The Role of Data Visualization in Activating Business Intelligence

Feb. 13, 2018

BY TARA CARTER

  

What can Business Intelligence (BI) actually do? Can it measurably improve company performance? Investors say yes. In a report by ATKearney, 60 percent of executive managers found BI to have a positive impact on shareholder value and give their firms an edge in Return On Equity (ROE) twice that of their competitors.

Can BI help managers peer into the future? Yes, according to statistics by the National Bureau of Economic Research. Averages of BI-driven data forecasts by non-experts outperform 96 percent of economic forecasts by individual industry experts.

Can BI help the average manager internalize market wisdom? This one is a bit trickier. BI has a vast power to change business reality, but only when managers fully understand what the data is trying to tell them. Data speaks its own language, and the best translator is a well-planned data visualization. With the right storytelling approach applied at the formation stage, data visualization has a proven ability to turn lifeless charts and figures into motivational models that impel action.

Visualization vs BI Visualization

There are two common forms of business analysis tools:

  • Front end solutions (commonly referred to as visualization solutions) and
  • End-to-end solutions (commonly referred to business intelligence visualization tools, or simply BI visualization tools)

Visualization tools create visuals and dashboards. BI visualization solutions involve a platform backend that consolidates, syncs and cleanses data from multiple sources in a central repository. BI visualization solutions are able to share real, actionable insights.

Visualization tools are restrictive - they are limited by the number of aggregations you can input in each formula, as these solutions focus on reporting data instead of actually analyzing it. Pre-calculations, such as summarizations, are necessary prior to launching calculations. For example, sum and average calculations cannot be done at the same time - each needs to be done separately, saved, and then considered together.

On the flip side, BI visualization offers data + insight. It’s an end-to-end solution in that it consolidates, syncs and cleanses data from multiple sources and allows for needed pre-calculations to be automatically compiled so users can immediately obtain their needed data.

The bottom line: don’t be fooled by the dazzle of a pretty visualization face when selecting a solution. It may not represent the best way to handle the requirements of your data. BI visualization, where all data is fed into one central database, removes the gritty background stages of data preparation and data joining, as these tasks are automatically done for you. In other words, BI visualization enables you to get comprehensive, confident and actionable insights from your company’s data.

Implementing a BI System

Aside from identifying Key Performance Indicators (KPIs) and how they should be presented in dashboards and reports, a BI implementation project that consolidates data in one central location (aka a data repository) should include data cleansing as well as a solid process for syncing new data. This often includes redefining the data hierarchy and formulas. An essential focus of the implementation should be a robust backend that can handle large amounts of complex data from various sources.

Source: ATKearney Report

BI Implementation Phase 1

As a first step in implementing a BI system, it's important to identify your KPIs and the overall level of business intelligence that is needed. Aligning to business requirements is key. This includes ensuring the KPIs reflect the company's objectives and making sure you have all the data you need to calculate the KPIs.

Identifying the purpose of the BI system is key. Is your BI system for finance only or will a supply chain management component be needed as well? And what functional areas of the business will be integrated into the project? It’s important to clarify scope up front, as KPIs for manufacturing, for example, can be vastly different from that of finance due to different business models.

BI Implementation Phase 2

After vetting through the business need, requirements need to be collected in order to build a proof of concept prototype. This includes identifying functional requirements of the tool as well as specs regarding report and dashboard design. Here, it's important to consider who will be accessing the BI tool and assess their needs.

For example, a dashboard for management will be very different from detailed reports for an analyst, which will be more time intensive to design. Once all is identified, a detailed concept is shared with a system integrator to implement. At this point, a proof of concept is key, as it allows both the developer to completely understand the requirements as well as gives stakeholders a means to view and test the solution design before a full system integration.

BI Implementation Phase 3

After a prototype is approved, the BI solution is implemented. Ideally, the solution is released in phases, the first of which should contain all needed management dashboard functions as well as some reporting requirements. Additional data and KPIs are then included in subsequent releases.

