Is Self-service BI a Myth?
In this series I cover business intelligence topics such as self-service BI, data-driven decisions, promoting a data culture, increasing user adoption, and more.
Imagine a world in which business users from across your company have, at their fingertips, an easy-to-understand data repository, gathered from both internal systems and external sources, that contains both live and historical information. A world in which they are creating reports, dashboards, and apps using friendly tools made just for them. A world in which the creation of shareable business analysis is no more difficult than creating a Word document or a PowerPoint presentation.
The benefits of Self-service Business Intelligence are clear, but most IT leaders and digitally minded executives know of this world only as an unrealized dream. A myth in the marketing.
It sounds easy: Step 1) Choose an easy-to-learn data visualization tool like Power BI, Tableau, or Qlik. Step 2) Connect the tool to your data. Step 3) Train business users to use the tool. So why isn’t self-service BI everywhere?
Failure to launch a BI program can often be traced back to unrealistic goals and expectations.
Good Reasons for Starting a Self-service BI Program
Analytics on the Ground
Business happens through constant decision-making. Hundreds and sometimes thousands of decisions are being made every day in your organization that will determine performance outcomes. Some of these decisions are made at the very highest levels such as by the Board and C-Suite, but most decisions are made on the ground at the day-to-day operational level.
Traditionally, reports built by the IT department supported decisions being made at the highest level (P&L, Balance Sheet, Cash Flow, etc.) or maybe senior leaders one rung down (Weekly Sales and Shipments, COGs and Margin, AR and AP Aging, etc.). IT would also build reports supporting more tactical decisions for managers on the ground, but the lead-time from request to delivery for these was long and other IT priorities like keeping the lights on and supporting larger strategic projects always seemed to slow down report requests.
With a self-service BI program, business users can build reports and dashboards themselves, cutting out the IT middleman. As you’ll see later, this is valuable for both the business users and the IT department.
Business users are in the best position to understand their own data and ask the best data-centric questions. Giving them direct access to a well-defined data model allows them to experiment with their ideas, something not possible or cost effective with a back-and-forth report development process owned by IT. For good reason, self-service BI is often called data democratization.
Data Dream Team
Now that business users are building and modifying their own reports and dashboards, you may be wondering what the report writing analysts in the IT department will do now. For one, they should be the best evangelists for your new BI program, providing tips & tricks to business users and performing training on new features that come out after your launch. On the technical side, they will spend more time on data enrichment (adding more sources and fields to the data model) and data governance (cleaning up data at its source and enforcing master data management rules). These are ongoing processes that will get easier and less labor intensive over time. At some point in the maturity of your program, these analysts may transform into data scientists who come up with new use cases for analysis using artificial intelligence and machine learning. This can often lead to the discovery of new revenue streams or business models.
Going Beyond Reports
As part of building your self-service BI program, you built a new data repository that is validated and well-described. Now you can start using it for applications beyond reporting. All best-in-class applications for functions like supply chain management, sales forecasting, demand planning, and marketing automation require a connected data source to feed their AI/ML algorithms with historical and real-time data such as sales orders, purchase orders, invoices, packing slips, etc. Rather than connecting these tools directly to your ERP which can have performance implications and very complex mappings, why not connect them to your new data repository? It was already well-documented for business users, so you can simply hand that documentation to your vendors and drastically reduce implementation time.
Business users can also develop apps on top of the new data repository using low-code/no-code workflow automation tools. Any app idea that does not require writing back to the original data source would be a good candidate such as a workflow that is triggered when certain thresholds are reached. Business users can experiment and try out their ideas just like with their reports and dashboards because the data is read-only, and the visual development tools are made for non-IT professionals.
Bad Reasons for Starting a Self-service BI Program
Staff Reduction
While it may seem like shifting the responsibility for report writing from the IT department to business users would allow for a reduction in staffing, the opposite is usually the case. The data team in IT will have the new responsibilities mentioned above, and if company culture is to truly change, the hiring of new data-minded staff in other departments is often needed. To get the most value out of a self-service BI program you must have people with the right innovative mindset and initiative across your organization.
IT Cost Reduction
Most new self-service BI programs have a one-time cost associated with the implementation of a new data architecture, but there are also ongoing costs to maintain the program. Continuous training of business users, data governance costs, cloud usage fees, and licensing for your chosen visualization tool are just some examples. That said, you should be able to include both non-IT cost reduction and revenue growth in your business case that will directly result from the insights your company’s business users will gain from the new program. After all, the entire point is to become a more data-driven organization which means discovering inefficiencies as well as new opportunities. Quantifying this ROI is not easy. This is where an experienced BI partner can really help early in the project.
Complexity Reduction
You may also be thinking that if business users are able to understand the data model and build analytics solutions with user-friendly tools then the overall system must be less complex. Unfortunately, the opposite is true. To enable the simplicity of the front-end, you must add at least some complexity on the back end. For example, there may be more ETL processes (extract, transform, and load), more environments to maintain (dev, test, prod), more data governance processes and more security overhead to deal with. And overall, you’ll end up with more points of failure that must be mitigated or at least more closely monitored than before.
…but most IT leaders and digitally minded executives know of this world only as an unrealized dream. A myth in the marketing.
Eyes Wide Open
I believe strongly that the value of a self-service business intelligence program greatly outweighs the costs in the long run. But to have a reasonable chance of realizing that value, management must go in with the right goals in mind and a change management plan that addresses cultural transformation.
In the next article, I’ll cover the technical challenges you may face launching your self-service business intelligence program.
Daniel Lucas is the founder of THRDparty Advisors which is on a mission to protect private capital from cyber risk, digitally transform portfolio companies, and maximize exit returns. They are IT executives advising PE firms and PE-backed companies through every investment stage. Visit https://thrdparty.com for service details, case studies, and pricing.