How to choose a CRM for a small business: seven questions
Before you compare feature lists, answer seven questions about who enters the data, how email gets in, where it lives and what it really costs. Then test on real deals.
Most small companies don’t need the longest feature list. They need a place where the next step on every deal is written down, and a team that keeps it true. A CRM that does less but gets used beats one that does everything and goes stale by March.
So before you open another comparison table, answer these seven questions. They decide more than any feature does.
1. Who will keep it up to date?
A CRM is only worth what’s in it. If two people sell and one of them hates forms, the tool has to cope with that.
Ask how a record gets written without anyone typing: does email land on the timeline on its own, can a website form create the contact, can you log a call in one line? Count the clicks to do the most common thing, which is writing down what happens next.
2. What does your pipeline actually look like?
Write your stages on paper before you look at any software. Most small teams need five or six, each defined by something the buyer did rather than something you hoped. We wrote a simple way to set up pipeline stages if you’re starting from nothing.
Then check that the tool lets you change those stages, flags deals that have stopped moving, and shows which deals have no next step. These alerts do more for your revenue than a dashboard.
3. How does email get in, and what can the tool do with it?
Connecting a mailbox is the feature people want most, and the one that deserves the most care. Three things to check:
- Read or write? A tool that can only read your mailbox can’t send anything by mistake.
- Whose email? Does it log everything, or only conversations with people already in the CRM? Newsletters and colleagues have no place on a customer’s timeline.
- What about the people who wrote to you? Creating contacts automatically from an inbox fills a CRM with noise, and with personal data you may not want to keep.
4. How many companies do you really run?
If you have one company and one product, skip this. But plenty of small businesses have two legal entities, or a product and a service sold under different names. Many CRMs treat that as one big account with filters.
Ask whether each company can have its own stages, products and mailboxes, whether someone can belong to one and not the other, and how the separation is enforced. A filter in the interface isn’t the same as a wall in the database. The row-level security entry in our glossary explains the difference.
5. Where does the data live, and can you leave?
Customer records are personal data, so the answer matters for GDPR as well as for peace of mind. Ask which region it runs in, who the hosting companies are, and whether you can export everything in a format another tool can read.
The test of a good exit is simple: export your records on the first day of the trial, and open the file.
6. What will it really cost in two years?
Add up seats, the plan you’d need for the features you actually want, integrations and API access. Then add the hours someone spends on it. With open-source software the licence can be free while the hosting and the upkeep aren’t, and we’ve broken down what self-hosting really costs.
7. Will you use it from an assistant?
More and more work starts in Claude or ChatGPT. If you’d like to ask for today’s follow-ups or log a call from a chat, look for a Model Context Protocol server, and read what to check before connecting one. What the assistant can’t do matters as much as what it can.
Test it on real deals, not on a demo
A trial on sample data tells you nothing. Spend two weeks on this instead:
- Import about fifty real companies and people from your spreadsheet.
- Put your five live deals on the board, each with a next step.
- Connect one mailbox, and see what lands on the timelines.
- Send one real quote, if the tool does quotes.
- At the end of the second week, count how many next steps are still true.
That last number is your answer. If half of them are out of date after two weeks, no feature will save it.
Where Soldlyne fits, and where it doesn’t
We make Soldlyne, so weigh this accordingly. It suits small companies that want the essentials (companies, people, deals, a pipeline, quotes accepted online), several companies on one install, and the choice of running it themselves.
It isn’t the right choice today if you need marketing automation, invoicing, or a mature product with a long track record: Soldlyne is under construction, and its screens are in English only. Better to know that now than in three months.