Expert advice on how to start with self-service analytics
In addition to an easy-to-use BI platform, keys to developing a successful data culture driven by business analysts include a data catalog, executive support and good training.
As of 2020, consulting firm QuantHub reported that demand for data scientists exceeded the available supply by 250,000. And that was before the COVID-19 pandemic struck and organizations that previously had not used analytics to inform decisions realized the importance of business intelligence when navigating economic uncertainty.
Not only are data scientists in short supply, but they're also expensive. The average salary for a data scientists was just over $100,000 as of May 2021, according to the Bureau of Labor Statistics.
"Talent availability is one of the top BI challenges," Elizabeth Espinoza, a research associate at Dresner Advisory Services, said on Dec. 8 during a session at Real Business Intelligence, a virtual conference hosted by Dresner. "Being able to provide self-service analytics can help address some of the strain that would otherwise be passed on to BI experts."
Self-service analytics is an approach to data and analytics in which business users are given tools that enable them to perform many of the data curation and data analysis tasks otherwise done by the data scientists and data analysts now in short supply.
Organizations opting for self-service analytics generally have data experts -- IT personnel -- that implement and oversee their organization's analytics operations. They set up end users with easy-to-use BI platforms that enable them to safely and confidently work with data and make data-informed decisions.
But getting started with self-service analytics is not simple.
While its aim is to make analytics accessible to business users with easy-to-use tools, underlying complexities can make the difference between a successful self-service analytics operation and failure.
Among the keys to success are tools that enable end users, commitment from organizational leaders and proper training that includes power users within different departments, according to Espinoza and her co-panelists at Real Business Intelligence.
A breakdown of self-service BI vs. traditional BI.
While setting up the platform and managing and preparing data may require the skills of a data scientist or data engineer -- particularly knowledge of coding languages like R and Python -- low-code/no-code capabilities enabling data exploration and analysis are necessary given that business users rarely know code.
Many BI platforms are designed for self-service use, including Tableau, Qlik and Power BI from Microsoft, which have been around a long time. Those from more recent startup vendors include ThoughtSpot and Sigma Computing.
But self-service analytics requires more than just an easy-to-use BI platform.
Data catalogs serve as libraries for an organization's data products such as models, reports and dashboards.
Talent availability is one of the top BI challenges. Being able to provide self-service analytics can help address some of the strain that would otherwise be passed on to BI experts.
Elizabeth EspinozaResearch associate, Dresner Advisory Services
They're where data products can be categorized and organized to make them easy for end users to find in order to inform decisions. They're also where system administrators can set up data governance frameworks that include access controls that simultaneously protect the organization from misuse of data while enabling business users to confidently work with data.
"You need to have really good documentation and spring for a data catalog," said Jonathan Sharr, data analytics manager at Dorel Juvenile, a manufacturer of products such as strollers and car seats for children. "Part of self-service analytics is governance, how security and access is structured, and who is going to be able to put new content into production."
"That balance between self-service analysis and data governance is so important," she said. "You do want your data to agree, and you want to make sure that the data sources you're using are reliable. Data catalogs can be a very useful tool in helping users to understand what data they're using."
Beyond enabling end users with the proper tools, self-service analytics requires organizational commitment. An organization that deploys a BI platform and sets up a data catalog but does nothing more -- one that provides tools but no other support -- is destined to fail.
Therefore, buy-in from the top is critical to self-service analytics success, according to Susan Gershman, chief customer innovation officer at Prophix, a corporate performance management software vendor
"Buy-in and engagement are so critical," she said.
One way to show buy-in from the top is for senior leaders to build BI into their presentations, she continued. There, senior leaders can demonstrate the benefits of self-service analytics to other potential users.
"Build [self-service analytics] into a boardroom presentation," Gershman said. "Look at some analysis, look at a graph or a dashboard and drill in right there. Start at the top, use it, and it will be much more likely to be adopted and used across the organization. It just becomes part of the fabric of how people work and think."
Similarly, Sharr noted that a C-suite-backed approach to self-service analytics is important. Without commitment from leadership, buy-in from business users is less likely.
At one organization Sharr is familiar with, a senior leader held a recurring meeting with employees. During each of those meetings, he used the organization's BI platform to discuss key performance indicators to show how self-service analytics can drive business decisions.
"Top-down is key," Sharr said. "In any organization I've been in, [self-service analytics] has to be a top-down function. If senior leadership is beating the drum and using the tools, that gets buy-in quite quickly. The best thing you can do is have your leader not just talk about [analytics] but do it. Then you will pretty quickly see adoption."
Training
Once the right tools are in place and organizational leadership is committed to a data culture, training is key.
That training includes not only teaching business analysts how to use their organization's BI platform and data catalog but also teaching data literacy. Data literacy is the ability to derive meaningful insights from data.
According to Valerie Logan, CEO and founder of The Data Lodge, data literacy training should include a blueprint that makes data literacy part of a data culture rather than a separate skill, workshops that engage end users as they learn, data leaders within departments who can give hands-on help as end users become more familiar with data, and a pragmatic launch that begins with easy wins to instill confidence.
Like Logan, who served as the keynote speaker at Real Business Intelligence, Sharr emphasized the importance of appointing data leaders who can lead at the departmental level.
"Find power users, super users, and really leverage them," he said. "Enable them to evangelize and train those around them. They're where [IT personnel] can't be to spread the word and do additional training."
Sharr added that training -- both data literacy and usage of BI tools -- should happen as soon as possible, whether an organization is just starting to adopt a data culture or a new employee is hired by an organization already committed to self-service analytics.
Gershman noted that overcoming resistance to self-service analytics is a challenge when transitioning to a data culture. Change is uncomfortable for employees used to doing their jobs a certain way, and change management is a process.
Communication, therefore, is a critical part of the training cycle.
"It takes time to get the buy-in and get people to realize that [self-service analytics] does drive efficiency over time. But it's so exciting when you turn the corner and it clicks and people start seeing what they can do," Gershman said.
She added that as business users buy in to their organization's adoption of self-service analytics, training can be incorporated directly into a platform to guide end users as they become more familiar with BI tools.
"Once they're self-sufficient, there's such an improvement in efficiency," Gershman said.
