Henrik Dolle - Fotolia

Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Top tips for a successful self-service BI strategy

As self-service BI adoption continues across enterprises, it's important to make sure you have a good strategy. Here are the top tips for a successful self-service deployment.

Self-service BI has been trending among enterprises, and a strong strategy is necessary to ensure success after adoption. Done well, a self-service BI strategy can empower users to discover hidden value and opportunities throughout enterprise data.

"Self-service BI and analytics strategies aim to enable users to easily develop, consume and use their own tools for the analysis of data," said Jorge Garcia, principal analyst of business intelligence and data management at Technology Evaluation Centers, an IT solutions advisory service.

It's easy to overfocus on the technology, but a successful self-service BI strategy also must include the users, their goals and their pitfalls. Garcia suggested self-service BI projects also address ease of use, training needs and overall data governance strategy.

"Business users are not necessarily BI or analytics specialists, meaning ease of use is a crucial element," he said.

Here are some top tips for actualizing these factors in a self-service BI strategy.

Find balance

Garcia said striking the right balance between ease of use, data availability and governance can be an elusive goal. It's important to perform preproject evaluations and self-assessments to establish what's already in place from technical and business perspectives. Next, consider what's necessary in order to deploy a successful self-service BI platform.

This is best achieved by assembling cross-functional teams for the planning, design and development of the necessary tools, which will also enable proper communication with management stakeholders.

"We need to ensure IT departments understand business data and processes and can provide users with the necessary tools while respecting data governance policies," Garcia said.

Create dedicated support

A strong self-service BI strategy depends on making sure BI tools are available to all users, no matter their positions or skill levels. Stephanie Duran, Qlik BI manager at JBS, a U.S. food processing company, said training is key. BI developers should train all staff to be able to develop, she said. It's important to note that other stakeholders -- including customers -- may also use the app.

"A dedicated support team for business developers and users is critical in a self-service BI strategy," Duran said.

JBS has a five-person team to support their use of Qlik across the enterprise. Additionally, the organization's other IT teams know what the BI environment delivers and supports the Qlik team.

Even as a manager, Duran works to find out what business users need and help them set up the appropriate guardrails. JBS also runs weekly collaboration hours where the entire IT team is online. During this time, business unit developers can call in with scripting and design questions with assistance from all business units.

Prioritize data readiness

There are a variety of hidden costs involved in getting data ready for a self-service BI strategy, said Abby Hao, electronics engineer at WellPCB, an electronics manufacturer. She has found installing the tools and software needed to handle and deliver custom self-service analytics can be expensive. The cost gets crazier when you factor in things like support and maintenance for these projects.

The best solution is to adopt cloud BI tools, Hao said. She found this improves budgeting because there is clarity around initial setup costs, and everything else has been covered through a consistent monthly subscription package.

Understand the business problem

"Understanding the business problem seems simple, but this crucial step can make or break your entire strategy," said Holly Rachel, co-owner of Rachel and Winfree Consulting, a data analytics consulting firm.

Understanding the business problem involves breaking it down into specific parts with an eye toward how they impact business users.

The best ways to overcome these obstacles is to first dedicate a good amount of time to defining and understanding questions you want answered. Nicole Kosky, director of services at AnswerRocket, an interactive visualization platform, recommended teams start with the outcomes that will drive value for the business. Good questions to ask include what you want to achieve with analytics and what business questions you are trying to answer.

These questions ensure your solution meets the needs of your end users and generates buy-in that can improve adoption rates. For example, a business outcome could be improving the performance of your brand. Business questions could be about what's driving brand performance and how your brand compares to competitors. The self-service BI strategy should answer these questions with meaningful visualizations and insights.

Don't start with your data

Many organizations start with a focus on data, attempting to perfect a data lake or warehouse and stand up a self-service BI project on top of it, Kosky said. But this approach can easily result in BI that doesn't provide business value because blending data can compromise what's most valuable to business users. End users often receive visualizations and computational or statistical insights, but the output rarely helps them make better decisions and move the needle on performance.

Teams often start from the data perspective and imagine the data will tell the story and find the patterns.

"This is just not true," said Geoffrey Lakings, sage strategist at Rippling Nature, a data analytics consultancy. BI users need to know the story so they can identify the sources of data they need to wrangle and transform, massage and convert into a proper data model that can be visualized.

Create appropriate access

Developing enterprise-wide roles can make it easier to manage access rights and improve governance for self-service BI projects. The right level of appropriate access to data sources is important for users to get the most out of the data, said Joseph Chong, the founder and CEO of Acxtron, an IT consultancy company that specializes in data analytics and business intelligence.

Common data sets that have companywide access should be identified, and other more selective permissions can be given to data sets. In Power BI, organizational data is usually classified as public, private or organizational. This provides better structure when it comes to setting policy while enabling users to bring in their own data sets.

Ensure people can trust the data foundation

People will only use vetted BI infrastructure if they can trust the results.

"Many times in my career people did not self-service their requests due to trust issues in data quality, hence preferring to stay out of it," said Dimos Papadopoulos, marketing analytics manager at PepsiCo.

Besides ensuring their questions are answered, a clear process needs to be set with regard to data cleansing and responsibilities in data quality.

Balance curiosity and confidence

A healthy dose of curiosity is required for users to undertake their own self-service research. This needs to be balanced with confidence they move in the right direction.

"Telling someone to be data curious doesn't mean they're immediately able to ask the right questions of the data," said Ben Schein, vice president of data curiosity at Domo.

Users need confidence in knowing they won't misinterpret a metric, see data that isn't allowed or even cause an error. Some of this confidence requires proper governance to enable data curiosity with the right guardrails.

Craft a comprehensive storyboard

A self-service BI strategy should make it easy to start projects with a storyboard to guide an exploration. A comprehensive storyboard can clarify the goals of a BI project, Lakings said.

A storyboard enables strategists to perform a thorough gap analysis and ensure they fully understand the problem. It can also ensure they apply the appropriate data sets, data modeling, analytical modeling and visualizations to explore the right story.

Dig Deeper on Self-service and collaborative business intelligence

SearchDataManagement

SearchAWS

SearchContentManagement

SearchOracle

SearchSAP

SearchSQLServer

Close