Five signs, each with a price attached
Not "we have a lot of customers". Something more specific:
- Two people work one deal without seeing each other's actions. The client gets the same offer twice, or nothing at all.
- You cannot say how many deals are open without asking someone.
- The correspondence lives in a salesperson's personal inbox. They resign; the history leaves with them.
- Nobody knows where a customer came from. Which makes cost per customer impossible to compute at all.
- Repeat sales happen by accident, because no list exists of who is due a call.
The fourth is the expensive one. Without it the entire advertising budget is spent blind.
The real threshold is hands, not customers
The common advice online is "you need a CRM from 50 active customers". A tidy number, and beside the point.
A spreadsheet stops working not because of how many rows it holds but because of how many people change them at once. One owner with two hundred customers lives happily in a Google Sheet. Three salespeople with thirty deals do not, because they have no shared version of the truth: who called, what was promised, what happens next.
So the threshold reads: two or more people touching the same deal. From that point the spreadsheet is not a tool but a place where information goes missing, and the only question is what you are willing to pay for that.
Why implementations fail more often than they land
Bluntly: most failures are not about the system. They happen because the process was never described before it.
If you have no agreed answer to "what stages does a deal pass through and what has to happen to move to the next one", a CRM will not invent it. It will record the chaos in fields and add the obligation to fill them in. Two months later the salespeople are back in spreadsheets and the system is an open tab nobody looks at.
The second most common failure is implementing for the manager's reports. A system that gives the salesperson nothing and only extracts data from them gets sabotaged quietly and reliably.
The objection that voids this: with one person in sales and the process in their head, there is nothing to document. Install the simplest thing and get on with it.
Choosing without overpaying
The order is the reverse of the usual one.
Write out your deal stages first. From enquiry to payment, in your own words, on paper. Fewer than four stages and the simplest system will do.
Then write out what must integrate. Telephony, email, messengers, the website, the till, the warehouse. This is where the real cost hides: an integration into non-standard bookkeeping costs more than a year of subscription.
Only then look at systems. And check two things rather than a feature list: whether you can export your database at any moment, and whose account it is. Same story as an ad account held by an agency — the question is not the subscription but what remains yours when you leave.
Do not buy the system you like everything about except the price. Buy the one whose base tier covers you for a year.
When you do not need one
- You are the only person talking to customers, and deals close in one conversation.
- There are fewer than ten customers and you know each by heart.
- The sales process is not settled — you change how you work every month. Find something that works, then automate it.
- The actual problem is too few enquiries. A CRM will not bring a single new one; it will only stop you losing the ones you have.
That last case is the most common. If the pain is an absence of customers, a system for recording customers does not treat it.