What do modern self-service BI and analytics really mean?

Self-service business intelligence (BI) has been a topic of conversation since I can remember.

It sounds simple. Let me start with an intuitive, naive definition:

Self-service business intelligence allows business users to access data and create their own insights.

Sound right?

It’s easier said than done.

Ask the internet about self-service data analytics. You will be met with the following schools of thought:

  • Naive (often promoted by vendors of data visualization tools)
  • Skeptical (coming from seasoned data warehouse folks)

The proponents of the naive approach cheer with optimism: “Self-service BI is about beautiful charts and dashboards. This is exactly what our product does!” (Quora is full of answers like this.)

The skeptics are… um… skeptical: “Self-service BI is impossible. It always leads to chaos. Like spreadsheet hell, just without spreadsheets but more pricey.”

Do either of them have a point?

Yes, they both do:

Sure, Mr. Naive: Nobody wants ugly dashboards. Beautiful and easy-to-use charts and dashboards are a must-have.

Yes, Mr. Skeptic: Let business users generate reports, and they will immediately start drifting toward a swamp of duplicate and untrustworthy content.

But which one is correct?


Skeptics’ concerns about the risks of the naive approach are obvious.

On the other hand, what if you take self-service capabilities from business users completely? The level of agility demanded by modern businesses can rarely be delivered by centralized IT teams.

Agile analytics are an imperative these days.

Agility and self-service

What does agility mean for that idea of self-service BI software?

Nobody implements self-service BI just for the sake of self-service BI. The real need behind this mysterious and ambiguous term is agility.

There are well-established agile methodologies adopted by the most famous software companies of today. Most of them have a number of things in common:

  • Simplicity. Avoid analysis paralysis. Prototype new ideas quickly.
  • Small, iterative, and frequent releases reflecting users’ needs.
  • Communication and collaboration. Business and technical people work together (this one requires some clarification in the context of BI and analytics - see below),
  • Process. Please don’t take “Individuals and Interactions Over Processes and Tools” from the Agile Manifesto too literally. Yes, it can be a lightweight process, but you do need one.

So… where are the self-service analytics?

We’re getting there.

First, one clarification. Modern BI is different from development of consumer software products: The distinction between business users and BI developers becomes blurry.

When you use Google to search the internet, it’s clear. You are the user while the folks in Mountain View wearing t-shirts, jeans, and flip-flops are the developers.

What if you are a user of a business analytic application who is somehow literate with data?

You don’t have to be a BI developer or a data scientist. You can still do some ad hoc reporting or data analysis, and you can prototype reports and dashboards that are worth sharing.

Of course, not everyone is like you. This is why it is good to start your self-service analytics efforts with user personas.


Get a demo
(Guided tour + Q&A)

Analytic personas

You may be literate with data to that point that you can create some promising prototypes, but you may easily miss a detail or two.

The lady from the marketing team may be on the top of it when it comes to data flowing from digital marketing systems, but she may be a little lost about product telemetry.

The receptionist may not be the right person to create a sales forecast dashboard. Neither is the CEO; she may have the knowledge but not time.

Finally, you have a data engineer or a data scientist who can crunch the data to get insights you never ever knew you may need. Or, more importantly, an engineer’s discipline when it comes to taking care of data quality, testing things out, and keeping them maintainable. However, their intuition will be somehow limited when it comes to which insights are actually useful for running your business.

The true agility happens when all these personas get the right level of analytical power and can work together in a well-defined collaborative environment.

Three principles of agile self-service BI

The best-kept secret of agile self-service business intelligence is in these three principles:

  • Give the right people the right set of tools and capabilities
  • Create a collaborative environment
  • Create a process to govern how this collaboration leads to continuous improvement of your analytics

If these principles are followed, each of your personas—like the marketing leader, the CEO, or data engineer—can use analytics and BI successfully. The marketing leader has the tools she needs to drill down into relevant data and use it to improve how her team functions. The CEO can reference high-level insights without getting bogged down by more detailed team-level data.

A more data savvy individual contributor can easily create new insights or new computations or even upload new data sets to prototype what he or his bosses need.

Finally, an administrator with an engineer’s discipline and a sense for software life cycle management makes sure that the good ideas come quickly to production on time while still meeting certain quality criteria.

Together, each of these distinct personas can use BI to improve their own performance, the performance of their team, and, ideally, the performance of their organization as a whole.

How do you get there?

I am working for a data platform company. It is tempting to close this blog post with a bold statement that our product is the only true self-service BI tool.

But if you read this blog to this point, you know I am not going to do that for two reasons:

  1. A sustainable self-service BI is not something you can just buy.
  2. The whole self-service thing (if done right) is just about making your data analytics environment more agile.

There is some obvious minimum set of analytics capabilities for a data platform that will help you achieve this sort of agile self-service environment (besides tablestakes features such as data visualization, pivot tables, or dashboards):

  • Semantic layer that enables business users to simply articulate powerful data queries across multiple datasets
  • Strong role-based access control that gives the proper level of content management rights to the right people
  • Data governance framework to prevent chaos from taking over your solution
  • Customization support so business users can extend the curated semantic layer by uploading their data or creating their own calculation while distinguishing them from the curated set of data
  • Prototyping support so power users can clone existing analytics artifacts or environments to prototype new capabilities

Again, no tool by itself will get you there; knowing your personas and setting up a clearly described agile process is the key.

May 29, 2019
Blog Bottom - Try GoodData

Want to ask about something specific?

Contact us