As data-junkie, I love Business Intelligence (BI) tools. With ease, you can query, explore and visualize data in tools such as Tableau, Qlik, or PowerBI. As someone who has dedicated his career to analytics, I used to evangelize these tools at all the companies that I work with. But, while consulting for Fortune 500 companies, I was surprised that most employees at these firms didn't even realize they had BI tools. Or frequently, employees had a license and didn't use it. In previous roles as an analytics executive in tech, I led two different BI integration projects for MicroStrategy and Tableau... and each time I failed.
Based on an abundance of anecdotal evidence, I realized BI Tools have an adoption problem. After some research, I found that my hypothesis was not just a hunch. A Gartner study1 shows that BI adoption is ~30% -- compared to >95% for tools like PowerPoint and Excel, which have more than a billion users. Office software continues to be the primary way that companies share and analyze data.
Why Employees Do Not Use Their BI Tools
Companies are clearly spending billions on these tools2, but why aren't their employees actually using them? In my experience, users only adopt new software when it can help them either save time or make money. So, I asked myself the following questions:
Do BI Tools make employees more efficient?
Not really. If executives want custom analytics on a large dataset, they tend to siphon off time someone who can code for the data pull. Although marketed as "self-service", BI tools are rarely used that way.
Do BI Tools generate content that I can present to my customers?
Rarely. Few companies trust their analytics teams to put unprocessed data in front of clients absent context. Most companies still use presentation software for this.
To understand the problem better, I again leaned into the breath of my experience in analytics and consulting. I estimated that for every 100 people at the companies that I worked with, each employee could be categorized in to one of three groups:
- 2-5 people who are BI power users that understand how and why data is structured a certain way and how to manipulate it
- 10 – 20 people (typically leadership) who view the dashboards created by the power users
- And the remainder of people who might be aware that the tools exist, and definitely never look at them.
Again, research supported the anecdotal evidence. Every few years the OECD surveys nearly 100,000 people3,4 across the advanced economies and tests their computer literacy. The results are shocking to any technology leader that isn't a UX designer:
A couple noteworthy data points:
- 66% of technology users cannot find data in a spreadsheet
- 94% of technology users cannot filter the data based on information they found in another app
Given this data, think about how complex a BI software product is for ~95% of software-using population. A user needs to be able to understand: a relational database's structure, the business logic underlying the relationships database tables, different table join types and how they effect aggregation function results, how imperfections in source data they manifest in the database, among many other complexities that need to be understood to use BI tools accurately. It's too much for the occasional user to understand. Knowing this, BI tools' lack of adoption boils down to a simple reason:
BI software is incomprehensible to most people!
Trends: How People Actually Meet Their Data Needs
Frustrations with BI tools aside, companies still demand more and more data analytics, so how are they getting the work done? Fortunately, organizations are adapting to a couple of growing trends that help leaders get the data they need to support decision-making.
Trend #1: Growth of Purpose-Built Analytics Tools & Low-code / No-code Applications
Increasingly, “purpose-built” analytics tools are quickly filling the demand that “general data access” tools built by BI providers cannot meet. Purpose-built analytics apps are easier for business users with domain-specific knowledge to use. For example, Datadog leads the cloud infrastructure monitoring space, the ELK stack is tuned for logging, etc. Even power users much prefer a simple drop-down menu, search bar and button click to get the data they need over manipulating a data-cube in a BI tool. Purpose-built tools meet a large and specific need, and consequently are growing rapidly.
Also, One of the best applications of low-code and no-code tools is reporting. A new generation of low-code and no-code tools makes it easy for business users create GUI-based apps that can issue simple queries to data sources. These tools allow user to apply the concept of a "purpose-built" analytics and make them specific to a company, a product or even a department. As these tools grow in adoption, the need for traditional BI tools will decrease.
Trend #2: Business Users with Coding Skills
Coding skills are proliferating beyond engineering teams. Increasingly, sales and marketing teams are hiring "Insights Managers" to be accountable for the coding work their centralized analytics teams do not prioritize. Professional services firms want junior team members that can write Python code, not Excel formulas. The number of students learning to code in high school is increasing exponentially6, and coding bootcamps are expected graduating more than 30,000 students this year7.
Unless your analytics leadership is trying to build a fiefdom inside your company, this trend is embraced. As a decentralized analytics model allows the routine work to be self-service, and it leaves the best data scientists and engineers at a company to focus on more-complex, strategic projects.
The Addressable User Base for Traditional BI is Getting Squeezed
The combination of these trends decreases decreases the addressable user base for BI Tools. If the skills required to use the most popular BI products often exceed OECD level 3, the addressable user base for even the most user-friendly of BI tools is a maximum of 2-3% of the working population. The number of potential BI users is also limited at the top of the skill curve. For users that can quickly access data via coding, BI tools become a cumbersome appliance. Why bother to prep, load, and maintain a data warehouse for a BI tool, when it's easy to write a SQL query straight off the source data? Ultimately as these trends advance, traditional BI tools will become obsolete.
Changing Your Approach to BI with Presalytics
I was disappointed when I figured out that traditional BI Tools will forever be impractical for daily use at most companies -- the inherent complexity of their design and current market trends make it so. But out of that realization and disappointment, I sensed an opportunity.
I built Presalytics.io to empower data analysts to automate their work and play larger role in a company’s go-to-market processes. Presalytics flips the analytics experience on its head, so producing reports and analytic content starts and ends in a format that your whole organization, rather than just the power users, can edit and share. Presalytics learns from and adapts itself to fit into its user’s workflow, rather than forcing its users to adjust the way they work to adopt a cumbersome new technology. With Presalytics, turning analytics into insight is faster, and training cost is lower.
To save users time, Presalytics is built on top of tools, such as PowerPoint, Google, Excel, etc., that employees throughout your company use each day. For your power users, Presalytics’ no-code and low-code tools make content automation and reporting easy, eliminating repetitive activity from their workload.
With its middleware and APIs, Presalytics integrates natively as a presentation and reporting layer into CRM platforms, productivity software and other no-code applications, so users can incorporate analytics automation and collaboration within the platforms they are already using. Working with Presalytics minimizes the new skills your teams must learn:
- Salespeople can work in PowerPoint,
- Business analysts can work in Excel,
- Data scientists can work in Python,
- and they all can collaborate on the same content for your customers.
OECD ‘Skills Matter’ Survey. 2019 Data. Population surveyed limited to workers aged 16-65 in OECD Advanced economies. Chart excludes 26% of population that is not computer literate.
OECD Skills Outlook. Summarized.
IDC Software Developer and Skilled Worker Estimate. Assumes this segment is a subset of the Level 3 segment in the OECD study and split out for use in this analysis.
Course Report Coding Bootcamp Study. Estimate for 2020.