Top 7 Must-Have Features for Your Modern BI Tool

Written by Tomas Gabik  | 

Share
Top 7 Must-Have Features for Your Modern BI Tool

Deciding which business intelligence (BI) solution to implement in 2022 is no small feat. The industry standard is higher than ever: Once highly prioritized features such as drag-and-drop functionalities and interesting chart visualizations are now considered standard, thus carrying less weight in the decision-making process.

So, which features shouldn’t be overlooked when seeking the best BI tool today? What are critical considerations that can improve the ease of adoption for both data engineers and business users?

Let’s examine the top seven must-have features of an excellent, modern-day BI tool.

1. Multitenancy That Enables Analytics Across Multiple Customers

When distributing and potentially selling your data product to customers, it’s essential that each customer can only see and access its corresponding data.

Unlike other BI tools, GoodData offers straightforward and painless sharing of personalized data for multiple, different tenants. The main advantage originates from one simple source of truth: the existing logical data model (LDM), which is loaded in the desired data structure. This source is replicated for each of the affected tenants, while showing only user-specific data. (For more details, dive into our documentation about multitenancy or a more specific implementation use case shared within our GoodData Community portal.)

In order to show GoodData’s multitenancy in action, we created a Github repository. With the COVID-19 dataset, you can segment the desired information into multiple workspaces (user-specific areas where metrics, dashboards, and reports are stored). For the COVID-19 data in the Czech Republic, the workspaces represent individual counties. Note that all insights, metrics, and LDM definitions apply to child-workspaces (workspaces with a dependency to a master workspace), as well, while filtered to show only user-specific information.

For the COVID-19 data in the Czech Republic, the workspaces represent individual counties.
For the COVID-19 data in the Czech Republic, the workspaces represent individual counties.

Additionally, your customers may want to benchmark their data against data from other organizations. GoodData allows comparison between aggregated data of different clients or industry standards. Adding a new dataset with aggregated data — to support your benchmarking needs — will do the trick. (Dive deeper via this GoodData Community thread.)

2. A Logical Data Model That Speaks the Business Language

Data modeling capabilities are a must-have for any competitive BI tool available in the market. The option to visually and easily decide the relationships between designated entities is an essential prerequisite for any BI project.

Directly linked to the aforementioned multitenancy use case, one of GoodData’s key benefits is its approach to data modeling — as demonstrated by its LDM feature.

Essentially, an LDM describes sets of used data in a meaningful (aka logical) way. The model can be set up independently to a source database that establishes a foundation for components of the semantic layer in data management systems. When created properly, an LDM enables the creation of new metrics, reports, and insights without relying on complicated joins or lookups.

The creation of an LDM may require some level of technical expertise as well as sufficient business acumen. However, once created, it provides business users and BI analysts with an optimal background to create all desired metrics and insights without the need of adjusting the existing data relationships. Essentially, it focuses on the most important data-related task: interpreting data and using it to help your organization. The metrics, insights, and dashboard sitting on top of the existing LDM could afterward be redistributed to multiple client workspaces, where individual adjustments of your clients/end users can take place. (For further information, please read our blog post on logical data models.)

Here is an example of an already existing LDM (prepared in the GoodData.CN Community Edition demo).

Example of already existing LDM
Example of already existing LDM

3. The Ability to Connect Nearly Any Data Source

You probably know the story: The business decides that a new software has to be implemented, and they want to track data coming from it … ideally starting yesterday.

Our recent introduction of Dremio integration aligns with GoodData's vision: connecting virtually any database to your BI solution should not be a problem. Dremio is the latest addition to a catalog that includes Snowflake, Redshift, BigQuery, PostgreSQL, and Amazon S3, among others. Plus, not only are these ready out of the box, it’s also easy to set up. Through the GoodData API, you can easily create a connection to any preferred schema in the existing database, all while ensuring you get your data at near real-time speed.

