This guide is aimed at helping business and IT leaders understand the key components to a successful implementation of Power BI in their organization. It provides a breakdown of how to align the business case, what key things to consider when planning and how to develop a strategy to accelerate adoption with the end users.
Using this guide your organization can better unlock the value potential of Power BI as an enterprise analytics tool. As the size and availability of data continues to grow at a rapid pace the need becomes clearer for defining a strategy to align how and when you can help our organization deliver insights from data. Building a good foundation around aligning business cases, planning for user interactions and content creation will better position your organization to capitalize on newly available data and future advancements in data technology.
At Decision Trek we have partnered with several clients to develop new Power BI capability and optimize the value of existing deployments. Through that journey we discovered some key considerations and approaches that are a best practice to follow when using Power BI to deliver insights to an organization. Distilled down, we’ve learned that the most successful implementations start by gathering a clear understanding of what behaviors or outcomes an organization would like to drive using data insights and then going back to the data to develop solutions that can address those needs in a suitable way.
Craft The Business Case
- Start with Why?
Any successful technology implementation starts with understanding why making this investment today will return value to your organization tomorrow. In defining the why, it’s helpful, though not always easy, to map a new or existing capability to one or more of the key strategic objectives your organization is trying to achieve. Doing so will serve to provide clarity to both your development teams and business users on what should be prioritized and how success should be measured in the early stages of implementation and beyond.
- Align Stakeholders
For many organizations, making the business case for Power BI will involve several top-level business objectives and competing priorities. If you are planning to address multiple business objectives at once, you will need a methodology for prioritization. In lieu of clear guidance from executive stakeholders, you can use a systematic approach that evaluates several factors of a given business objective and ranks the importance of one objective against another. For example, a sales dashboard might rank high on revenue and overall business impact compared to an operational dashboard which might rank lower on business impact but higher on expense reduction.
- Define Success
When crafting a success statement, it is key to describe how Power BI will address a particular business challenge by defining what behaviors or outcomes will be driven from the insights generated. For example, if growing sales is a strategic objective for your organization, you might define success as, “Develop dashboards that unlock channel insights and identify opportunities for our sales team to address underperforming channels”. Another example of a strategic business objective might be to improve operational efficiency. In this case you could define success more broadly as, “Develop a capability to analyze operational data in near-real time across multiple domains of data to identify opportunities for reducing costs”.
Plan the Implementation
- Consider Deployment Methods
As an enterprise business intelligence application, Power BI comes with the capability to provide content and insights in a variety of different ways. You may decide to choose one or multiple engagement methods based on your organization’s goals and data maturity. The key is balancing the need to unlock timely insights while preserving trust in the insights generated.
Traditional Enterprise BI
All analytic content is owned by a single IT team or multiple IT teams. The cost-benefit to this method is controlling a single version of the truth across the organization but at the expense of scalability and timeliness of information.
Centralized Semantic Model BI
IT and business users can generate analytic content from a centralized dataset developed by IT. The cost-benefit of this method is dissemination of information and insights are less constrained by IT bandwidth. This comes at the cost of complete control of what information is developed and used by the business to make decisions.
Fully Democratized BI
Business users are free to combine datasets from existing sources and generate their own content sometimes with little to no IT involvement. This can lead to rapid insight generation, but without significant controls in place can erode end user’s trust in the data insights produced.
- Aligning Data Governance
In planning for a Power BI implementation another important consideration is understanding how data is currently managed at your organization. The size and scope of our organization’s data will influence how to best implement Power BI to meet your organization’s goals. Your deployment could involve integration with an existing data warehouse platform or developing a new data warehouse capability to support a Power BI deployment. As Power BI becomes a centralized platform for your users to access insights it’s critical to align development efforts and data policies. Additionally, to ensure appropriate access controls, it’s best practice to develop a security model that can govern access to sensitive data. Power BI comes with a robust set of security configurations that can control access to specific content in the service as well as data level security.
- Develop a Product Roadmap
The product roadmap serves as a tool to track and communicate timelines for the release of new capabilities. This phase of planning should include identifying dependencies tied to your deployment efforts. Dependencies can include availability of data, securing developer resources, training end users, and setup and configuration of new environments.
The roadmap should be structured to place a bias toward delivering insights as quickly as possible when taking all your dependencies into consideration. In practice this could mean using Power BI’s built-in ETL capabilities such as Power Query or Dataflows to source data for an analysis before that data is made available for general use such as in your organization’s data warehouse. These dependencies may also become a deciding factor when ranking your organization’s business objectives and lead to reprioritization.
Another key consideration when developing your product roadmap is forming a plan for managing technical debt. In most cases there will be trade-offs that need to be made to ensure insights are provided quickly at the cost of incurring some technical debt upfront. As your user engagement grows, make sure to document these trade-offs and allocate time in your product roadmap to address technical debt that may have accrued during content prototyping.
Create a Content Deployment Strategy
- Content Management
A content management strategy will provide clarity on how to better target specific audiences and increase overall adoption of your Power BI deployment. Some of the key considerations in developing a content strategy include identifying your consumer audiences and content creators, understanding the types of content that will be created, reviewing access management requirements, and licensing needs. To guide this process, it helps to start by reviewing all the stakeholders involved and deciding on the best way for them to interact with the platform. Each stakeholder can be classified into one of three types, Consumer, Power User and Engineer.
Consumer
A Consumer is typically a business user who will be making business decisions based on content created. For these users Information and insights can be pushed directly via an App or can be controlled through a specific workspace in the Power BI service. Licensing for these users can be either on an individual basis or as part of shared capacity.
Power User
A Power User can be either a business or an IT user who is generating analysis or content based on existing data sources or datasets. Each power user will require a developer license to share content in the Service.
Engineer
An engineer is typically an IT resource who develops reports or creates datasets that enable power users to perform self-service analysis or consumers to view insights. These users will also require individual developer licensing.
When reviewing your consumer audience, it’s also important to understand each user’s data literacy and what actions or outcomes are intended when consuming specific content. As an example, for consumers that need to use data to make daily operational decisions but may not be trained to navigate dashboards, you can choose to use Power BI Paginated Reports to display data in a simple table format. For mid-level management consumers, you may choose to deploy curated datasets to facilitate self-service insights. Executive management consumers may need not have time to perform self-service analysis, in this case developing dashboard that display corporate KPI’s would be a good approach.
- Rapid Prototyping
One of the great value propositions of Power BI as an enterprise analytics tool, is the speed at which insights can be generated. The focus of engineers should be to practice good data modeling techniques that enable rapid prototyping of visualizations or analysis. Additionally, Power BI comes with a suite of AI capabilities that can further expedite the generation of valuable insights from data. In keeping with the ethos of rapid development of insights, it’s best practice to follow an agile development process. The goal should be to rapidly prototype solutions for end users while continuously gathering feedback to make iterative improvements to content. This also serves to help mitigate the risk inherent in older generations of BI solutions where business requirements change prior to the completion of content.
- Monitoring Engagement and Continuous Innovation
To understand the success of your implementation it is best practice to define some success metrics for your deployment. Common success metrics include overall user adoption, distribution of user types, counts and types of content created, end-user sentiment scores, number of new insights unlocked and an effectiveness score around driving desired behaviors or outcomes linked to the corporate goals.
Power BI is an ever-involving tool with product updates released monthly. As your organization continues to mature its capability to deliver timely and actional insights there are new features that can continue to unlock value. One immerging feature that can further improve timely delivery of content is using DevOps for continuous integration and continuous delivery of content. With the recent release of Power BI Projects, developer teams can now control and deploy updates to reports in code.
