Starting a BI project in 4 simple steps
What would it mean to you and your enterprise, if you could start getting useful business insights from your data in literally five days or less, using four simple steps?
As exciting as this seems, it’s actually just what a good business intelligence platform should be able to do for you. While BI projects can be short term or long term, straightforward or sophisticated, they should all bring actionable results as soon as possible. Business moves fast nowadays, and there isn’t enough time for months of preparation, data modeling, IT platform planning, management decisions, and implementation.
Fortunately, these four clear, do-able steps will allow you to publish your first BI dashboard in five days, keeping up with the pace of your business without needing specialist help or extensive resources.
STEP 1: Map out your BI project with small, practical milestones (half-a-day)
Why do certain BI projects fail? Often because they try to bite off more than they can chew. Start off by focusing on one insight of value, and your BI project can already be a success in just days. Afterwards, there will be plenty of opportunities to derive further insights, making sure each additional step brings you a measurable benefit.
So, let’s begin! Here’s how to do step one:
- Start with a standard business process you want to understand better or improve
- Keep data sources few at first, with a just 2-3 reports that hold the answers
- Get an initial, useful result, before iterating to go deeper or wider into your business processes
This also means using a business intelligence system that lets you start simply, and then scale to any level of BI that makes sense for your organization.
Allowing half-a-day for step one, your BI project map will then look like the following steps for the rest of the week (the 4.5 days left)
- Business planning to define useful questions to answer (step two, below)
- Setting up your data model to bring your data sources together properly (step three)
- Designing and publishing a dashboard to display the results (step four)
Remember that as you progress with your BI projects, your BI tool should let you go beyond just automating any manual business reporting you are doing currently (Excel spreadsheets included). A little business rethinking may show you even more important questions to answer, for which your BI tool will then become even more valuable. That’s when you start reaching beyond the realm of standard reports and into the realm of BI.
STEP 2: Collect requirements (half-a-day)
To get your first successful BI project off the ground in five days, requirements should be modest. On the other hand, business, result, and technical requirements should be stated clearly and precisely enough to keep your BI project on track for success:
- Business requirement: state the question that is to be answered. For example, 'what are the trends in the monthly revenues of the organization?' Or, 'which product lines can use more marketing budget to generate higher profits?'
- Result requirement: decide how a result from the BI system should be displayed or communicated, so that the business teams involved can understand and act on it as quickly and as easily as possible
- Technical requirement: what hardware and software will be needed for the BI project? If you can use standard PC hardware, for instance, you can meet technical requirements that much more easily. Sisense, for example, both runs and scales on a standard PC, handling up to terabytes or billions of rows of data with full BI functionality quickly and efficiently, as needed.
STEP 3: Select and compile your data sources (2 days)
Business intelligence needs inputs of data to produce outputs of results and business insights. Data can come from many different sources (some BI tools have built-in data connectors that make it super easy use data from different places). Remember, data must be correct to start with. Otherwise, the end results will be flawed. Here’s your to-do list with detailed examples:
- Select the data sources you want to use to answer your business question (see step two above). You might choose your organization’s sales database, existing Excel spreadsheets with financial data, Google Analytics data on the number and type of visits to your enterprise web site, or some combination of such data sources.
- Understand the correlation between the data sources you want to use. For example, your sales database and your financial spreadsheets might both list your products: the sales database showing how well they are selling, and the spreadsheets showing how much they cost to make. Using the two data sources, your BI tool could show you how to maximize profit by putting more marketing resources on specific products.
- Join the data from different sources for one version of the truth. Sisense lets you use simple 'drag and drop' to bring different data sources and tables into the same central, high-performance database, called an ElastiCube. Everybody then uses the same version of the collected data, avoiding arguments and allowing people to focus on the results and conclusions of the data analysis.
STEP 4: Build and display your first BI dashboard (2 days)
Remember the result requirement from step two above? In this final step, it’s time to create the displays that help your own users understand the results from the BI tool and the data it has analyzed.
Sisense gives you numerous options to produce web-based dashboard displays, reports that can be distributed to groups of users, and interactive analytics to let users ask new questions and explore further. Here are some great dashboard templates by industry. Your goals in step four are:
- Identify your target audience. Seek to understand, before trying to be understood! A business management audience may want more intuitive overviews and indications of trends, compared to a technical audience looking for more detail. So, use a corresponding approach to your dashboard.
- Design your dashboard. Sisense provides options for graphs, charts, and filters that can also be accessed by dashboard viewers to make the dashboard as useful and as engaging as possible. Dashboards are also accessible using a standard web browser, meaning that your viewers do not have to use any additional plugin or download.
- Information design. Common sense will play an important role here. Looking to show a trend over time? A line chart might be the simplest and most effective way. Or perhaps you want to show how overall sales of different products compare? A pie chart may be the right choice. If in doubt, remember the KISS principle (Keep It Simple, Stupid!).
Actionable results from data using BI in one week
By following the steps above, business users can start their business intelligence journey simply and effectively. They can also rapidly accomplish data management and analysis tasks that would otherwise have taken months of IT resources.
Author: Elana Roth