The BI stack shift inside European mid-market: where Looker is losing ground
The licence renewal that started the conversation
A pattern shows up in the conversations European mid-market CTOs were having in late 2024 with their finance teams. Looker, acquired by Google in 2019 and folded into Google Cloud, was up for renewal. The licence quote came back materially higher than the previous cycle, with the now-standard Google Cloud bundling pressure attached. Most of the European customers running Looker had landed there during the 2019 to 2022 window when the modern BI category was effectively a choice between Looker and Tableau, and the architectural argument for Looker's modelling layer felt strong. The renewal quote in 2025 surfaced a question those customers had not asked recently. What does the modern BI category actually look like now, and is the replacement cost lower than the renewal delta.
The companies that ran that analysis seriously discovered three things. First, the alternatives had matured. Second, the cost gap was real, sometimes two or three times the licence price. Third, the data residency conversation had quietly become a procurement requirement that Looker's Google Cloud entanglement made awkward to answer. The combination produced a category shift that is now visible across the European mid-market.
What the alternatives actually look like
Lightdash, the British open-source BI company that raised a Series A in 2023, made the cleanest pitch for the displaced Looker user. The modelling layer is dbt-native, which is where Looker's customers had increasingly been doing their data work anyway. The licence model is open source for the core product with a hosted offering for teams that do not want to self-host. The product feels like a Looker successor rather than a different category, and the migration path from LookML to dbt models is the kind of one-time pain that finance teams will tolerate when the renewal arithmetic is two or three to one.
Metabase, the older San Francisco-built tool that built its European foothold through self-hosting and a permissive open-source licence, sits in a different architectural space. The strength is in the speed of going from connection to first dashboard. The weakness historically was the modelling layer, where Looker's semantic consistency had no real analogue. The 2024 release of Metabase Embedded SDK and the improvements to its query language closed enough of that gap that European mid-market accounts who needed self-hosted BI without an enterprise sales process landed there in increasing numbers.
Cube, the New York-headquartered open-source semantic layer, took a different approach. Instead of competing with Looker as a BI tool, it competes with Looker's LookML as a semantic layer that any BI tool can read from. The strategic implication is that a company can run its modelling logic in Cube and let Metabase, Lightdash, Hex, Mode, or even an internal app consume the same semantic definitions. For European mid-market customers worried about being locked into another BI vendor's modelling language, Cube's positioning solved a real procurement objection.
The European-specific shape of the shift
The category shift would still be happening without the European specifics, but it would be slower and would land differently. The European specifics that accelerated it are three. The 2022 to 2024 CNIL and BfDI rulings on data transfer to the US made any tool whose default vector pushed data through a US-incorporated provider harder to justify on the procurement side. The licence inflation across the broader US software market hit European mid-market accounts harder because they had less budget headroom and a smaller share of internal data infrastructure ready to absorb migration cost. The growth of dbt as the default transformation tool inside European data teams created a generation of analysts who were already writing models in code, which made the move to a code-first BI tool less disruptive than it would have been for an analyst who had only ever used a drag-and-drop tool.
The companies that ran the migration in 2025 are a small but informative set. A €200M ARR Dutch e-commerce platform moved from Looker to Lightdash over six months and reported a sixty percent reduction in BI licence cost. A German B2B SaaS company at €60M ARR moved from Looker to Metabase plus Cube and described the migration as the unglamorous infrastructure cleanup the data team had been deferring for three years. A French fintech moved off Looker not because of the cost but because the Google Cloud lock-in made its Schrems II posture defensively weaker than the regulator-facing diligence required.
What the buyer needs to ask
For a European mid-market data leader looking at the next renewal cycle, the productive question is not which tool to buy. It is which architectural property the team needs to preserve. If the semantic layer is the load-bearing component, Cube plus any of the consumption tools is the cleanest path. If the speed of analyst self-service is the load-bearing component, Lightdash or Metabase is closer to what the team actually uses every day. If the existing Looker setup has been hollowed out by dbt and the modelling has migrated out of LookML already, the renewal conversation is even easier to answer because the cost of moving is lower than it looks.
The category has stopped being a two-vendor decision. The European mid-market is moving accordingly, and the next twelve months will produce more case studies of that move than the previous five years combined.