Adding self-service BI tools to your BI portfolio may decrease the development and support burden on IT. But the real value is in creating a partnership between IT and business groups, and empowering the people doing analysis to get the information they need in a timely manner.
Although acquiring the self-service tool(s) is a necessary step, it is just one of many in setting up a successful self-service BI program.
Gartner describes self-service BI as “end users designing and deploying their own reports and analyses within an approved and supported architecture and tools portfolio”. Today most database and BI vendors have a self-service offering. In the current Microsoft BI stack, Power BI (Excel, Power Pivot, Power Query, Power View, Power Map, and Office 365) provides self-service capabilities. Tableau is another popular choice. Both of these vendors offer on-premises and cloud options.
Self-service BI has been a hot topic for the last several years because of the potential benefits it can offer organizations. It can alleviate the burden of report development on IT/BI teams and empower the people who know the data to quickly obtain the information they need.
Self-Service BI as an Ongoing Program
To realize these benefits, you must set up a self-service BI program to go along with the new technologies. BI/IT staff can use self-service tools for quicker and easier development, but the adoption of these tools and processes by business users is what moves your company toward a culture of analytics and data-driven decision making. Purchase and setup of self-service tools is arguably the easy part of implementing a self-service BI program. Many times when IT is sponsoring the self-service initiative, they may build the first models and reports. But if the BI team members are the only ones using the self-service tools months later, it isn’t really self-service BI.
It is possible for IT to use self-service tools like Power Pivot and Excel/Power View to create parameterized reports for business users. Then IT deploys the report to O365 and users only have permissions to view, not download and alter. This may be quicker than using more traditional tools, and may have more data discovery features built in. But these tools come with technical and functional limitations for experienced technical users who are accustomed to the level of control found in traditional reporting tools.
The level of involvement of IT and enterprise data models in the self-service program can be different for every company, but business users’ involvement as report builders and/or data modelers is required to meet the definition and reap the benefits.
Growing a strong self-service BI initiative is not as easy as the marketing materials may make it sound. Defining the partnership between business users and IT is a common challenge. The balance can be different for each company and possibly each business group within the company. It is necessary to answer questions such as:
- What infrastructure will be put in place and supported by IT?
- Who will help with complex calculations and data models?
- What priority should be given to requests for help?
Although each company may define this partnership differently, it’s important to realize an improper balance between governance and flexibility can reduce the value for the business users and hinder adoption.
Tools don’t eliminate the need for report writers and data modelers to understand the data and the model in order to be effective. Building models and reports may require a good amount of time, which can be difficult if that is considered a supplemental part of an analyst’s job to be performed in her spare time.
Self-service BI usually intensifies the need for data governance and master data management rather than alleviating it. It can be more difficult to get a consistent answer and ensure common understanding of terms and calculations across the organization.
Here are some key recommendations to proactively address these common challenges as you begin your self-service BI program:
1. Decide what type of self-service BI culture you wish to cultivate. Refer to the article by Melissa Coates for more guidance on this subject.
2. If the self-service BI initiative is started by the IT group, make sure to have a conversation with business groups to ensure they are ready and willing to make the commitment. To begin, pick an initial business group partner who will provide feedback and help communicate successes to other business groups who are waiting to get self-service capabilities. Business groups who do not assume ownership of their data will have a harder time than groups who already take responsibility for data quality and process improvement.
3. Understand that not every business person will be a report builder. Some will just be consumers. It’s best if there is an analyst in the business group who is fairly tech-savvy. These are your Power Users. Also understand that not every group will immediately want the responsibilities that come with self-service BI. Each team must decide what to do with their limited time and budget, and self-service BI may not be at the top of their list. These groups may come around after seeing successes with other groups, or when the burden of going through IT for every report build or change outweighs the burdens associated with self-service BI.
4. Have at least one person on the IT/BI side who understands the self-service tools and the data available to help with complex data modeling and calculation challenges. As Rob Kerr points out in his article, self-service BI tools may have lowered the learning curve for building reports, but they have done little to make complex data relationships and statistical problems easier to understand and model.
5. Provide periodic training and support opportunities, and continue to follow up after initial training. Training doesn’t sink in until users have to apply it. Create templates for common data questions. Use these as starting places for data discovery using the new tools and data models, and reference them during training.
6. Facilitate conversations among data owners, report builders, and data modelers to identify terms that have various definitions across the organization, or where a unified data model supported by master data management would improve the quality of analytics. As Javier Guillen wrote in a previous post, the most successful and far reaching self-service BI implementations commonly have well-defined data models as a supporting infrastructure.
Ensure that you are setting up a self-service BI program to go with your self-service tools. It should be one that IT can support and in which business groups are willing to participate. Want even more self-service BI insights? Download this free blueprint for self-service BI success.