Summary - In this article (the second in a two-part series), we dive into the opportunities that metric-centric analytics tools provide to data teams, decision makers, and analysts. We also explore the future of analytics, where BI systems work equally well for machine and human consumption and where everyone understands and leverages the power of data.
Recap from Part 1: The evolution of Business Intelligence (BI) software design has been fuelled by the need to understand and effectively use data to make good decisions. A third generation of BI tools is emerging. It’s centered around metrics – BI artifacts that capture purpose and data and address the immediate needs of decision makers – in a way that reports don’t. Metric-centric analytics solutions, like Klipfolio PowerMetrics, treat metrics as first-class artifacts and are designed for independent analysis by business users.
Everyone benefits from metric-centric BI
The first two generations of BI tools manipulated and presented information but didn’t focus on the real purpose of data – enabling and supporting good business decisions. Metric-centric BI builds purpose into the data so even less technical users can make better decisions. Let’s look at how this new generation of BI solves business challenges for data teams, decision makers, and analysts.
Why metric-centric BI is good for data teams
Business users faced many usability challenges when working with 2nd generation BI. As a result, data teams spent a lot of time authoring reports, often based on ad hoc requests. With metric-centric BI, instead of authoring reports, data teams choose the right data for each metric and have more time for other meaningful work.
Each metric has a specific definition. It’s easier (and faster) for data teams to fulfill a well-defined need versus managing a multi-purpose set of data, as is the case with traditional BI models. Each metric can be fulfilled separately and from different data sources. This reduces or eliminates the need for advanced modeling to combine data. It’s also easier for decision makers to see the information they need so they can ask for specific metrics from their data teams, streamlining the request process. In addition, the ability to allow and restrict access on a per-metric or dashboard basis means data teams can be confident the right metrics are being exposed to the right users.
Why metric-centric BI is good for decision makers
Metric-centric BI consumption tools are easy to use due to the highly structured nature of metrics. There’s no need to expose complex features for selecting data from a data model. Instead, these tools provide user-friendly features for visualizing and combining metrics. This gives decision makers more options and the flexibility to self-serve than static reports do. Instead of hunting in models and reports, decision makers can also easily access data using internal metrics libraries and catalogs.
Why metric-centric BI is good for analysts
Traditionally, data analysts spent a lot of time collecting, cleaning, and assigning meaning to data. Only after that was done could they start to process the data and uncover insights. Since metrics represent specific business concepts to which clean and pre-modeled data has already been applied, these time-consuming steps are unnecessary.
The consistent structure of a metric makes it easier to query. This also means it’s more immediately consumable in analysis toolkits. For example, most business analysis relies on time (for example, comparisons between time periods). The structure of metrics means they typically include temporal context. This makes the analyst’s job easier – time is modelled into the metric even when it’s not available in the raw data.
Driving data adoption by machines and people
With the widespread availability of data and ongoing advancements in AI, the ability to make effective data-driven decisions is more important than ever. A metric-centric approach helps machines and humans interpret and work with data, a necessity when designing future-proof BI.
How metrics enable AI-driven BI
The main challenge in applying AI to data to gain insights is the data itself. Data is often messy and lacks structure, making it unclear to humans, let alone an AI system. AI can only identify patterns and insights in data it understands.
Metrics include well-structured data. Their definitions provide context for what the data represents. Query results returned from metrics are equally well-defined and structured. These qualities make metrics immediately consumable by most AI systems. The AI can understand and draw connections between the metrics being analyzed and expose the data and context as meaning to end users.
Why metrics enable the socialization of BI
Business intelligence infrastructures have traditionally been isolated. Each organization had its own BI infrastructure built up around their specific needs. These unique setups made it difficult to establish standard best practices for BI and to define common BI artifacts. Sure, KPIs could be identified and discussed in abstract terms by industry experts, but each business would implement them separately.
Since a metric-centric BI artifact captures a KPI definition even if it’s not connected to data, it can be published and consumed across organizations. Public libraries of metrics can be made available that define best practices for common business KPIs. Reports and dashboards that use these abstract artifacts can also be defined, providing a toolset even before the data is connected. These artifacts can then be visualized either using known queries against common APIs or, with minimal guidance from data teams, by connecting organizational data to the metrics. MetricHQ is an example of a publicly available metric knowledge base.
By enabling the creation of consumption BI artifacts that are independent of the data for metrics, dashboards, and reports, metric-centric BI has established a sharing culture not possible with previous generations of BI tools. Businesses no longer require a data team to create and use common metrics. Instead, they can access metric definitions and data connection templates created by experts in public libraries.
Why metrics enable effective decision making
The decision-making process starts with identifying an opportunity or a problem, investigating alternatives, choosing a solution, acting on it, and, finally, monitoring to ensure a satisfactory outcome. Having the right data, and understanding it, is key to the first step (identifying the opportunity/problem) and the last step (monitoring).
Metrics have a defined purpose. The clarity in metric data makes it easy to identify opportunities or problems. Each metric has meaning built into it, so the patterns that surface in the metric data aren’t only interesting, they indicate actionable, meaningful activity in your business.
Metrics are ideal for monitoring the situation after a solution has been implemented. A metric can be set up to monitor outcomes for a specific decision. For example, if the decision was to increase advertising spend, a metric could be defined that monitors website traffic and measures the decision’s success from that perspective. Metric-centric tools can include additional monitoring features, such as the ability to define metric goals and be notified when the metric value changes relative to the goal.