With a CRM, the feature list decides nothing. The export does.

Customer data is the estate where switching costs the most — and the one where hardly anyone checks in advance whether switching is possible at all.

The reason is a misunderstanding about what a CRM actually holds. Contacts are the easy part; every system exports those. The value sits in the rest: the note from a phone call three years ago, the attachment on a quote, the link between company, contact and deal, the history that explains why an order was lost.

That is precisely the part missing from most exports.

The test that belongs before the choice

It takes half an hour and works with any vendor offering a trial.

Create five records that match your daily reality: a company with two contacts, a deal with a quote attached, three dated notes, one closed and one open task, an email filed against the deal.

Then export — the way you would have to in an emergency, without asking the vendor.

Now open the export and check four things. Are the attachments included, or only links to addresses that stop resolving when the contract ends? Do the relations between records survive? Are the timestamps there? And is the format readable without the vendor's software?

Anyone who can answer those four questions has learned more about the vendor than any comparison site will tell them.

What is on the table

Odoo (Belgium) is the strongest all-rounder for companies between ten and fifty people: CRM, quotes, invoices and inventory in one, with an open-source core and the option to self-host. The catch rarely appears in the proposal: you are not introducing a CRM, you are introducing an ERP. That is a different size of project and needs a different decision.

Twenty (France) is young, open source and visibly built for people leaving Salesforce. Workable for small sales teams, still thin on automation, reporting and integrations.

Dolibarr (France) has been developed since 2003 and carried by an association since 2009. The interface visibly comes from another era, and much depends on add-on modules of uneven quality. In exchange there is no investor here who will eventually need an exit — in this context that is a feature, not an accident.

CentralStationCRM (Cologne) stays deliberately small: contacts, notes, tasks, no rollout project. For teams under twenty it is often exactly right.

weclapp (Marburg) is a subscription product with inventory management, strong in trade and manufacturing.

Two names the lists get wrong

Pipedrive is still filed as a European alternative. It was founded in Tartu in 2010, has been majority-owned by Vista Equity Partners of Texas since November 2020, and is headquartered in New York.

EspoCRM appears on many of the same lists. Development happens largely in Ukraine; the parent company is registered in Delaware.

Neither is a criticism of the products. It is a sign that origin data in such lists is rarely maintained — and that headquarters is secondary anyway when the export path is sound. A company that self-hosts is far less exposed to a change of ownership than one holding a subscription and no way to extract its data.

What to put in the contract

Three sentences that cost little before signing and are almost unobtainable afterwards:

The export covers master data, notes, attachments, relations and timestamps in a documented, machine-readable format.

It can be triggered at any time by you, without vendor involvement and without a separate charge.

After the contract ends, data is handed over within a fixed period and then deleted — both with dates, not with "on request".

Since September 2025 the EU Data Act backs these points for data processing services: switching charges are limited, notice periods capped, and the provider must assist with a switch. The legal position is better than contract practice. You only have to use it.

The real point

A CRM is the software with the strongest emotional attachment in the building. Sales chose it, management sees its numbers in it, and nobody replaces it for fun.

That is exactly why it is where Vendor lock-in gets most expensive — and where preparation pays off most. An export nobody has ever opened is not an export, it is a hope. Open it once before you need it.