the columns.
| Need | Spreadsheet | Structured system |
|---|---|---|
| Isolated calculation | May be adequate | May be too much structure |
| Multiple users | Requires discipline and version control | Can apply access control and history |
| Several Units | Manual consolidation | Can separate Units and consolidate for management |
| Approval | Depends on an external flow | Can record state and person responsible |
| Link between areas | Copies and formulas | Can reuse the same record |
| Audit | Requires files and external records | Can preserve events and changes |
Prepare the migration before importing
Importing inconsistent data only makes the problem faster. Before migrating, choose the source of truth and clean up the master data for ingredients, units of measure, suppliers, recipe sheets and Units. Duplicate records need to be reconciled.
Also define the cut-over. During a parallel period, you need to know which system receives new changes and how the results will be compared. Without this rule, the team creates two competing sources.
- Define the authoritative source for each set of master data.
- Clean up duplicates, units and conversions.
- Choose the initial scope and the people responsible.
- Import in preview mode and handle rejections.
- Validate balances, documents and sample calculations.
- Run a parallel period with a defined editing rule.
- Reconcile the results before the cut-over.
- Limit editing of the old files and keep a backup copy.
How to validate that the migration worked
Counting imported rows is not enough. Validation should compare relevant records, totals, balances and relationships. A recipe sheet can exist and have no ingredients. A balance can match in total and be spread across the wrong item.
Use samples that cover every stage of the process. Pick a menu, calculate the requirement, check the stock, simulate a requisition and validate the meal count. The goal is to prove that the relationships the routine needs have been preserved.
- Number of records accepted, rejected and ignored
- Units and conversions of the main ingredients
- Physical balances and values by item and Unit
- Recipe sheets with ingredients and yield
- Roles and access by Unit
- History and documents needed for the cut-over
- Calculations reproduced on real samples
Access, backups and personal data
A spreadsheet or a system needs controls proportional to the risk. Backups, software updates, authentication, least-privilege access and incident response are important practices. Where there is personal data, the guidance from the ANPD (Brazil's data protection authority) can help structure security measures.
Adopting a system does not guarantee compliance with the LGPD (Brazil's General Data Protection Law). You need to understand which data is processed, why, for how long and who needs access. Nor should you assume that every operational spreadsheet contains personal data; that depends on the content.
How MEPO handles this topic
MEPO separates data by Organization and Unit, applies role-based access and records events in its flows. Ingredients and measurements can be imported by CSV or XLSX with preview, validation and confirmation. Implementation does not promise automatic migration from any ERP and should define scope, master data and validation according to the operation.
References
The sources support the general concepts. Specific rules can vary by contract, technical lead and local regulations.
- [1]Spreadsheet Errors: What We Know. Raymond Panko.
- [2]What We Don't Know About Spreadsheet Errors Today. Raymond Panko.
- [3]Revisão da literatura sobre erros em planilhas (review of the literature on spreadsheet errors). Tuck School of Business.
- [4]Guia de Segurança da Informação para agentes de pequeno porte (information security guide for small-scale data processing agents). ANPD.
- [5]ISO 10013:2021 e informação documentada digital (ISO 10013:2021 and digital documented information). ISO.