Our metric-centric product vision – Business Intelligence ahead of its time
Allan WillePublished 2023-06-21Summary: In 2018, when the world was still obsessed with monolithic dashboards and sprawling data lakes, we began building something different. We launched PowerMetrics as the first metric-centric analytics solution—a concept that went beyond the traditional "semantic layer" to anticipate the needs of a more agile, cloud-native future. Today, that vision has become the industry standard. This is the story of how we moved beyond the data warehouse-centric BI stack to embrace a future where data finally has meaning.
Moving Beyond the Warehouse-Centric Stack
There is a fundamental shift occurring in business intelligence. Businesses are moving away from the traditional model—where data is dumped into a warehouse and "visualized" without context—and embracing a metric-centric vision.
Focused on defined metrics, rigorous governance, and clear lineage, this approach provides the trust and confidence needed to make data-driven decisions at speed. While industry thought leaders have only recently begun to champion the "metrics layer," Klipfolio has been refining this architecture for nearly a decade.
What is the ‘semantic layer’ trend all about?
In the traditional approach to BI, data lacks inherent meaning. Even when measures and dimensions are named in a report, that naming is often loose, inconsistent, and stuck in a silo.
In a modern data stack, the semantic meaning (the metrics layer) is inserted between the raw data and the people using it. Metrics are defined at a global level, abstracted from the source, and shared across the entire organization. This ensures a stable, single source of truth.
Properly defined metrics—stored in a central layer—present the first real opportunity for true self-serve data analysis. We call this metric-centric BI. While we pioneered this path, others have since joined the movement: dbt Labs now promotes the dbt Semantic Layer, and Google’s Looker launched Looker Modeler to solve these same consistency challenges.
Why We Rethought BI Five Years Ago
To understand our vision, you have to understand our history. In 2012, we launched Klips. It was a revolutionary, flexible product that connected directly to cloud services, performing transformations in real time at the visualization layer.
Klips was (and still is) powerful, but it had a steep learning curve. As our customers matured, they told us that managing "raw" data was becoming an overwhelming challenge.
Data needs to have meaning
With our first product, customers could access a lot of data, but it lacked context. Things got messy, fast. It was like having a thousand sticky notes of information without a filing system. As data grew, customers struggled to manage thousands of individual assets.
We realized the future was clear: Data must have meaning. This led us to the concept of the metric. All metrics should have business definitions, a catalog of shared dimensions, and consistent time grains. We shifted our focus from "dashboards" to "opinionated metrics"—the solution to a world where data was noisy and directionless.
Analytics must be self-serve
In the world of traditional BI, business teams (Marketing, Sales, Finance) have had to wait on technical teams to "prepare" data. This creates an endless feedback loop of requests and revisions that frustrates everyone.
Our vision was a "buffet" of metrics: a world where a business leader could walk up to a prepared set of centrally defined KPIs and serve themselves. However, we noticed that even the best self-serve tools required a "data person" to ensure accuracy. This sparked our CTO’s realization in 2018:
- Users often don't know how a metric is calculated or what action to take.
- Without semantic meaning, there are no "network effects"—you can’t easily compare your performance against benchmarks or see relationships between different parts of the business.
PowerMetrics: A metric-centric architecture
We decided to build PowerMetrics from the ground up, making the Metric a first-class citizen. Our architecture is defined by two core ideas:
- Semantic Meaning: Data becomes meaningful only when business logic is applied. "Net Profit" isn't just a number; it is an understood business term calculated as Sum(Revenue) - Sum(Expenses).
- Time as the Universal Dimension: Almost every business question involves time. We built PowerMetrics to handle snapshot, periodic, and transactional data natively, ensuring that every metric is comparable across any time grain.
In 2020, we doubled down on this by launching MetricHQ, the world’s first community-driven dictionary for metrics. It allows data and business teams to learn from industry experts and add pre-built metric definitions directly to their accounts.
Why the world wasn’t ready for PowerMetrics. Until now.
Early on, our vision was so ahead of its time that even we struggled to message it correctly. We focused entirely on the end user, who often saw the "metrics layer" as an extra step in the way of their dashboard.
Then came our eureka moment: Trusted data is the common goal. Data teams want governance; business teams want accuracy. When they work together through a metrics layer, the friction disappears. Experience has taught us that working with data engineers and analysts leads to the best outcomes for the executive suite.
Metrics: The language of data AND business teams
Today, we see the purpose of the data team and the business team finally aligning. In this new world:
- Data teams manage the complex query landscapes and ensure integrity via the metrics layer.
- Business teams apply those trusted metrics to their own dashboards and reports in a true self-serve environment.
The metrics are defined once and used everywhere. No more queuing up for IT to build "one more report." No more arguments over why two different spreadsheets show two different versions of the truth.
The future is data metric driven
We are moving beyond the limitations of the old warehouse-centric stack. We are joined by a growing group of believers—leaders from dbt, Mode, and DuckDB—who agree that the future isn't about "Big Data," but about Valuable Metric Data.
The metric-centric approach offers a more efficient workflow, better performance, and superior governance. But most importantly, it empowers people at every level of the organization to make well-informed decisions with absolute confidence.
Data needs meaning. The future of business intelligence is here, and it’s all about metrics.