The quiet exodus: how Zammad ended up running support for some of the world's biggest brands.

The interesting thing about enterprise software is that the biggest decisions almost never get announced. A Fortune 500 logistics company doesn't issue a press release when it migrates 800 agents off Zendesk. A global NGO doesn't publish a blog post explaining why it dropped a six-figure SaaS contract. It just happens, on a Tuesday, and the only people who notice are the support team and the finance team. For very different reasons.

If you go looking for it though, you can find the pattern. The system most of these quiet migrations land on is called Zammad. It is open-source, web-based, and most people working in customer support have never heard of it. The companies that have heard of it tend to be running it.

What Zammad actually is.

Zammad is a helpdesk and customer-support platform. Tickets, queues, SLAs, knowledge base, the usual shape. It is released under the GNU AGPLv3 license, and its source code is owned by the Zammad Foundation, a non-profit independent of any commercial vendor. The day-to-day development is done by Zammad GmbH in Germany, in collaboration with the community.^[1]

That last bit matters more than it sounds. Most "open core" enterprise software is owned by a company that can change its mind. Zammad cannot be quietly relicensed, sold, or rug-pulled, because the entity that holds the code is structurally not allowed to do those things. For a large enterprise picking a system it intends to run for ten years, that is an unusual amount of certainty.

01 · The customers

Who is actually running it.

Zammad publishes its customer reference list openly, and the names are not what most people expect from an open-source ticket system.^[2] A partial list, all confirmed on Zammad's own customer pages:

These are not toy deployments. Schenker is one of the largest freight-forwarders in the world. Amnesty operates in dozens of countries. swissQprint handles support in German, English, Spanish and Japanese from a single Zammad inbox. They switched to it in 2017 and described their previous shared-mailbox setup as "unstructured and time-consuming."^[3]

02 · The math

The cost story is real, but it isn't the whole story.

The easiest answer to "why are large companies leaving SaaS helpdesks" is the obvious one: per-seat pricing eats you alive at scale. A published case study from the cloud integrator Kubedo walks through a European SaaS company with a 30-agent support team handling roughly 1,800 tickets per day. They migrated from Zendesk to a Zammad stack and reported a 90% reduction in annual support-tooling cost.^[4]

You should treat any single case study with appropriate skepticism. The pricing gap is real though. Zendesk's enterprise tier sits in the $115–$215 per-agent-per-month range as of mid-2026, and a 30-agent team paying $150 each costs $54,000 a year before any add-ons, integrations or sandboxes. The same team on Zammad pays for a server, a database, and one annual update, typically under $3,000 even on premium cloud.

What gets the decision made is the second-order stuff. The CFO is happy. But the CIO is happy because the data lives in their tenant. The DPO is happy because there is no cross-border processor to argue about. The CISO is happy because the audit trail is in a Postgres table they can query. The procurement team is happy because there is no contract to renew under duress.

03 · The checklist

The enterprise checklist, all in the box.

For a long time, the answer to "can you run open-source software at enterprise scale" was a polite no. The features that matter to a Fortune 1000 security review (SSO, audit logging, role-based access, encrypted email) were either missing or required a paid plugin from a third party. Zammad simply ships them:

• LDAP / Active Directory for centralized user and group sync
• SSO via SAML, OpenID Connect or Shibboleth
• Two-factor authentication for agents
• S/MIME and PGP for signed and encrypted email
• Role-based access with custom roles per business unit
• Audit-proof ticket history: every change, who and when
• REST API and webhooks for integrations on either side
• Multi-channel: email, chat, phone, Microsoft Teams, WhatsApp Business, Telegram, Facebook
• SLA matrix with escalation triggers
• Multilingual knowledge base for agents and customers
• Grafana / Kibana integration for custom reporting
• Self-hostable on any Linux box: Docker, RPM, or Debian package

All of these are documented in the official feature catalog^[5] and verifiable in the open-source repo. There is no enterprise edition gating SAML behind a paywall. The community edition is the enterprise edition.

Zammad's own marketing claim about being "audit-proof. That's one reason why it is often used in banks"^[6] is not as bold as it sounds when you look at what it means in practice: every ticket holds an immutable change history of who edited what and when, queryable directly from the database. Many SaaS helpdesks sell that as a top-tier add-on.

04 · The trade-off

What you give up.

The honest part. Zammad is not free in the sense that matters. The license is. The labor isn't. You take on the operational burden of running a stack that includes PostgreSQL, Elasticsearch, a Ruby on Rails app, Memcached, a reverse proxy and a TLS cert.^[7] None of those are exotic. All of them have to be patched, monitored and backed up by somebody.

For a 500-agent enterprise with a platform team, that's a Tuesday. For a 30-person team without a dedicated infrastructure engineer, it's a real cost. The trade-off Zammad asks you to make is the same trade-off every self-hosted system asks: do you want to rent a service, or own a system? SaaS makes one of those answers very easy. The other one used to be very hard, and Zammad is one of the more honest attempts to make it less so.

The companies migrating to it have mostly already done that math. They have either grown past the point where per-seat pricing makes sense, or they have a compliance posture that makes "your data leaves the building" a non-starter. For everyone else, the SaaS answer is still rational. It just becomes less rational every year.

05 · The takeaway

Why it matters.

The reason this is worth paying attention to isn't Zammad specifically. It's that the category of "serious open-source business software that a global enterprise can actually run" is no longer aspirational. Zammad is a working example. Odoo is another. There are more. The supply side has caught up.

What hasn't caught up is the demand side. Most procurement teams are still defaulting to the SaaS option because the SaaS option used to be the only one that worked at scale, and habits outlast their reasons. The companies that have noticed, quietly, on a Tuesday, are the ones already running Zammad.