Getty Images/iStockphoto
Why a dashboard audit matters before BI cleanup
Managing large BI estates takes more than deleting stale dashboards. To keep order, leaders need to assess analytics usage and assign ownership to prevent clutter from returning.
In theory, a dashboard cleanup effort sounds straightforward. In practice, deciding which dashboards still matter to users -- and which ones don't -- can be one of the more difficult aspects of managing enterprise analytics.
To make informed decisions on keeping and removing redundant data assets, BI and analytics leaders need a dashboard audit that goes beyond a simple inventory. They need to understand who uses these views, what business purpose they serve, whether they overlap with existing ones and who owns them. This also requires a process for assessing whether the visualizations help drive business decisions or only create more work and unnecessary costs for analytics teams.
It's less of a technical cleanup project than a cross-function exercise in governance, business alignment and change management. And once order has been restored, these processes must be kept in place to prevent dashboard sprawl from happening again.
Sripathi Jagannathan, head of data engineering at digital services company UST, has seen those challenges firsthand. He pointed to one client, a global hospitality company, that had more than 300 dashboards and used a cloud migration effort to audit which ones were still needed. The company removed about 250 dashboards after determining that just 59 were worth keeping, Jagannathan said.
"We performed a detailed discovery and clustering exercise, identified redundancy and low/no usage and had business stakeholders prioritize dashboards into high, medium and low value," Jagannathan said. "The retained set reflects the dashboards that were actively used, materially relevant and needed to support near-term operations -- even as the client considered longer-term platform changes."
Over the past two decades, organizations have expanded access to data and analytics tools, such as BI platforms, enabling more business users and teams to create dashboards. That broader access made it possible to analyze data faster, but it also contributed to inconsistent outputs, conflicting reports and higher costs as duplicate or low-value dashboards consumed licenses, refresh cycles and compute resources.
"Now [organizations] want to rein things back in," said Georgia O'Callaghan, a director analyst at Gartner.
That's a tough task. Rationalizing dashboards requires careful analysis, effective change management and stronger governance to prevent further sprawl.
"It's not necessarily about having fewer dashboards. It's about making sure you have the right dashboards," said Alan Cecil, data analytics manager at professional services firm BPM.
Obstacles to dashboard cleanup efforts
Determining which dashboards to keep and which to retire presents several practical challenges during a dashboard audit, O'Callaghan said.
Documentation for dashboards is often missing -- particularly when business units create them rather than data teams. Visibility into how they're used is typically opaque. And users might resist cooperating with the consolidation if they think they will lose their pet analytics tools or reports. In other cases, the original users or owners have left, leading to dashboards that no one uses or manages.
In some organizations, the harder problem is finding someone to lead the dashboard audit and consolidation effort, because they don't have anyone who owns the governance of BI tools.
"The audit process itself is not easy," Jagannathan said. "You have to know how you will go about identifying redundancies in the dashboards, and they tend to be proprietary black boxes."
Who should lead the dashboard review?
To rationalize dashboards, organizations should start by assigning clear ownership of the initiative and of dashboard management.
O'Callaghan said the dashboard environment should ideally be overseen by "the people with the right expertise." In some organizations, that means the BI team. In others, enterprise architecture may take the lead if it owns the data platform.
However, O'Callaghan and other analytics experts stressed that dashboard rationalization must also involve IT and business units.
"It's not as simple as saying, 'Who owns dashboards?'" Cecil said.
He said each dashboard should have an owner responsible for making sure it supports business objectives, while the IT and data teams should manage the infrastructure, practices and policies that keep the dashboards available, reliable and trustworthy.
How to review and consolidate dashboards
The experts offered the following recommendations to streamline dashboard environments:
- Determine how much self-service access to grant. O'Callaghan said some organizations tightly control access to data and analytics tools, while others have relatively few restrictions. Both approaches have pros and cons. Strict limits might slow decision-making, while unrestricted self-service access can lead to dashboard sprawl. Determining an organization's position will guide subsequent actions.
- Consider usage in context. "You can't judge a dashboard only by how often it's used. It might look like it's not that important by its low usage, but it could be super critical," Cecil said. The people leading the rationalization effort must consider not only how often a dashboard is used, but also why it matters to the business. Nonetheless, dashboards with little to no use over an extended period are likely candidates for retirement.
- Identify duplicate dashboards and overlapping functionality. O'Callaghan said this step is difficult but essential to decide which dashboards can be combined without disrupting the business.
- Determine the value of individual dashboards. Organizations should prioritize dashboards that generate the most business value, Cecil said. "If you've identified a group of valuable reports or dashboards, invest in getting those into broader usability. You still will need to clean things up, especially if you're burning compute resources, but focus more energy on the dashboards that are useful," he said.
- Select the right processes and tools for the assessment. In most organizations, no single individual or team can review every dashboard, so leaders need surveys, usage data and other metrics to determine the value of individual dashboards. O'Callaghan said many organizations use questionnaires to learn which dashboards workers use and how they use them. Jagannathan recommended that teams also use monitoring and scanning tools to help identify duplicate or unused dashboards.
- Use change management practices. Encourage users to complete surveys and support the work to revise the dashboard environment. "There is a human component that needs to be managed here," Jagannathan said. Workers might resist retiring familiar dashboards even when it's the best option for the organization, so leaders should clearly communicate the business case for the consolidation.
How to prevent dashboard sprawl from returning
Dashboard rationalization can reduce sprawl and many of its related problems, but that won't stop them from multiplying again.
To keep that growth in check, experts advised leaders to:
- Decide whether the goal is only to remove unnecessary dashboards or also to redesign the remaining ones to better meet user needs and reduce demand for new dashboards.
- Create a searchable repository of dashboards so users can determine whether an existing asset will work for them before requesting or building a new one.
- Establish or strengthen governance to ensure each dashboard has an owner. The governance program should also include periodic dashboard reviews to classify dashboards into four categories: keep, consolidate, retire and redesign, Cecil said. In addition, it should establish a review process for new dashboard requests to limit unnecessary ones.
Mary K. Pratt is an award-winning freelance journalist specializing in enterprise IT, cybersecurity strategy and data management.