Most CRM evaluations compare feature lists. Feature lists are close to useless, because every product in the category has a pipeline, contacts and reporting, and the differences that matter never appear in a comparison table.
What predicts regret is the stuff you find out in month eight. These nine questions surface most of it before you sign.
1. What does this cost at the plan I will actually be on?
The plan you sign up on is rarely the plan you stay on. Find the cheapest tier that includes everything on your genuine must-have list, ignore every tier below it, and price that.
Then add the one-time fees. Some vendors charge mandatory onboarding on their mid-tier plans, which can add several thousand to year one and is easy to miss when comparing per-seat numbers.
Follow-up worth asking: which features live one tier above the one I am quoted?
2. Is there a seat minimum, and are seats sold in bundles?
Some products require a minimum number of seats. Some sell seats in fixed blocks, so a team of four pays for five and a team of six pays for ten.
Neither is unreasonable, but both change the arithmetic significantly for small teams, and neither is prominent on a pricing page.
3. What is sold as an add-on?
Ask for the full list of paid extras and their prices. Documents, e-signatures, lead generation tools, email campaigns, web visitor tracking and project management are commonly separate.
You may not need any of them. The point is to know the ceiling before you commit, rather than discovering it the week you need a feature.
4. How do deal values get into the system?
This one sounds mundane and it is the best predictor of whether your forecast will be worth anything.
If deal value is a free text field, it will drift. A rep enters an estimate early, the scope changes, nobody updates it, and by the end of the quarter your pipeline total is a collection of guesses of different ages.
Ask whether values can be assembled from a product or service catalogue instead, so the number is calculated rather than typed. If not, you will be managing that drift manually forever.
5. What happens after a deal is won?
If your revenue recurs, this is the question that separates products.
Ask specifically: does the system track what was sold, does it know when the term ends, and does it create the renewal, or is that a reminder somebody sets for themselves?
Most CRMs treat the win as an ending. If yours does and your business is renewal-based, you are buying a tool for half your lifecycle.
6. What can I export, and how easily?
Ask whether you can export every module to CSV, on demand, without contacting support.
The answer tells you two things. Whether you can leave, obviously. But also how the vendor thinks about the relationship, because products confident in their retention make leaving easy and products that are not make it a support ticket.
7. Who can see and delete what?
Get specific about permissions. Can a rep delete a contact? Can a manager change pipeline stages? Who controls billing?
Then ask how those permissions are enforced. Some products hide buttons in the interface but do not enforce the rule at the server, which means the restriction is cosmetic. It is a fair question and the answer is revealing.
8. How long until my team is actually using it?
Not how long to set up an account. How long until reps are working real deals in it without prompting.
Ask the vendor for an honest answer and then ask a reference customer the same question. If the honest answer involves a consultant, a phased rollout or a training programme, factor that time and cost in, because it is real even when it is not invoiced.
9. What does this product deliberately not do?
The most useful question on the list, and the one that gets the most interesting answers.
A vendor who cannot name anything is either not listening or selling you scope you will pay for and never use. A vendor with a clear answer has made deliberate choices, and deliberate choices are what keep a product coherent as it grows.
You are looking for a product whose omissions you agree with. That is a much better fit test than a feature list, because the features are largely the same and the omissions are where the philosophy lives.
Three questions not worth asking
How many integrations do you have? The number is meaningless. Ask instead whether it connects to the two or three tools you actually use, and whether it offers webhooks so you can connect anything else yourself.
Is it AI-powered? Almost everything is now. Ask what the AI specifically does and whether it costs extra.
Can it be customised? Everything can be customised. The better question is how much configuration is required before it is usable, and who on your team will own that work.
How to run the evaluation
Shortlist two products, not five. Take a real trial with real data rather than a demo, because a demo shows you the product working perfectly with data that was prepared for it.
Give it to the person who will use it most, not the person who is choosing it. Their reaction after a week is the most reliable signal you will get, and it is the one most evaluations never collect.
Then price both at the plan you would be on in six months, not the one you would start on.
If you want to run these questions past us, the short answers are: $29 per user with everything included, no seat minimum, no add-ons, values calculated from your catalogue, renewals created automatically, CSV export on everything, permissions enforced server-side, and we do not do marketing automation, service desks or project management. See what else we refuse to do.