Business Intelligence is the means of turning raw data into useful insights. It allows you to collect data from different sources, organize it, and then enjoy the analytics. Do you need it? Probably yes, as it’s the most balanced view of the business you can get. But as you can guess, starting such a complex endeavor requires some preparation and we will hopefully help you with that. And if you've already introduced some BI procedures in your company, this article will help you get organized as well. Now, let’s talk about your Business Intelligence strategy.
A BI strategy will allow you to address all your data problems and needs, develop a cohesive system, and keep it maintained. What happens when you start implementing BI without a strategy? Basically, you focus too much on getting those graphs but no one in the company has an understanding of why and how to use them. Here’s a handy illustration of life before and after a BI strategy.
Previously, we posted a guide for implementing the BI practice in your organization. Now, we go into detail about one of the stages of your implementation plan - documenting a BI strategy.
Your BI strategy has three main elements:
Vision. Why are you building the BI practice in your company and what do you want to achieve?
People and processes. Who will define and run the BI strategy? And how?
Tools and architecture. Which dashboards and solutions do we want to build? For which areas? And how will they impact those areas?
So, as you look at your company’s corporate strategy, you envision what BI initiatives you want to start. Then that vision helps you choose the right people, those will use and maintain the chosen processes. To support people and processes, you use software tools. And finally, you establish the architectural blueprint for its development. Now, let’s go into detail on those steps. And we’ll start with those businesses that already have some form of BI.
To know where you’re heading, you must set a baseline. Say, you know that several departments have been using analytics, but the data has been mostly siloed - marketing people don’t have access to sales information, and customer support is tracking user feedback for their own internal purposes, or maybe there’s no analytics at all - basically, it seems to work but how effectively is unclear.
A vision is a combination of purpose and direction. There’s no strategy without a vision. It manifests itself in the form of many crucial decisions, like what data we will be sourcing or who will get access to the insights.
A vision also has a very mundane purpose: explaining to people in your organization -- who already have their favorite tools and processes -- why they need new ones, and how the transformation will happen.
Mike Biere, a former BI Analytics tools specialist for IBM, in his book The New Era of Enterprise Business Intelligence, offers this template for your BI vision:
Our corporate vision for BI is to create and support an infrastructure with secure and authorized access to data held anywhere in the enterprise. Our corporate standard for a BI tool is ________. We staff and measure our BI competency center based upon end-user satisfaction surveys and successful deployments. An important segment of our end-user community requires near real-time data access. Therefore, we have provided such an infrastructure to accommodate them. We currently support ___ users representing ___ % of our user population. Our goal is to increase the usage by ___ % by (date). We weigh the potential costs of increased BI usage against the business value and ROI we receive. Thus, we have a clear view of our success that is measured, accountable, and defensible.
Biere also notes that a vision articulated in this way will help fend off the executives whose only goal is to provide staff with cheap tools and get back to “real work.”
From rare reports to the role of Chief Analytics Officer
Define BI governance members, their roles and responsibilities, functions, goals, and relationships to different structures in your company. Involve people from all levels, from executives to end users to bring all their perspectives to the table. So, it’s not really a group of BI experts, but rather a board of representatives from different company areas.
Larger enterprises, however, would benefit from an all-expert group in a Business Intelligence Competency Center (BICC). BICCs help identify data needs, establish data governance structures, oversee data quality, and general data integration processes. These are programmers, data scientists and analysts, specialists in relational databases and reporting tools. We recommend establishing such a team if you have a set of specialists to spare or resources to hire them.
BI lifecycle is a framework supporting BI efforts or specifically the architecture and tools used for it. Normally, the architecture looks like a pipeline starting at data sources (your ERP, external sources, etc.), then following a data integration process or ETL, where data is transformed and loaded to the repository (data warehouse and data marts) before the data is finally displayed in dashboards and interactive tools. There are a few architectural styles with different configurations of system elements.
The data warehouse + data mart configuration of BI architecture
Choosing BI tools. Depending on your level of confidence, you can get an end-to-end platform or create your own mix for every phase of the BI process. You can also find a perfect match price-wise or depending on whether you want to deploy on-premise or in the cloud. Use our guide to the best BI tools to help you with this task.
Drafting the data integration process. Define data sources and make sure your BI tool can assess them. Make sure the data is of high quality and set up processes for data preparation. Consider the architecture of a data warehouse.
Ensuring data presentation. Establish what types of reports and dashboards your system will display based on end-user needs and KPIs. Use our guide for some data visualization ideas where we also list some handy tools and libraries.
Performing user acceptance tests. User acceptance testing is an often overlooked but critical process, where you ask end users to perform a few tasks and collect output information about the system’s usability and performance. Then you prepare test cases, choose the time of testing and pick the required tools.
Performing training. End users should be trained to understand data basics and use a visualization platform. Before that, non-BI proficient members of the governance team should be trained to understand data transformation phases. Basically, recognize where there’s a knowledge gap and make sure to fill it as soon as possible.
Apart from training, prepare to serve users and address their problems. Set up the feedback process and decide how you can react to it as quickly and cost-effectively as possible. Having a user support framework means that you cover them from three stand points:
Data education support - provide the knowledge base to let users find answers they have about incoming data: metadata, data purpose, metrics, data source, and so on.
Tool support - if possible, choose the tool with end users, then establish the agreed timeframes for issue response, and communicate what channels or contact they can use to get support.
Business support - make sure end users not only understand data but know how to derive value from it. Assign mentors in each department to help users learn how to provide value via BI tools, what metrics and dimensions to look for, and how to pinpoint data trends.
Having decided on how you approach all these tasks of the BI lifecycle, prepare your roadmap.
Here, a roadmap is a visual document demonstrating deliverables at different stages of implementation within the timeline. By this step, you already have all the data you need to organize and schedule on the map, you only need to set up time frames and deliverables for every task.
The roadmap can encompass only high-level tasks like “Find a BI vendor” or be narrowed down to “Create a list of top ten best matches,” but for strategic mappings, the high-level overview will be enough. Below is an example of such a roadmap spanning three quarters. It has deliverables and milestones and tasks are divided by teams.
This roadmap has planned activities, KPIs, and, of course, implementation dates Source: Roadmunk
How will you measure the success of your BI strategy? Quantitative metrics alone won’t do - you want to know how much better you understand your business or even how strong end users’ relationships have become.
You can revisit your BI Maturity model to determine whether you’re going in the right direction and establish metrics. The main one would be ROI, which describes if BI actually brought expected value.
Qualitative metrics would include things like the number of data access requests made by end users, increased productivity, or that deadlines are met more often.
You should also measure the effectiveness of the BI governance group - whether all goals were met and if the priorities haven’t shifted.
As we mentioned a few times in the article, a BI strategy is always created with a corporate strategy in mind. How will BI help you achieve your business goals? This is also the task of enterprise architecture - the practice of finding how tech can boost your overall goals.
To investigate a company’s current situation and propose changes, enterprise architects use frameworks - decades-old methods that help seeing the company from all its angles. TOGAF or the Zachman framework are also popular for BI strategizing. Filling in the templates, you can draft your requirements and align new processes without losing the vision. They are also useful for communicating your idea to the team and upgrading your project indefinitely.
As economist and Harvard Business School professor Michael Porter said, “The essence of strategy is choosing what not to do.” Indeed, even if developing a strategy is a time-consuming process, it allows you to eliminate errors that will inevitably arise without it.