Data Visualization and Goal Setting

With the backing of a robust BI visualization solution in place, your company is able to implement insights to drive more intelligent business decisions. Before the kickoff any collaborative business project, a management team is far more successful when they hammer out a shared vision of what they want to accomplish at the end of it all. Data analytics can come into play at every stage of a SMART goal-setting process. SMART goals are those that meet the following criteria:

  • Specific – target a specific area for improvement.
  • Measurable – quantify or at least suggest an indicator of progress.
  • Assignable – specify who will do it.
  • Realistic – state what results can realistically be achieved, given available resources.
  • Time-Related – specify when the result(s) can be achieved.

While this more traditional approach to team alignment has proven motivational and effective in some verticals, managers in every department at Google use a modified version called Objectives and Key Results (OKRs).

OKRs differ in that the emphasis is on the data-related aspects of the KR portion. Each employee has three to five objectives and each objective should be measured with three key results. Objectives tend to be short, personal (addressing assignability) and simple (emphasizing realistic).

In defining their objectives, managers are encouraged to be ambitious and come to terms with uncomfortable uncertainty. “The most difficult goals produced the highest levels of effort and performance,” according to a comprehensive review of academic goal-setting research.

The key results to be measured for each objective are highly important to note. Identifying these results before project commencement and monitoring throughout the process, makes the stakeholder feel personally accountable for its success. This translates to action on their part to ensure goals are met.

This shifts the emphasis from intent to measurable steps in progressing toward the goal. It’s a data-first mode of thinking that enables more strategic business decisions and helps managers see their next steps more clearly.

This is part of a larger trend in business of eliminating IT from the chain of data to information to knowledge to wisdom. Data visualization has proven to be key for effective BI. The best BI data visualization software in the hands of a data storyteller can transform the most non-technical manager into a data-driven specialist.

Translating Data Into Action

The vast majority of people need a model to visualize the results of data analysis in order to absorb the essential information quickly and convert it into deep knowledge about the findings. The wisdom piece emerges from recommendations on how to apply that business knowledge in making more prescient decisions.

BI data visualizations – which go beyond simple charts and graphs – can hold the viewer’s interest long enough to grasp the subtleties in the data. They allow decision-makers to (often literally) connect the dots to see patterns, trend lines, inter-relationships, stacked dependencies and opportunities for growth.

The secondary benefit is that BI data visualization more closely aligns with the collaborative nature of modern distributed teams, condensing what used to be covered in a time-intensive face-to-face meeting into an instantly communicative image. These images are easy to share and distribute to team members across time zones and devices. The data dashboard has taken on the role of the modern data visualization staging area.

Elements of BI Based Data Dashboard Design

In the old days, managers or their assistants would just select a collection of data points in Excel and run the chart wizard. Three-dimensional bar graphs or elevated pie charts were the primary tools of data visualization at the corporate level and they were easy to cut and paste into a 50-slide PowerPoint.

Companies that depended on consuming their data that way have, for the most part, tended to languish or fade away in the face of more agile, global, mobile competition.

BI based data dashboards, customized to the priority concerns of each stakeholder, communicate urgency and make it clear which steps can help to redirect the course of business. Building those data dashboards is an art, however. C-level visualizations and dashboard arrangements will have a very different look and feel than those of a frontline manager or active data analyst.

No matter whether the team chooses to work from SMART goals, OKRs or some other form of KPI management, the dashboard will serve as both a trusted digital advisor and a central command center where business leaders can hold stakeholders accountable for achieving their goals.

Within the framework of the dashboard, data visualizations require another sort of artistry. Even minor variations in the presentation of data can end up telling radically different stories to the viewer. Data visualizations emerge not from the data itself, but the actions the viewer will be impelled to take when they understand the ramifications. The most successful data visualizations begin with viewer's goals, indicate the best option for moving forward, tell the story in a personal way and then offer the viewer a way to drill down deeper into layers of detail.

The augmented manager, combining expertise in the field with the wisdom of a BI dashboard, is a living symbol of the future of work. Data visualizations will be the life support system that managers will rely on to swim in the sea of data that defines this new reality.

BY TARA CARTER
BUSINESS INTELLIGENCE, DATA VISUALIZATION, DASHBOARDS, ANALYTICS, FORECASTS, KPIS, OKRS, SMART GOALS