Tools
Spreadsheet, ERP or custom web app: when should an SME move off Excel?
You are not looking for a "web app." You are looking to get out of a file nobody dares to touch. An honest decision grid, maintenance included.
Key takeaways
- 77% of French very small businesses and SMEs use their data to run their business; among them, 58% use a spreadsheet and 67% use industry software, an ERP or a CRM (France Num 2026 barometer).
- A spreadsheet remains the right answer as long as one or two people use it and mistakes are easy to catch.
- There are three options: stay on the spreadsheet, adopt off-the-shelf software, or have a custom tool built. Custom only makes sense if the way you work is genuinely distinctive.
- Four questions settle it: how distinctive the process is, the number of users, the integrations, and the expected lifespan.
- A custom tool without maintenance becomes a risk. Maintenance is a selection criterion, not an option.
Everyone knows this file. It is called "Job tracking V7 final (2)". Only one person really understands it. It crashes when two colleagues open it at the same time. And every Monday, someone spends an hour putting it back in order. Should you replace it, and with what?
What are the signs that a spreadsheet has reached its limit?
A spreadsheet has reached its limit when it costs more time than it saves: multiple versions, re-entry, errors that spread, and only one person able to maintain it. One of these signs alone is not enough. Three are.
The five symptoms we see most often:
- Versions multiply. "The latest version is in Thursday's email." Nobody knows for sure which file is the reference.
- Data gets re-entered. The same information is typed into the quote, into the tracking sheet, then into the accounting system. Every re-entry is a chance for an error.
- An error is spotted too late. A broken formula, an overwritten row, and the month's revenue figure is wrong for three weeks.
- Only one person understands it. When she is on vacation, the file waits. If she leaves, it becomes a black box.
- Several people use it at the same time. A spreadsheet was designed for one user. With five, each with different rights, it can no longer keep up.
You are not alone. According to the France Num 2026 barometer report from the French Directorate General for Enterprise (September 2026, page 49), 77% of very small businesses and SMEs use their data to run their business. Among them, 58% use a spreadsheet and 67% use industry software such as an ERP or CRM, which is about 44% and 51% of all businesses. The two often coexist: software for accounting, spreadsheets for everything else.
What are the three options, and when is the spreadsheet still the right answer?
The three options are: keep the spreadsheet and make it more reliable, adopt off-the-shelf software, or have a custom tool built. The spreadsheet remains the right answer when one or two people use it, the volume is low and the process changes often.
Stay on the spreadsheet, but do it better. A single version shared online, locked tabs, drop-down lists instead of free-text entry, automatic backups. Sometimes a small automation is enough to eliminate re-entry between the spreadsheet and another tool: it is one of the seven tasks an SME should automate first. Low cost, quick results. And some files should stay spreadsheets: a budget forecast, for example, is for testing assumptions. Locking it into an application too early also locks in the reasoning.
Adopt off-the-shelf software. CRM, job site management, inventory, scheduling software: there is software for almost every trade. Advantages: ready to use, updates and security handled by the vendor, a predictable subscription. Limitation: you are the one who adapts to the tool. To test it, prepare a demo driven by your real cases: ask the vendor to create an order, modify it, handle a cancellation and export the file, in front of the people who will use the tool. Then classify each gap: essential, adaptable, or simply a habit.
Have a custom tool built. A web application designed around the way you work, accessible from a browser, with the right screens for each role. Advantage: the tool adapts to you. Limitations: a higher upfront cost, and a long-term maintenance responsibility.
Our view: most SMEs should look at off-the-shelf software first. Custom makes sense when none fits your trade, or when you would have to combine three of them to cover a single process. A fictional example: an SME prices its quotes based on combinations of dimensions, materials and delivery constraints, by hand across several files. A targeted tool can prepare the pricing and track its approval, without touching the accounting. The right scope is often a bridge: you keep your invoicing software and only build the business screen it is missing.
How much does each option cost?
There is no public price grid for a custom business tool: its cost depends on the number of screens, rules and connections. The closest benchmark is France Num's guidance for websites (France Num is the French government's program for small-business digital transformation): 5,000 to 10,000 euros for an "online platform" (our translation), excluding hosting, domain name and maintenance. Off-the-shelf software is paid monthly, often per user. A spreadsheet costs almost nothing in licenses, but a lot in time.
We have gathered these ranges, with their sources, in our article on the cost of an AI or digital project for an SME. Above all, remember the right way to compare:
- The spreadsheet: multiply the hours spent each week maintaining it, fixing it and re-entering data by the hourly cost of the people involved. Three hours a week is about 150 hours a year.
- Off-the-shelf software: the per-user subscription over three years, plus the time for configuration and data migration.
- The custom tool: the build cost, plus annual maintenance, over the tool's expected lifespan.
The calculation takes half an hour. It should be done before signing, whatever the option.
