SensorFlow Migrated from Looker to Holistics in Two Weeks and Achieved 88% MAU

Max Pagel, CTO at SensorFlow, had used Looker for three years before he made the switch to Holistics.

SensorFlow, a Singapore-based company providing smart energy management solutions for hotels, was growing fast at the time, powering 40,000 IoT devices across 18,000 hotel rooms in Asia. As the company grew, Looker’s limitations became more pronounced: stagnant product development and a pricing model that penalized scaling, especially for Embedded Analytics, a critical component as SensorFlow expanded client-facing data access.

The transition from Looker to Holistics was smooth and straightforward. Holistics offered a similar code-based semantic layer, a self-service UI, core features that Max valued, without the weight of restrictive enterprise pricing or the product inertia that had begun to characterize Looker.

Within two weeks after the switch, SensorFlow migrated 45 dashboards and 124 data models to Holistics, with just a two-person data team.

SensorFlow operates in a high-volume IoT environment, ingesting 2,000 rows of sensor data per second from over 18,000 hotel rooms across Southeast Asia. Internally, this data powers their operational dashboards for tech support, customer success, and engineering teams. Externally, it underpins embedded dashboards used by hotel engineers to monitor HVAC load, room automation, and energy configurations in real time. SensorFlow’s previous setup with Looker was straining under these requirements.

Embedding dashboards came with a steep price: Looker’s per-embed-user licensing model forced the team to purchase user licenses for each customer, even if they only needed access to a single dashboard.

Even more frustrating was the architecture: Looker required two separate instances for internal and embedded use cases, each with duplicated LookML models.

Compounding the frustration was Looker's stale development pace after being acquired by Google. For a forward-looking company like SensorFlow, that lack of momentum raised long-term concerns.

Frustration built up, and Max knew it was time for a change.

When Max began evaluating alternatives to Looker, he was searching for something fundamentally similar to Looker, but better designed for modern operational demands. The team explored all the most popular tools (Domo, Tableau, Power BI, and Grafana). Some were too static, others lacked support for real-time querying, and many fell apart when it came to embedded analytics.

Holistics emerged as the standout. It shared Looker’s core strengths—a semantic modeling layer, live query support and self-service capabilities—but crucially, it fixed two of Looker’s most protruding shortcomings: dynamic data source switching and embedded pricing.

Max’s team runs multiple databases with shared schemas. Without the right tooling, that would have meant building the same models and dashboards multiple times, just to support different databases. "A critical capability for us was being able to work seamlessly across our different databases," explains Max.

This kind of dynamic connection switching based on user attributes or environment was not natively supported in Looker. While Looker offers user attributes and Passthrough Authentication as partial solutions, they fall short in key areas:

• They can’t change database dialects (e.g., SQL Server to Snowflake).
• They don’t allow model-level switching between connections.
• Workarounds often require duplication or complex logic.

Holistics, on the other hand, offered native support for this use case through its Dynamic Data Sources feature, which allows the same models and dashboards to point to different data sources based on runtime conditions. This meant SensorFlow could maintain a single set of models while dynamically switching between production and analytics databases as needed.

"We always wanted to set up embedded dashboards for our customers," Max shares. "That was always the goal, to give our customers direct access to their data—see the dashboards, track performance, take actions from it."

Yet, Looker's pricing model made that goal unreachable. The cost of giving dashboard access to every user, many of whom only needed it occasionally, made the ROI impossible to justify. When every dashboard user requires a separate license, the costs add up quickly, even if they only log in once a month. This pricing structure created a significant barrier for SensorFlow's vision of providing analytics to all its hotel clients.

"That just doesn't make sense for our business model," says Max. "We needed a pricing approach that would allow us to experiment freely and scale our embedded analytics without constantly worrying about the financial implications."

The pricing model of Holistics allows Max's team to enjoy unlimited embedded viewers and view counts. Instead of charging for every user or query run, Holistics charges based on the number of concurrent dashboard queries (via embedded workers). This pricing model gives SensorFlow cost predictability, with no surprise overages, and lets them roll out embedded dashboards widely, without racking up costs for users who only check in occasionally.

This approach gave SensorFlow the freedom to experiment and scale its embedded analytics offering without financial constraints, finally enabling Max’s vision of making SensorFlow a true action delivery platform.

With Holistics meeting all the key requirements, Max was ready to begin the migration process. Because both platforms share the same foundational approach of running live queries against a database through a semantic layer, SensorFlow's team could leverage their existing knowledge without rethinking their entire analytics strategy.

The transition from Looker to Holistics felt very intuitive from day one.

With a small team of five people, SensorFlow successfully migrated 45 dashboards and 124 data models from Looker to Holistics in just two weeks, with another two weeks for final adjustments.

The migration process also became an opportunity to optimize their analytics portfolio. "We probably retired at least 50-60 percent of our dashboards that we had on Looker because they weren't relevant anymore," Max notes.

One of the biggest improvements Max noticed with Holistics was how much easier it was to create metrics on the fly, without having to go back and update the semantic model every time.

In Looker, every ad-hoc calculation required editing the LookML model, saving, validating, and pushing changes before a new aggregation or calculation could appear on a dashboard. That meant analysts constantly switched between the modeling layer and the visualization layer, even for basic analytical questions.

With Holistics, analysts could work directly within the dashboard layer to apply new aggregations or calculations without detouring into the model. This reduced context switching and made the development workflow significantly more fluid.

This flexibility was especially obvious during their migration when Max asked his tech lead to start porting dashboards into Holistics. “Within minutes, he was like, ‘Oh, this is great. So much easier than Looker.”

After SensorFlow moved to Holistics, their data team finally had a way to unify both internal and embedded analytics without juggling multiple tools or chasing down developers.

Max, who leads the initiative, has long envisioned SensorFlow as more than just a smart building platform. He’s working toward making it an Action Delivery platform, one where the right people get the right data, fast enough to actually do something about it.

Holistics made that speed real. Now, the analytics team, and even customer support, can build and share dashboards on their own. This shift has helped the team respond to customer needs more quickly, and made the whole experience smoother for both internal teams and SensorFlow’s users.

What once took week-long dev sprints now takes minutes, and the ROI of the platform speaks for itself.

SensorFlow now has a new operating rhythm: fast, flexible, and customer-centric. The platform now powers real-time internal insights, proactive customer interventions, and experimentation at scale, all while keeping engineering overhead low.