How to choose a CRM — an evaluation that survives reality

Start from the failure, not the feature list

Every CRM evaluation that goes wrong starts the same way: with a feature comparison. Feature tables are the least useful artefact in the process, because every serious product has every table-stakes feature, and the differences that matter are in behaviour rather than presence.

Start instead by naming what is actually broken. Write one sentence: the reason we need this is that something specific is failing. Common honest answers:

• Deals are dying from missed follow-up.
• Nobody can see the pipeline without asking a person.
• History lives in individual inboxes and leaves when people do.
• We are not having enough new conversations.
• Reporting takes a day a month and nobody trusts the result.

Those five point at different products. The fourth in particular — not enough new conversations — is not a CRM problem at all in the traditional sense; it is an outreach problem, and buying a records system to solve it is the most expensive way to not fix it.

The order to evaluate in

Evaluate in this sequence, and stop as soon as a product fails a step. Most shortlists collapse from eight to three by step three.

1. Does it model your business?

Not can it be customised to. Out of the box, does it have the objects you sell against? If you sell to organisations, does it treat the organisation as a first-class record or as a text field on a contact? If you sell recurring revenue, can a deal represent that, or only a one-off value? If you sell through partners, is there anywhere to put the partner?

Customisation can close most gaps. It also creates the configuration you will maintain forever, and configuration is where CRM projects go to die.

2. Does history capture itself?

Ask specifically: when a rep sends an email from their normal mail client, does it appear on the contact record without any action? What about a call from their mobile? A LinkedIn message? A WhatsApp?

This is the highest-leverage question in the entire evaluation and it is almost never on the feature comparison, because every vendor can answer yes to does it log email while meaning wildly different things. A CRM where logging is manual will contain a partial history within a month and a misleading one within three.

3. What does it cost for everyone who needs to see it?

Count the people who should be able to look at the pipeline, not the people who will edit it. Founders, finance, marketing, delivery, customer success. Then price the tool for that number.

Per-seat pricing is the industry norm and it quietly converts visibility into a budget decision. The predictable result is that the people who most need context — the ones joining a call cold, or forecasting cash — do not have access. Usage-based models price the other way: Autocloz includes unlimited users on every plan including the free one, and charges on per-channel sending volume instead, because that is where the real marginal cost sits. Neither model is universally right; what matters is that you price the real user count rather than the licensed one.

4. Can you get your data back out?

Before you put anything in, confirm what comes out: contacts, companies, deals, custom fields, notes, activity history, attachments. In what format, on what timescale, and can you do it yourself or does it need a support ticket?

A vendor that makes leaving hard is telling you something about how they expect to keep you.

Why demos mislead, and what to do instead

A demo shows you a system configured by someone who knows it perfectly, populated with data designed to look good, driven through a path they have rehearsed. Nothing about that predicts your Tuesday.

Three things to insist on instead:

Use your own data. Ask for a trial and import a real slice — two hundred contacts, ten open deals, with your actual field mess included. The import experience is itself diagnostic: if mapping your fields is painful in the trial, migration will be worse.

Have a rep drive, not the vendor. Give someone on your team a specific task — log a call, move a deal, find every account nobody has touched in thirty days — and watch without helping. Count the clicks. The click count is the adoption forecast.

Run one real week. Not a parallel test, a real one: one rep, one segment, actually working in it. A week reveals friction that a month of evaluation calls will not.

The costs that appear after you sign

Budget for these explicitly, because they are real and rarely quoted:

• Migration. Contacts and deals move in an afternoon. Custom fields, automations, email history and integrations do not. Assume weeks of part-time work for anything beyond a simple system.
• Integrations. Check whether the connectors you need are native, third-party, or available via our API. The third answer means you are building it.
• Add-on modules. Marketing, service, quoting, dialing and AI credits are frequently separate lines. Price the bundle you will actually be on in a year.
• The seat creep. Teams grow, and per-seat bills grow ahead of value because the newest seats are usually the lightest users.
• Admin time. Somebody will own this. If nobody is named, the answer is that data quality will own it, badly.

Four questions that eliminate most shortlists

If you only ask four, ask these:

1. What happens to a record when the rep who owns it leaves? The answer reveals how much of the system's value is actually stored versus remembered.
2. Show me a deal that has gone quiet. If finding it requires building a report, nobody will find it in practice.
3. How does a reply stop a sequence? In any tool with automation, this is the difference between professional and embarrassing. Ask what happens with an out-of-office too.
4. What is the smallest plan we could run this on, honestly? A vendor willing to talk you down a tier is one worth trusting on the rest.

When the answer is not a general-purpose CRM

Two cases worth naming, because the standard evaluation misses both.

If your bottleneck is new conversations, the deciding factor is not pipeline features but whether the tool can actually reach people — sequences that pace themselves, mailbox warmup, authentication monitoring, a dialer, and a compliance gate for calls and texts. A conventional CRM will store the outcome of outreach it cannot perform. See the best free CRM comparison and email automation software for the tools built the other way round.

If your business is delivery-heavy, the handover from won deal to delivered work matters more than the pipeline itself, and a CRM with project management attached will beat a better pure CRM. Insightly is the obvious example; our Autocloz versus Insightly comparison lays out that trade honestly.

The decision

When two products are genuinely close, pick the one your team will keep current. Data quality compounds and features do not: a simpler CRM with an accurate pipeline beats a sophisticated one describing last month. Everything in this evaluation is ultimately a proxy for that one question.

Ready to move? The CRM migration checklist covers the switch itself without losing history.