What are the four questions that settle it?
Four questions are enough to guide the choice: is your process really different from your competitors', how many people will use the tool, which software must it talk to, and how long must it last. The answers almost always point to one option.
1. Is your process distinctive? If the way you work resembles the rest of your industry, off-the-shelf software probably covers it. If it is what sets you apart (a pricing method, a quality tracking process, a way of organizing job sites that is your own), custom becomes relevant.
2. How many users? One or two: a more reliable spreadsheet is often enough. Five to thirty, with different roles (sales, technicians, management): you need a real tool, with per-person access rights.
3. Which integrations? Does the tool need to exchange data with your invoicing software, your email, your accounting? The more connections there are, the more you need to check that off-the-shelf software offers them, or plan to build them. The French e-invoicing reform of September 2027 makes the question urgent: your invoicing tool will have to connect to an approved platform.
4. What lifespan? A temporary need does not justify development. A tool that will carry your business for five years or more pays back its upfront cost better.
What are the risks of a custom tool?
The risks of a custom tool are dependence on the provider, ownership of the code, missing documentation and hosting. All of them are settled in the contract and in the way you work, before the first line of development.
Check these points before committing:
- Code ownership: are the source code and its usage rights transferred to you? Where is it kept, and could another provider take it over? A tool whose code you do not own ties you to a single provider.
- Accounts: are the domain name, hosting and technical accounts in your company's name? What happens if the provider, or the employee who manages them, leaves? If the tool contains personal data, also check where it is stored.
- Data: test the export for real, with attachments and the links between records. A contact file is not enough to take over an order history.
- Documentation: not all the code, but installation, backups, recovery after an incident and the main business rules.
A serious provider answers "yes" to these questions without hesitating. If they hesitate, you have your answer.
Why is maintenance a selection criterion?
A tool without maintenance degrades: the technical components it relies on receive security patches, browsers evolve, your needs change. After a few years, a tool that has never been updated becomes fragile, then risky. Maintenance must therefore be planned from the start.
It covers three things:
- Security: applying component updates, monitoring access, backing up data.
- Operations: fixing bugs, monitoring availability, running backups. A backup is only useful if restoring it has been tested: ask who checks it, and how the team works during an outage.
- Changes: adding a field, a screen or a connection when your business changes.
This is also a point where off-the-shelf software has a natural advantage: the vendor maintains the tool for all its customers. For a custom tool, compare the options on their total cost over time, maintenance included. Also get the support terms in writing: hours, how to report an issue, response time. The France Num 2026 barometer notes, incidentally, that only 26% of very small businesses and SMEs have access to external digital expertise, down 11 points: a tool must be able to live without depending on a provider for every minor change.
How do you start small?
To start small, pick a single process, aim for a first tool that is usable in about three months, and measure the gain before going further. A tool that covers one process well is worth more than a company-wide project that never gets finished.
- A single process. The one that costs the most hours or errors: job site tracking, quote preparation, scheduling of on-site work.
- A simple version first. The essential screens, for the main users. Everything else waits.
- A before-and-after measurement. Hours spent, errors, lead times. Without a starting figure, there is no way to know whether the tool kept its promise.
- Then a decision. Extend to a second process, stop there, or switch to off-the-shelf software if the need turned out to be standard.
If you are still unsure about the role of AI in this kind of project, our article on where to start with AI when you run a team of 10 to 30 people offers a method that starts from the problem rather than the tool. And if your need is limited to connecting existing software, n8n or Make is sometimes enough, with no development.
Polaris designs business tools for SMEs, as part of the projects we run with our clients. You can see examples in our case studies.
Tell us about the file you want to replace. In 30 minutes, we tell you whether it is better to make it more reliable, adopt off-the-shelf software or build a tool, and what each option would involve. Book your slot.
Frequently asked questions
- Is Excel really a problem for an SME?
- No, as long as one or two people use it, the volume is reasonable, and mistakes can be caught. It becomes a problem when several people work in it at the same time, versions multiply, and important decisions rest on formulas nobody checks.
- Off-the-shelf software or a custom tool: which should you choose?
- Start by looking for off-the-shelf software: if it covers 80% of your needs without bending the way you work, it is often the best choice. Custom makes sense when your process is distinctive or when covering it would mean combining several pieces of software.
- Can you start with a prototype?
- Yes, to test the user journey and business rules with future users. But a prototype does not validate security, data migration or backups: say so in the quote so you compare like with like.
- How long does it take to build a business tool?
- For a first version focused on a single process, expect anywhere from a few weeks to three months depending on complexity and the number of integrations. The projects that drag on are often the ones that tried to cover everything at once.
- Is maintenance really essential?
- Yes, for any tool used daily. Technical components receive security patches, browsers evolve, and your needs change. A tool that is not maintained eventually breaks down or exposes your data.