There is a conversation that happens in every growing company, and it is one of the strangest artefacts of how business software is sold.
A warehouse supervisor needs to check stock. A part-time bookkeeper needs to see two invoices a week. A seasonal hire needs to log their hours. In each case somebody works out what a licence costs, multiplies it by twelve, and decides the person can just ask a colleague instead.
That decision is rational. It is also the moment the company stops being able to trust its own data.
What per-seat pricing actually optimises for
Seat licensing is popular because it produces beautiful revenue curves for the vendor. Every hire your customer makes is an automatic upsell that requires no work. It is the closest thing software has to a tax on growth.
The trouble is what it does inside the customer. Once a login has a price, access becomes a budget line, and budget lines get rationed. In practice that means:
- Shared accounts, so nobody knows who actually posted the entry
- Data entered by proxy, one person typing what another person told them
- Spreadsheets running alongside the real system, because the spreadsheet is free
- Whole departments, usually the warehouse and the workshop, working from printouts
Every one of those is a data quality problem that the vendor created and the customer pays for twice: once in licence fees and once in the cost of untrustworthy records.
What actually costs us money
We run the platform, so we can see the bill. Nothing about hosting a company’s Site scales with headcount in any meaningful way. What scales is work.
Three things genuinely cost us money:
- Compute. The processing your Site consumes when it posts a document, rebuilds a report, runs a scheduled job or answers an agent.
- Database size. The live business records you hold: orders, ledger entries, employees, stock movements and their history.
- File storage. Everything you attach: contracts, scans, product photography, personnel files.
A hundred people reading dashboards costs us almost nothing. One nightly stock valuation over two years of history costs us something real. So that is what we charge for.
The uncomfortable part
Usage pricing means we make less money from some customers than we would if we billed per seat. A hundred and twenty person manufacturer running the whole suite pays us under a hundred euro a month. Under seat licensing that same company would be a five-figure annual account across four vendors.
We know. That is the point.
We are betting that a pricing model which encourages you to put your entire company inside the system produces a customer who never leaves, rather than a customer who spends every renewal cycle counting licences they can cut. And we would rather compete on whether the product is worth keeping than on how difficult we have made it to go.
What it changes in practice
The first thing companies do after switching is give everyone an account. Not as a policy decision, just because there is no reason not to. Within a month the warehouse is entering stock movements as they happen rather than on a clipboard, the part-time bookkeeper stops emailing questions, and the seasonal staff log their own hours.
None of that is a feature. It is the absence of a disincentive, which turns out to be worth more than most features.
The honest limits
Usage-based pricing is not free of edges. If you import a decade of transaction history you will use more database. If you run agents continuously across every department at volume, you will use more compute. Both of those can move you up a plan.
The difference is that both are visible on your dashboard before they arrive on an invoice, both track something your business actually did, and neither of them is triggered by hiring a person.
When a company outgrows a plan we contact them and move them up. No overage rate, no penalty, no suspended Site. It is a conversation about growth rather than a bill designed to punish it.