Try this for yourself by using the GoodData.CN edition and replicating this POST command:

{
  "data": {
    "attributes": {
      "name": "prod-db",
      "url": "jdbc:postgresql://localhost:5432/prod",
      "schema": "public",
      "type": "POSTGRESQL"
    },
    "id": "prod-ds",
    "type": "data-source"
  }
}

4. A Metric Editor With Intelligent Query Completion

The ability to perform complex calculations and aggregations is key to any BI tool. GoodData’s metric editor provides the end user with the ability to create custom metrics for reporting. The Multi-Dimension Analytical Query Language (MAQL) is the engine of the machine.

The key advantages of MAQL include the following:

  • No joins or sub-joins as MAQL works on top of LDMs and its queries are context-aware.
  • Any metric can be immediately used for reporting, reused again, or deployed to assemble other metrics.
  • MAQL makes multidimensional analysis simple by abstracting any data complexities. You do not have to specify the fact or attribute origin as it is done automatically for you.

After the creation of a set metric, the ability to format appropriately is key. GoodData provides you with out-of-the-box features to select whether your metric should be a currency, a number with several decimal points, or any other format you like. (For more information, visit GoodData University.)

5. The Ability to Drill to URL

In general, driving engagement with your data product can be tricky. One foolproof way to do so is by keeping user experience top of mind. For example, your end user may spot an irregularity or an interesting development in a report, and then want to further examine this issue in the source software. By providing the ability to drill to URL, you can help facilitate — and streamline — this process.

Let's see how easy it is to set up in your dashboards:

6. The Ability to Attribute Filters That Are Metric-Specific

Creating a report — which is going to contain a metric filtered by some attribute values — is nothing new. When selecting a metric, you would just apply a filter in the Analytical Designer (GoodData’s environment that allows users to create their reports and visualizations, as well as further data exploration) and then adjust based on your preferences.

However, what if you’d like to select the same metric multiple times, but with each time filtered by something else? GoodData has you covered: Instead of creating a new metric in MAQL with the filter covered there, you can simply select attribute values, which will impact only the chosen metrics.

Instead of creating a new metric in MAQL with the filter covered there, you can simply select attribute values, which will impact only the chosen metrics.
Instead of creating a new metric in MAQL with the filter covered there, you can simply select attribute values, which will impact only the chosen metrics.

This makes it easy to create a new quick insight (without the need of generating a new metric); it also provides the ability to compare multiple attribute values against each other.

The finalized insight may look something like this:

Finalized insight in GoodData
Finalized insight in GoodData

7. Different KPIs Affected by Different Dates

When creating a new dashboard, the BI analyst will realize that the reports added to the dashboard are likely to be filtered by multiple different date dimensions.

Let’s consider a report that shows how many rentals of DVDs your business accrues, and then add a second report that demonstrates how many DVD returns occurred per month. We would like to show both of those next to each other on the same dashboard, but each one must have a different date dimension. GoodData allows the user to easily select which date dimension (present in the data model) is going to affect the filtering’s end result.

A report that shows how many rentals of DVDs your business accrues
A report that shows how many rentals of DVDs your business accrues
A second report that demonstrates how many DVD returns occurred per month
A second report that demonstrates how many DVD returns occurred per month

Both of the reports defined above are present in the same dashboard, and users can filter them by different data dimensions. The same goes for filtering the dashboard by attribute values.

Next Steps

When considering the integration of a new BI tool, one should be aware of its technical requirements and features. Understanding the current and, more importantly, future use cases for your data is key, too. Falling into the trap of tantalizing visualization widgets and other superficial features may prove costly; the priority always should be to provide the business with accurate and fast information while making the lives of data engineers and BI professionals as easy as possible.

Header photo by Tara Winstead from Pexels

Written by Tomas Gabik  | 

Share

Subscribe to our newsletter

Get your dose of interesting facts on analytics in your inbox every month.

Subscribe