Business Intelligence Best Practice
Foundational Concepts
At the core, Business Intelligence is a communication tool. We create visualizations, tabular summaries, dashboards, graphs, and other visualizations and summaries for one simple purpose, to provide a clear and accurate representation of data to strategists and decisions makers. To create meaningful and useful information for customers, follow best practice. In this process, the customer relationship comes first. When that foundation is solid, the next phase is to understand the relevant business rules and create clear definitions. After that, it is time to choose the right technology for the task. During development, it is important to follow the five C’s of Data. When the project is complete, it is time to deliver the product to the customer with clarity and objectivity. Finally, a follow up appointment with the customer allows for any additional requests, or a conversation about the next project.
Develop the customer relationship
When building relationships with customers it is important to effectively portray your integrity to the customer in a way that is not only consistent with your character, but also communicates clearly to the customer that you are committed. Showing predictability, reliability and availability will allow customers to feel comfortable with you and with your abilities. Set a schedule and follow it (within reason). When you are away, provide an out of office message via email and voicemail, especially if you are in the early stages of new engagement. Make a commitment to answer any messages within two business days, even if your reply is a short message to update the customer that you are still in the development process or that the request is in queue. When your customer calls, answer cheerfully and greet them warmly. Let them know when your schedule is open. Be available to them.
Understand the Business Rules
Business rules require definitions. Definitions are often tricky topics. Throughout a business, definitions and their connotations and attachments may differ. I once kept a table of definitions and added a column for each business center because customers in different states talked about different concepts in different ways. No matter how many trainings I held, their own definitions were deeply ingrained in them and it would take more time than the project allotment to socialize new definitions. Creating working definitions within the team and mapping them to a more central definition can be helpful in those scenarios. This can be tricky when communicating with multiple teams who have very nuanced understanding of the same concepts. When the business definitions are worked out, map the logic that governs the business process model. I start with a pencil and a large piece of paper, move to a whiteboard, take pictures, then map in an application (Microsoft Vizio). This process can be quick when there is clarity, or it can be slow when there is ambiguity. There are times when clarity will have to be found or created in the process. When the process and the definitions are recorded, take time to clarify the details with the customer.
Use the right technology
Relational database structures are the best for data sets. Avoid data lakes and other non-relational structures. It is important to be mindful of buzzwords when creating business intelligence solutions that are reliable. Too often, buzzwords will get the best of well-meaning managers that want to create the next big thing and jump in the front of the line at market. It is possible to use the right technology, even if it is simple, and jump in the front of the line also. Make balanced decisions about technology. Create impressive, eye appealing analytics packages that are factual and appropriate for the scenario. The hot topic of the day will inevitably change, but accurate and meaningful information is always in style.
The Five Cs of Data (Sherman, 2015)
Clean
Data should always be as clean as possible, allowing decision makers and leaders to sign off on their conclusions with confidence. Assure that all entries are valid and complete. Most data are collected with machines during regular business processes. If data is manually entered, ensure that user controls are in place and review the data before reporting. Create quality checks and run statistics on data to assure that it is complete.
Consistent
When data is consistent, customers can focus on their projects and results. Align business definitions and process maps with database elements and database logic to create a seamless process. While there may be many versions of many reports available to customers, when the process elements are aligned, you are able to create a consistent product from start to finish. Check totals against each other when there are multiple visualizations to assure that each version presented relates to the same consistent information.
Conform
Conform data to the business scenario and to the larger organization or project. When the analytical package reflects the big picture as well as the relative project, team members can share information easily. Reports are often passed from one team member to the next, especially when they are particularly informative or useful. Conforming data means that when this information is shared, the information is clear across teams.
Current
All databases update with various timing depending on processing. Current data is relative to the assignment. Data should be current for the situation at hand. When working with monthly or quarterly reporting, current data is last month’s data. When you are working with a dashboard reflecting ER wait times, current data is up to the minute. Establish what is current for the context.
Comprehensive
Comprehensive data is data that fits the decision-making process to the level of specificity and granularity required to make the best decisions. Including the option to “drill” into the details of the report line by line from the simple line graph created as a summary allows for operational teams to understand factors that relate to their job functions and meet their goals more effectively.
Delivery to the customer
A concise presentation to the customer with enough context, meaning and clarity assists the customer in the decision-making process. As the analyst, it is important to present information in a neutral manner, rather than persuasive. Keep in mind that the strategy of the decision maker may be unique or different from what you expect, and the information presented is an objective aide in that process. Maintain professional distance from the information and practice observing its nature. Communicate evidently what the data reveals and delicately resist any inquiry that would require you to become the decision maker. Data analytics, when executed correctly, are not meant to be influential.
Follow up
When the analytical package is delivered to the customer, schedule a follow up inquiry, or ask if a follow up survey is preferred. Surveys are a great way to allow customer the space for honesty if they prefer it. Others may want to chat for a while about the project. Be sure to find out if the product met their needs or if they require a little more clarification. Ask if there are any new projects on the horizon and discover what you can help with next.
References
Sherman, R. (2015). Business Intelligence Guidebook. Waltham: Elsevier.
