A metric is a new type of BI artifact that captures the data and meaning around a single business concept. Unlike traditional BI models and reports, metrics don’t expose any end user configurations that could change the meaning of the metric. Once metrics are defined, you know they can be trusted. And, the ability to retrieve data using contextual queries makes for flexible, comprehensive analytics. With metrics, you get the best of business intelligence analysis – trustworthy, all-encompassing data for effective decision making.
Why complex BI tools may not be as powerful as they seem
Over the last few decades, BI products have focused more and more on flexibility and rich feature sets. More capability was assumed to mean more power and more value. Yet, studies show most companies that could take advantage of these advanced BI tools continue to use spreadsheets instead.
As we all know, BI software has evolved tremendously, with capabilities that could make any business user seem like a data superhero. So, why aren't more business teams adopting these feature-rich BI tools into their daily decision making? It all boils down to trust, or, more specifically, the lack of it.
Business users often lack confidence in the data – Is it accurate? Is it the most recent? Is it the right version? Sometimes they’re not sure if they can trust the sources of the data or if the data team fully understood their request, saying the data doesn’t look like what they asked for.
On top of this, most BI tools include features that require a deep understanding of the data to really see what’s going on. This stumbling block makes it difficult for the casual user to trust the tool or that they’re using it appropriately to solve their immediate issue. An unfair expectation has been placed on the average business user to become a data expert in order to effectively use a sophisticated, modern BI tool.
Instead of an over-powered analytics tool that includes features they don’t need or comprehend, business users are looking for a streamlined solution that empowers them to independently make daily decisions, fuelled by trusted data in a format they can analyze and understand.
Where does the data team fit in?
As a company grows and tries to solve larger, more complex data challenges, they turn to data experts to collect, manage and prepare their vast amounts of data in a way business users can consume. Eventually, they mature to a point where a dedicated data team, whose primary role is the day to day management and distribution of these artifacts, is required.
The data team is made up of experts who manage, and thoroughly understand, the company’s data. They don’t consume the data; instead, they prepare it for business teams to use in their decision making. This preparation can include a wide range of activities from configuring and creating tables in data warehouses and databases, and generating BI data models to building reports and dashboards for their business users.
Over time, business users lean more and more on the data team for their BI assets. They feel they don’t have time to learn, or have tried unsuccessfully to learn, how to use the company’s BI tools. So, they leave it to the data team because “they know what they’re doing”. As a result, data teams end up spending more and more time doing non-data related configuration, such as making small changes to reports and dashboards because the business users want to see something a little different.
In a worst case scenario, the data team becomes overwhelmed with support tasks and can’t meet the business teams’ demands. Faced with a delay in getting information, business users throw up their hands and go back to the familiarity of spreadsheets, willing to suffer the inherent challenges and lack of features so they can get timely results.
Cementing the relationship between data and business teams
Data experts (e.g., data engineers, analytics engineers, and data analysts) and business users (e.g., marketing, sales, finance, and customer success team members) have distinct roles within an organization. As a result, they also have different perspectives on how data should be managed and used.
Data experts make sure the organization’s data is accurate, up to date, governed, secure and, of course useful. Data experts are not concerned with how the data is ultimately consumed as long as the end users have access to, and are using, the right data.
Business users need to make data-driven decisions. They’re concerned with having access to correct, current data, as well as being able to trust the tools they use will give them accurate results. They want to understand what the data is telling them without having to deeply understand the finer details of the data.
Between the two roles is a sacred contract, where data is secure but also accessible and trusted. Unfortunately, this isn’t supported very well by traditional BI tools.
What is needed is a set of BI artifacts that fundamentally enforces this contract. One that encapsulates all the details around data configuration and management behind an abstraction that has meaning. And, one that exposes as much flexibility in configuration of presentation and context as possible without violating that meaning.
The answer? Metrics. By their very nature, metrics support the trust contract between data and business teams, unlocking business intelligence within a modern data environment.
Metrics are the BI artifacts business users need
Metrics deliver trusted data. Metrics are purely data artifacts, containing only data and exposing only data-based choices. They can be used in data presentation and analysis tools, but enforce no consumption constraints of their own. They differ from other BI data models in that they don’t represent structured sets of data that relate to each other. Instead, they contain only the data needed to support a single business measurement. All of the data configuration required to ensure the values are correct and accurately represent that business measurement are built into the metric and not exposed to the consumer. As such, no query can be written against a properly configured metric that will give the wrong results, or results that don’t match the metric definition. It is possible to get no results if there’s no data for the query that was sent and it is also possible to get query errors when things go wrong in the system. However, if the metric query returns values, those values can be trusted to be the correct values for that metric as it was defined.
