The real cost of a new tool isn't on the pricing page

They show you €49 a month, and you make the decision on that number. It is the smallest figure you will ever pay for that tool, and just about the only one you get in writing before you sign. This is for the owner about to buy something new — a CRM, a job-sheet system, a booking app — who wants the whole sum now rather than in March. It isn't an argument against buying software. It's an argument for knowing the price.

The sticker price is the only cost anyone shows you

Vendors compete on the visible number, so that one is sharpened to a point. The rest doesn't vanish — it moves into your team's week, where no invoice picks it up.

Here's what's missing from the quote you were sent:

• The licence you'll actually be on. Not the shop-window plan: the one you'll need three months in, when you find that automatic reminders or per-user permissions live one tier up. Count real seats, including the person who logs in twice a month and still takes a licence.
• Migration. Getting what you have out of where it lives and into where it's going. There's nearly always a manual slice left over, and someone on your payroll does it.
• Learning. Hours per person — and not the same hours for everyone.
• The month you run two systems. The overlap where old and new are both live.
• Glue. Connecting it to what you already run: invoicing, the calendar, WhatsApp. If it doesn't connect, somebody copies and pastes, and that's a new permanent task.
• Getting out. What leaving costs in two years. Ask on the way in, because on the way out nobody replies as quickly.

None of those six is a theoretical quibble. They are precisely the six things that make the €49 tool feel expensive by March, without your being able to explain why.

Learning hours land unevenly, and the heaviest ones are invisible

In a team of five or six, a new tool gets learned at three different speeds.

Someone picks it up in two afternoons. Good news — until you notice what happens next: that person becomes internal support for the other four. Every question routes through them. It's a real cost, it appears nowhere, and it's eaten by someone who had their own work to finish.

Someone else uses it halfway. They log in, do the minimum, leave fields blank or filled in wrong. This is the most expensive of the three, because half-entered data is worse than no data: you trust an incomplete list and make decisions from it.

And someone never opens it at all. They carry on with the notebook, the spreadsheet, the mail folder — and for months nobody notices, because their work still comes out fine. By the time you do notice, the history for that part of the business isn't where it should be.

So the sum isn't "two hours of training per person". It's two hours each, plus weeks at half speed, plus the hours of whoever ends up being the help desk — in a six-person business, easily two or three full days scattered across the calendar. The till pays for them either way.

Migration, and the month you run two systems

Migrating isn't exporting a CSV and uploading it. What exports cleanly is the record: names, phone numbers, amounts. What doesn't travel is exactly what made the record useful — the note that this client pays in 60 days despite agreeing to 30, the email thread where the discount was settled, the photo of the broken part they sent over WhatsApp. That lives outside the system, and it will carry on living outside it afterwards.

Two decisions remove most of the pain:

1. Only migrate what's active. Customers with movement in the last twelve to eighteen months. The rest doesn't move: it becomes a readable archive — an export, a PDF, the old database in read-only — and that's it. Migrating ten years of dead records means paying to clean data nobody will open.
2. Put a date on the cut-over. The overlap is necessary; nobody jumps off a cliff on a Monday. But it's the most expensive state your business can be in: double entry, constant doubt about which system is right, and mistakes that are nobody's fault. Two or three weeks is healthy. Three months means the switch never happened.

The warning sign is specific: if people still open the old system "just to check something" after the cut-over date, there was no cut-over. Either a field is missing in the new system, or a conversation is.

Do the first-year sum properly

Take a sheet of paper and write five lines. The numbers below are illustrative — use yours, because using yours is the whole exercise.

A six-person business buys a tool at €29 per user per month:

• Licences: 29 × 6 × 12 = €2,088 in year one. If the feature you wanted sits in the €39 tier, it's €2,808.
• Migration: 15–25 hours of someone on your payroll, or a vendor set-up fee.
• Training: 2 h × 6 people = 12 h, plus 10–15 h from whoever becomes the internal help desk.
• Overlap: three weeks of doubled work on the part of the process it touches.

Add up the hours and multiply by what an hour of your team genuinely costs — fully loaded, not gross salary. In that example you land somewhere between 45 and 60 hours, and that's where the bulk of the spend sits: the euros on the invoice are usually less than half of the first-year cost.

Don't run this sum to scare yourself. Run it because it changes the question. It's no longer "can I justify €29 a month?" — it's "does this tool give me back more than fifty hours in year one?" That second question tends to answer itself.

When the new tool is clearly worth it

None of the above says don't buy. There are four situations where the numbers come out comfortably ahead:

• It replaces two or three others. If the new one absorbs the work of three tools and all three are cancelled the same day, the whole sum changes — you save licences and, more importantly, you stop paying the "where was that again?" tax.
• It's the system of record for something the law requires. Invoicing, time records, clinical notes. There's no efficiency debate here: you need a serious place for it to live.
• What you have is genuinely holding you back. Not "it's clunky": it falls over, it won't let two people in at once, or the vendor has stopped updating it.
• The vendor does the migration and shows it to you before you sign. Your data, already inside, in a trial. That moves the risk off your side of the table.

And the inverse, worth saying even though it doesn't help us sell: if the process is broken, a tool won't fix it and automating it won't either. If nobody knows who answers quote requests today, quoting software will hand you unanswered quotes on a nicer screen. The same goes for an AI agent — automate a muddled process and it will be muddled faster. Settle who does what first, even if it's one sheet of paper, then decide which tool you need, if any.

Three questions before you sign

Take these to the sales call. The answers tell you more than the demo does:

1. Which plan do I need to do this specific thing? Name the feature that made you pick up the phone, and ask them to confirm in writing which tier it's in.
2. How do I get my data out the day I leave, and in what format? If the answer is vague, you've just learned your exit cost.
3. Which part of this process is still manual afterwards? It's the least popular question in the room and the best predictor of your March.

And a fourth — this one for you, not the salesperson: is the problem that you're missing a tool, or that the work of joining everything together is being done by hand, by you? If it's the second, buying software won't solve it. That's how we build custom automation: on the process and the tools you already have, without making you start over.