Transform: Making Metrics the Interface for Data

Business stakeholders want accurate metrics such as active customers, revenue, and acquisition to make informed decisions. That is why companies must ensure a consistent definition of all these metrics. Data analysts take raw data sets and ultimately turn them into useful metrics or time series of data or single data points that become the foundations for important decisions. But these metrics are prone to errors due to the involvement of the human element, and the process is also time-consuming. Secondly, data analysts often struggle with improving productivity. A data analyst becomes a gatekeeper for the data inside of a modern business. Today’s systems speak the language of data, which is SQL, and programming languages such as Python, which only data analysts understand. That is why business stakeholders need to go through data analysts to obtain the data sets they want.

We are trying to define a layer on top of logic that allows business stakeholders to express their data sets in metrics without the help of data analysts. They can ask to obtain answers for any question such as revenue by product category, week, or country. Even for the straightforward questions, the logic that gets expressed against the underlying data warehouse is complicated. We are building a layer that can translate business stakeholders’ simple requests into the language that is compatible with the underlying storage and compute engines for turning into data sets.

Could you elaborate on the features and functionalities of your platform?

We are bridging the gap between the data analyst and the business stakeholders or the data producer and the data consumer. Data analysts define metrics using a combination of structured abstractions to express complicated logic in a relatively simple format. Then, they manage and govern the definitions of the metrics using version control and various modern software engineering systems. These metrics get picked up by Transform that exposes a variety of interfaces for different consumers to use. We have a metrics catalog, which allows business stakeholders to come and look at all of the different metrics that are defined within Transform. They can see the definition, a chart of the data, metadata, ownership, tiering of the metric, and everything around governance. Data analysts can also create annotations and analyze that a particular event happened at this specific time, which is why there is a fluctuation in the metric.

Stakeholders can ask questions, and data analysts can answer those. Our aim is to build structured information around metrics, which has historically not existed in a formalized context in the data space. From data analysts, engineers to anyone trying to build different applications on top of metrics can utilize Transform. We have a Python interface, command-line interface, and all the generic interfaces that software engineers and data analysts consume. We enable data analysts to pull metrics into everything, including BI tools, forecasting platforms, anomaly detection for experimentation, executive reports, internal tools, and any applications where they want to consume metrics.

What does the future look like for Transform?

In June this year, we expanded our company into several markets but didn’t open access to the product yet. At present, we are still completing the foundational components of the product. We have worked with several partners and shaped our product based on their feedback to deliver more value to clients. Recently we have made a few new partnerships to broaden and generalize our product. These partnerships are proven to be valuable for enhancing the product’s capabilities and features. We have reached a point where many data teams are interested in the metric store. We are scaling our team, engineering capabilities, and product development to support our customers with a product that will immediately bring value to their organization.