| Overall score | 4.4Best value | 4.4Best value | 4.1Best value | 4.0Best value |
|---|
| Ecosystem | 4.7Best value De-facto standard of the transformation layer with a broad package landscape. Since 1 June 2026 dbt Labs belongs to Fivetran, so the transformation layer is no longer independent of the loading vendor. | 4.5Best value Over 600 connectors, including long-tail ad platforms; the Starter tier covers 121 of them. | 5.0Best value Over 750 sources and more than 200 destinations, including reverse ETL since the Census acquisition. Since the dbt Labs merger closed on 1 June 2026, the transformation layer belongs to the same vendor. | 5.0Best value A huge community connector landscape deep into the long tail. |
|---|
| EU data sovereignty | 4.7Best value dbt Core runs in your own infrastructure and transforms inside the warehouse. No payload data reaches the vendor. dbt Cloud is the US-hosted alternative. | 5.0Best value Swedish vendor with EU hosting and an EU contractual basis. | 3.5Best value US vendor with an EU region, but all data passes through Fivetran’s infrastructure. There is no self-hosting option. | 4.2Best value US vendor with an EU region. The core can be self-hosted (Elastic License 2.0, free for internal use), so there is an exit path out of the cloud. |
|---|
| Reliability | 4.5Best value Tests on every build catch silent data errors before they hit a dashboard. | 4.5Best value Managed pipelines that simply keep running. | 4.5Best value The managed heavyweight: pipelines run without babysitting. | 3.0Best value Community connectors vary in quality, plan for maintenance. |
|---|
| Value for money | 4.5Best value dbt Core is open source; costs only start with Cloud or orchestration. | 3.5Best value From mid-size volumes the licence becomes a real line item. | 3.0Best value MAR-based pricing gets expensive at scale. | 4.5Best value Self-hosted without licence costs, far cheaper than managed ELT. |
|---|
| Ease of use | 3.5Best value Assumes SQL and git discipline, otherwise the value stays on the table. | 4.5Best value No-code harmonisation, marketing teams manage without engineering. | 4.5Best value Pick a connector, choose a destination, done. | 3.5Best value Setup and upkeep need a team that runs pipelines. |
|---|
| Pricing model | Core free; Cloud per user | Subscription by data volume | Usage-based (monthly active rows) | Open source; usage-based cloud |
|---|
| Vendor | dbt Labs · USA | Funnel Oy · Sweden | Fivetran · USA | Airbyte · USA |
|---|
| Subcategory | Transformation | Ingestion & Connectors | Ingestion & Connectors | Ingestion & Connectors |
|---|
| Operating model | Self-hosted | Managed · EU | EU-Region | EU-Region |
|---|
| Last reviewed | May 2026 | May 2026 | May 2026 | Jun 2026 |
|---|
| When it fits | When SQL transformations should be versioned, tested and documented. | When marketing data from many sources should flow harmonised into the warehouse. | When you want a broad, managed connector portfolio and end-to-end data movement from one vendor. | When long-tail APIs need connecting and engineering capacity exists to run it. |
|---|
| When it does not | When there is only one report and nobody maintains SQL. | When only two or three standard sources count; then native exports cost less. | When only marketing data counts. Then a focused tool like funnel.io is leaner and cheaper. | When nobody wants to babysit pipelines; then Fivetran or funnel.io. |
|---|
