alotofpeople - stock.adobe.com
Maximizing the value an organization can get from its BI technology starts with putting the right business intelligence team in place.
BI emerged from a need to dig deeper into the numbers previously only available in operational or management reports. A standard report presents business data in a fixed format reflecting what a developer or IT team has decided is important.
For example, an organization may see weekly sales by category and region. However, if sales numbers are falling, reports really don't help an organization understand why because they don't enable exploration of the data.
A report represents one hypothesis of what matters and how to approach it. Business intelligence enables the organization to examine its own hypotheses, finding new answers to new questions.
Some organizations may already do this in some form or another. So, the first question to ask is why an organization needs a specific business intelligence team.
The BI team and the enterprise
The simplest answer may be that the skills don't exist today in an organization and therefore they're not doing much in the way of formal business intelligence. Nearly all organizations have some operational and management reporting.
Most organizations do have some form of analytics going on, using spreadsheets as an entry-level tool. It's still common to find organizations that lack a formal analytics capability. So, that's an excellent reason to build a business intelligence team.
However, it is probably more common today to find that these functions -- building reports, dashboards and analytics -- do exist within an organization but are not well-defined and have some functions that are very informal.
With self-service BI technologies, it's common to find users who have discovered a data visualization tool online and adopted it for their own use. One department could visit another department and find they adopted a different tool because they discovered it separately.
In fact, although software vendors love to describe their land-and-expand strategy, most often they land-land-land with different departments adopting software independently. This can feel tactically effective for each department but rarely leads to good strategy.
Whether building a BI team from scratch or recognizing an existing loosely coupled or virtual capability, it's best to clearly define the relationship of that team to the rest of the organization from the outset. The more complex and more common scenario arises when BI roles already exist more-or-less formally throughout the organization and need consolidation.
Ask whether the team will be virtual with members in different departments or a new formal organization -- but before that, consider the team's functional structure. Four specific roles are critical to business intelligence team success.
Key roles in the BI team
Any good business intelligence team comprises the following four roles in some way. It is even common for one individual to wear a couple of hats. That's OK, as long as they don't feel overstretched or undersupported.
The subject matter expert
First, someone needs an excellent understanding of the business and how that relates to current IT processes and applications. In other words, it is someone who understands the business process well and who also feels familiar with the data architecture or the IT infrastructure.
An organization can find this person from two directions: the business specialist who has informally learned the infrastructure and the data architecture; or an IT specialist, probably already working in data management, who has been in their role long enough to really understand the business.
This role can easily be under stress with demands, so it is important to share knowledge of data systems and business practices widely as part of the BI strategy.
Another role is an analyst who can create effective models and representations of the business with the database or BI tools of choice. This captures the relationships between data, the specialized calculations and the hierarchies of data by geography, tax jurisdiction, sales region, product categories and so on. Some of this modelling may already be in place within the IT infrastructure, especially if the organization has a data warehouse.
But even then, business intelligence teams often find they need to create models ad hoc for specific purposes, such as an experimental advertising campaign or a special project for executives. That model probably does not exist within a data warehouse designed to provide an authoritative long-term record of the business. So, the team needs someone capable of bringing data together from multiple sources, building the right relationships and creating the right definitions and calculations.
An analyst can possibly build some reasonable dashboards and visualizations. They may not be the most aesthetically pleasing, but they may prove quite functional.
But that's not enough today. As the data literacy of business users increases and as executives grow used to working with newer and more compelling technologies in their personal lives, they generate more demands for excellent designs -- not just functional reports. When an exercise bike offers better analytics than the organization's analytics department, they need a designer.
However, this role goes beyond pleasant charts. Compelling visualizations prove easier to understand and enable more effective communication. This in turn leads to more engagement, better decisions and generally more executive interest in the work.
The data steward or governance director
A data steward oversees policies for corporate governance, ethics, privacy and security as they flow through the IT team to the BI team, and from there to business users. This is a demanding task, but in the beginning, it may not be a full-time role. Database architects and business analysts can both take on the work because, like the subject matter expert, this role may start as a technical one with acquired business knowledge, or as a business role that also supports technical knowledge. This may even be a shared function.
Very often, a data steward needs support from legal counsel to understand legislation, from HR to review employee privacy implications, and from IT managers to define how policies and permissions can be practically applied.
Sustaining and retaining the BI team
Whether creating a business intelligence team as a new organization or simply a virtual commitment from existing roles, the team must have continual support and development.
For example, if the team is virtual, at some point it may need to transition some members to full-time commitments as the volume of work and its importance to the business grows. With both organizational and virtual teams, an organization may need to add new members with specific skills, such as machine learning.
Analytics skills are very much in demand today, so businesses should also consider how to reward and retain the business intelligence team. For many analysts, the most important recognition is not financial -- they can move elsewhere easily if they need that -- but they do look to be significant advisors to the business.
Acknowledging the contribution of analytics to operational, tactical and strategic decisions can add greatly to the sense of purpose in a BI team. The ultimate measures of success -- increased sales, reduced costs and so on -- may have many complex factors behind them.
Whether an organization is thinking about starting a BI team today or is already beginning to invest in architecture, tools and people, the future of analytics depends on recognizing and rewarding the significant contribution it can make.