DESAmedia ·

Short answer

A spreadsheet is enough while one or a few people work with the data and a mistake does not cost much. A business has outgrown it when the same file is sent around by email, data is entered twice, reports are compiled by hand and management only sees the figures at the end of the month. At that point, business process digitalisation, that is, moving the process into a shared system with a single database, usually pays for itself faster than it first appears.

Why a spreadsheet seems sufficient for so long

Almost every computer has Excel or Google Sheets, there is nothing to install, and you can build your first table in an hour. That is why most processes in a company start in a spreadsheet: the client list, the order log, stock levels, the staff rota.

The problems do not appear straight away. The table grows, more people start using it, copies appear, and formulas pile up that nobody remembers writing or why. From the inside this looks like normal work, so the decision to change tools is usually made too late.

Spreadsheet errors were studied for decades by Raymond Panko, a professor at the University of Hawaii. In his summary of audits of real company spreadsheets, the five most methodologically rigorous studies found errors in 91% of the 55 files examined. In a 2016 paper he concludes that an error in any single cell is rare, but in a large spreadsheet at least one incorrect bottom-line value is almost inevitable, and that authors are far too optimistic about the accuracy of their own files.

Below are seven signs that your process has already outgrown the spreadsheet.

1. The file is sent around by email

If a spreadsheet called "Orders_2026.xlsx" is emailed to colleagues every week, each recipient has a separate copy. After a few weeks the copies differ, and nobody can say which one holds the latest data.

Microsoft Excel does let several people edit the same file at the same time, but only if the file is stored in OneDrive or SharePoint Online and is in .xlsx, .xlsm or .xlsb format. The vendor's documentation states that if someone opens the file in a version of Excel that does not support co-authoring, the file is locked for everyone else. In other words, shared work only functions when everyone follows the same rules.

2. Nobody knows which version is correct

"Final", "final_v2", "final_REAL": familiar file names. When decisions are made from a spreadsheet and there are several spreadsheets, meetings begin with working out whose numbers are right.

A system solves this structurally: the data lives in a single database, and every change is recorded with its author and timestamp. You can see who changed a client's address or an order total, and when.

3. Data is entered twice

An order arrives by email, is retyped into a spreadsheet, then into the invoicing software and finally into the stock records. Every retyping costs time and is one more chance to make a mistake.

This is one of the clearest signs that you need a system. In a custom system, a piece of data is entered once and passed to other applications automatically through integrations. We explain how different applications are connected on our system integrations page.

4. Reports are compiled by hand

If a monthly report requires merging several files, copying columns, recalculating pivot tables and checking formulas, it takes a day or two. On top of that, the quality of the report depends on who prepared it.

Manual copying between spreadsheets has also played a part in large financial losses. In the JPMorgan "London Whale" case, where the bank lost more than 6 billion US dollars on trading in 2012, a new risk (VaR) model ran as a chain of several Excel files in which data had to be copied and pasted by hand. A later review found spreadsheet errors that understated the calculated risk. Researchers note that the model errors were not the main cause of the losses, but they allowed traders to go longer without seeing the true risk.

5. Errors are noticed too late

A spreadsheet does not warn you about an error. A formula covering the wrong rows, or a value deleted by accident, goes unnoticed until someone spots a strange result.

Two well-documented examples:

  • The Reinhart and Rogoff study. In 2013, economists Herndon, Ash and Pollin at the University of Massachusetts tried to replicate the results of an influential study on public debt and growth. They found a spreadsheet error that left five countries (Australia, Austria, Belgium, Canada and Denmark) out of the average, as well as questionable data selection and weighting. Once all the problems were corrected, average growth for countries with debt above 90% of GDP was 2.2%, not -0.1% as published.
  • England's COVID-19 test data. In October 2020 it emerged that 15,841 positive cases between 25 September and 2 October had not made it into the official statistics. Laboratory data was being imported into files in the old .xls format, which holds only 65,536 rows. Each test took up several rows, so one file could hold about 1,400 cases, and the rest were simply cut off without any warning.

The newer .xlsx format holds 1,048,576 rows, so the limit is rarely reached today. The lesson is different: in a spreadsheet, data validation is optional and easy to bypass, so the process does not stop when something is wrong. In a system, you can make sure that an order without a client or an invoice with a negative total simply cannot be saved.

