What a CRM really costs in a small business
The subscription is the most visible line and almost never the most expensive one. The real costs sit in migration and adoption.
When you evaluate a CRM you look at the price per user per month. It's the line that reads most easily, compares against a pricing table on a website, and multiplies in five seconds. It's also, almost always, the cheapest part of the spend.
The other lines aren't in any price list, but you pay them all the same.
The costs that don't appear in the quote
Data migration. Customer records are never clean: the same client exists three times with three slightly different company names, phone numbers come in incompatible formats, half the email addresses bounce. No system imports dirty data and hands it back clean. Cleaning is human work, it has to happen first, and the software vendor doesn't do it.
Configuration. A CRM that's «ready to use» is ready for a generic process. Yours isn't, otherwise you wouldn't need to manage it. Statuses, fields, permissions, automations: that's days of work, and days of work by someone who knows your business — meaning one of the people you need doing their actual job.
The month you work twice. During the switchover the data is over there and the habits are over here. For a few weeks everything gets done in duplicate, and output drops. It's normal, it's expected, and nobody quotes for it.
Adoption. The highest cost and the least visible. A CRM your salesperson doesn't update isn't a CRM: it's an incomplete archive that gives the illusion of knowing. If the screens ask for ten mandatory fields that are useless to the person filling them in, those fields will be filled in at random — and from that moment your reports are worse than nothing, because they look like data.
Integrations. A CRM on its own is an island. The moment it has to talk to invoicing, to the online store or to your existing system, that piece has to be built. It isn't in the price list.
Getting out. Ask yourself now how you take your data away if you change your mind in three years. If the answer is a partial CSV export with no activity history, the exit cost is high and you'll find out at the worst possible moment.
The calculation that makes sense
Take three years, not one. Add up: subscription for the number of people who will really use it, days of configuration, days of data cleaning, the internal hours of the switchover period, the integrations you'll need in year one. Compare it with what not having it costs you today: retyping, forgotten enquiries, quotes never followed up.
It's a calculation that sometimes says buy something off the shelf, sometimes build it, and sometimes do nothing for another year. All three answers are legitimate. The one that never is: «let's take what everyone else uses and see how it goes».
What to ask before signing
Three questions, and insist on the answers in writing.
Who does the migration, on what timeline, and what happens to the data that doesn't fit the new model. How much it costs to add the field that isn't there today — because you will add it. And in what format the data comes out if one day you leave, history included.
If the vendor answers all three happily, the subscription is almost a detail. If they get defensive on the third, you already have the information you needed.