Metrics enable comprehensive analysis. Metrics can be queried under different contexts, giving lots of control to the business user over which data to consume. For example, a query for a revenue metric can include different time periods, geographies, or departments. You can ask for the revenue in France for January, broken down by weeks, or the yearly revenue for North America, broken down by quarters. All of these queries give you the revenue numbers, but under different contextual choices. There is no way to get the revenue metric to give you anything but revenue, for example, profit instead. If your organization doesn’t have a profit metric, you can ask your data team to add one.
Metrics can be combined for presentation in reports and on dashboards. The same rich reports and dashboards can be created from metrics as with traditional BI artifacts, however it may take a few metrics to get there. Fortunately, metrics can be fluidly combined, making report building as easy as putting together your favourite Lego set. You know what all the pieces do and how they work, so you can trust they’ll come together perfectly.
Metrics are the BI artifacts data teams need
Metrics make for efficient data management. Initially, it may seem like a lot of work for data teams to build and maintain a set of metrics for an organization, but, when done properly, managing data as metrics actually leads to far less effort and frustration. For example, each metric is unique; if a metric breaks, only that metric needs to be fixed, reducing the time it takes to get data back up and running for end users. Also, if desired, the data team can build each metric independently, allowing them to divide their work into smaller, more manageable pieces.
Defined metrics bring clarity to data requirements. Metrics represent a business measurement that is understood by the organization, so it’s easier for data teams to fulfill requests appropriately (as each metric is clearly defined and communicated).
Build data-secure metrics instead of end-user reports and dashboards. The consumption of metrics by business users is carefully controlled through the metric query. This ensures end users can independently access metrics without compromising data security. The data team can focus on building and maintaining the data artifacts (a catalog of metrics), moving report building and data consumption into the capable hands of business users. It’s a win-win situation – Data teams spend less time configuring reports for business users and business users get fast access to data for time-dependent decisions.
But, what can you do with metrics?
Reports and dashboards. As mentioned, metrics can be used just as effectively as other BI data sources in reports and dashboards. However, where metrics outshine traditional methods is the ease with which you can create and configure those consumer assets.
Primed for AI consumption. The context built into a metric makes it easier for AI systems to consume and provide value from them. Data is often confusing for people, let alone AI systems. Much of this confusion boils down to a lack of context about what the data means. Metrics not only provide clarity for people, they also make data more consumable by AI systems. There have already been several demonstrations of chat systems enabling users to interact with natural language to get insights from a set of metrics.
Streamlined advanced data analyses. As carefully crafted, single-purpose sources of clean data, metrics are perfect for advanced analyses. In traditional BI, much of an analyst's time is spent preparing and cleaning data for analysis, but, with metrics, that’s already been taken care of. When new metrics are required, they only need to be defined once and can be reused indefinitely.
Metric-based analytics are trusted analytics
For business users, a key component in effective data analysis is trust. Trust in the data. Trust in their ability to use the BI tools. Trust in the data team. Unfortunately, the element of trust is missing for business users in many of today’s BI solutions. While traditional BI software may offer advanced features for data experts, who know the data well and understand how to perform advanced analytics, the needs of business users are often neglected.
Metrics have the potential to solve the trust problem in BI today. There are several technologies that enable metric-style BI, including metric platforms, like PowerMetrics, and semantic layers, such as those offered by dbt Labs or Cube.
Data teams are beginning to realize that existing analytics solutions are not right for the majority of their users (aka business teams). Most importantly, they’re looking into tooling that won’t require stakeholders to become data experts. By enabling their business users to perform self-serve analysis using metrics, data teams will also benefit. They’ll get to spend more time on the data-centric tasks they enjoy and are trained for and less time building reports and dashboards. Let’s close the gap between data and decision makers. It’s time to empower everyone to work with data and let metrics tell the story of your business’ journey to success.
Brought to you by Klipfolio and PowerMetrics
Klipfolio has helped thousands of people worldwide succeed with data. Designed and developed by Klipfolio, PowerMetrics puts data analysis and dashboard creation into the hands of business users with curated metrics, governed by the data team. Learn more about PowerMetrics on getpowermetrics.com.