6. Onboarding a new employee takes weeks

When the process rules are not written down anywhere and the file has dozens of sheets with colour codes, a new employee has to learn a host of unwritten conventions: which column not to touch, which sheet to fill in first, what a yellow cell means.

In a system, those rules are built into the interface. Fields are either mandatory or not, actions are only allowed in a certain order, and each employee sees only the functions meant for them. Training gets shorter because most mistakes are simply impossible to make.

7. Management only sees the figures at the end of the month

If answering "how many open orders do we have today?" means asking a colleague to update the spreadsheet, decisions are being made from last month's picture. A drop in sales or a stock shortage is noticed when it is already too late to react.

In a system, the figures are calculated from the same data employees work with, so the dashboard shows the current state. This is one of the main reasons companies build business management systems: management needs to see the process while there is still time to act.

Personal data in shared files

There is also a legal question to add to these signs. If a spreadsheet holds personal data about clients or employees, the General Data Protection Regulation (GDPR) applies. Article 5 requires data to be processed with appropriate security, including protection against unauthorised processing, loss or damage, and to be kept no longer than necessary. Article 32 requires ensuring that anyone with access to the data processes it only on instructions.

This is hard to achieve with a file whose copies have been emailed around. There are no role-based permissions, it is difficult to say who has seen the data, and a request to erase personal data has to be carried out in every copy separately. In a system, access is restricted by role, actions are recorded in an audit log, and data retention periods can be applied automatically.

When a spreadsheet or a ready-made tool is enough

Not every table needs a custom system. A spreadsheet is the right tool when:

  • one or two people work with the data and the file is not sent to others;
  • it is a one-off analysis, calculation or plan, not a daily process;
  • the table contains no personal data, or very little;
  • the process is still changing and needs to be tested. Here a spreadsheet makes a good prototype, because it shows which fields and actions are really needed.

If the process is standard and works in much the same way in thousands of companies, for example bookkeeping, a simple sales pipeline or time tracking, a ready-made application with a monthly licence is usually the better choice. A custom system makes sense when the process is your competitive advantage, or when ready-made tools would force you to change it in an inconvenient direction.

Situation Most suitable solution
One person, one-off analysis Spreadsheet
Standard process (accounting, simple CRM) Ready-made application
Several departments, the same data, specific logic Custom system
Data must reach several other applications Custom system or integration
Process not yet clear Spreadsheet as a prototype, decision later

An example from our own work is a tool tracking system with QR codes. In construction and service companies, tools are often tracked in a spreadsheet or a notebook by the storeroom door, and the records fall behind reality as soon as tools leave for a site. In the system, an employee scans a code with their phone and hands the tool over to a colleague, and every handover is saved in the history.

What staying in a spreadsheet costs

A spreadsheet looks free because its cost does not show up on any invoice. It is hidden in employees' time. Below is an illustrative calculation you can repeat with your own numbers.

Example (hypothetical figures, not a specific company):

  • Three employees each spend 6 hours a week retyping data between files and preparing reports. That is 18 hours a week in total.
  • We count 46 working weeks a year (after holidays and public holidays): 18 × 46 = 828 hours a year.
  • Hourly cost. According to State Data Agency (Lithuania's national statistics office) figures, the average monthly gross wage in the national economy (excluding sole proprietorships) in Q2 2026 was €2,628.2. Adding the employer's social insurance ("Sodra") contribution (1.77% for an open-ended employment contract, 2026 rates), this costs the employer about €2,674.7 a month. Assuming about 165 working hours a month, an hour costs about €16.2.
  • Result: 828 h × €16.2 ≈ €13,400 a year.

This is direct time only. The calculation does not include fixing errors, worse decisions made on late information, or training new employees. On the other hand, a system is not free either: it has to be built, maintained and updated. So a fair comparison is the cost of manual work over several years against the cost of building and maintaining a system over the same period.

If your figure is similar or higher, the alternative deserves serious consideration. If you arrive at a few hundred euros a year, a spreadsheet is probably still the right choice.

What to do next

Start with the one process that involves the most manual work, and count how many hours a week go into it. If you recognise at least three of the seven signs, take a look at how business management systems work and how they are built as custom web applications.

Sources