Dashboards.those graphical displays of business intelligence that blink on executives’ PC screens, always had a high cool factor. Even Luddite CEOs could use them. If the light shone red, it meant trouble in River City. If it showed green, you could go back to sleep. But have dashboards actually delivered on their early promise? Do they pay real dividends? Judging by two user organizations we polled, the answers are yes and yes.
“Ultimately, like all business intelligence, dashboards are about better decision making,” says consultant Royce Bell, the UK-based CEO of Accenture Information Management Services. “A lot of companies are also now dashboarding their compliance status, but that’s something that has only grown up in the post-Enron era. Mostly they help [managers] make better, faster decisions.”
The term dashboard was probably a poor choice, implying as it does the simplicity of a car. “The light on a BI dashboard is not fed by a simple sensor like a car,” Bell points out, “but usually by several sources of data.” Accenture and some vendors now talk instead about metric management systems. Dashboards address just one part of what business intelligence (BI) is about: event-driven metrics. This event has occurred – sales in the west fell below this threshold or the flow of raw materials to this factory slowed – and you need to do something about it.
In fact, dashboards can be a life saver, particularly in industries with complex relationship chains, such as custom manufacturing and engineering, Bell says. For example, they can alert managers when there are bottlenecks in a supply chain that might cost the company lost production time or even contractual penalties. “I’ve seen cases where [dashboards] saved [companies] a lot of money.”
Dashboards have evolved since they first appeared. They’re easier to implement and easier to integrate with real time data sources. In the early days, they provided a monthly, weekly or at best daily snapshot of the organization. Today, they increasingly present near real time information. And dashboards are being driven down the organizational chart. It’s not just senior executives who use them anymore. In some organizations, as we’ll see, even front-line workers are getting access to dashboards.
The Workers Compensation Board of Alberta (WCB Alberta) implemented a BI system from Cognos Inc. in 2001, part of a larger Y2K initiative. The entire project, including development of a data warehouse, cost about $10 million, says Dirk Smith, director of program development in the operations group at WCB Alberta. The Cognos implementation represented about $1 million of that. The system has paid big dividends.
One of the first metrics the organization focused on was the number of compensation claims still outstanding after three months. Those claims represented a huge chunk of benefits payments. Sometimes they were outstanding because they involved very serious injury. Often, though, they were stalled because not enough attention was being paid or the agent wasn’t pursuing a resolution in the most efficient manner. Monitoring that metric drew attention to the problem.
“As a result, in the last four years we’ve cut the number [of claims still outstanding after three months] in half,” says Smith. “We’re not doing that by cutting off [claimants], but with right interventions, by resolving [claims] fairly. It has saved us from $60 million to $90 million in compensation costs annually.”
Not all of that can be directly attributed to dashboards or even to BI and the data warehouse, he says, but the technology was certainly part of the solution.
WCB Alberta began deploying dashboards in 2003, building them with add-on Cognos tools that cost about $200,000 for licenses providing dashboard access to 300 users. The dashboards use a classic presentation style.
An onscreen display shows red, green or orange lights for a variety of metrics – including number of three-month-old claims. Clicking on a light allows the viewer to drill down into detail.
Initially, the organization only dashboarded its balanced scorecard and a corporate risk watch list – a total of 25 to 30 metrics. Today there are dashboards for every division and group covering more than 600 different metrics.
It’s mostly divisional, group and team leaders who have access, but managers often use dashboards in staff meetings to drive agendas, Smith says. “You start at the top and walk through a discussion first of all the metrics in red. Are we worried that it’s red? What can we do about it?”
One of the great benefits of dashboards is that they force managers to focus on problem areas first. And the WCB dashboards allow them to zero in and work on underlying problems – not just that a group’s number of unresolved claims is creeping up but that these two agents have an inordinate number of unresolved claims.
WCB Alberta began experimenting this fall with giving front-line workers access to team-level dashboards. “It will let them see their performance and where they are relative to others in the team, and where their team is relative to other teams,” Smith says. “They’ll be able to take action before the supervisor comes around to speak to them.”
Dashboards have also eliminated a lot of often-ignored email traffic at WCB Alberta. In the past, super users tasked with producing reports from Cognos queries would periodically mass e-mail data. Today those reports are all posted to dashboards and available centrally.
SaskTel, the phone company in Saskatchewan, implemented a business intelligence (BI) system from Hyperion Solutions Corp. in the mid-1990s. Around 2000, it started producing simple dashboards based on templates of its own design and construction – intranet Web pages with pull-down lists for selecting parameters to produce pre-formatted reports. “It’s sort of like self-serve business intelligence,” says Gary Calcutta, a business analyst in SaskTel’s BI group.
The company started with a couple, but in the last four years, dashboarding has taken off. Calcutta estimates there are 50 now, each able to generate 10 to 50 semi-customized reports with a few mouse clicks. Virtually every department uses them – sales to track customer churn, for example, operations to track network loads, security to detect fraud and so on.
“A dashboard is just a way of delivering business intelligence,” Calcutta says. “But because it’s even easier to use [than the Hyperion query system], you can target employees with a lower level of technology sophistication and who don’t know as much about the data. In fact, they don’t need to know anything except how to point and click.”
Dashboard alerts have had a significant impact on decision making up and down the management hierarchy. If the dashboard used by operations staff shows that network traffic at a central office has passed 90 percent of capacity, for example, the manager knows immediately and can take steps to install or activate additional infrastructure. Churn reports in the sales department can alert reps within 24 hours that a customer has jumped ship so they can begin trying to win him back.
Quite aside from the benefit business departments and users are getting, deploying dashboards has radically simplified management and support of BI. SaskTel has reduced BI support staff by more than half in the last few years, while increasing the number of users that have access to BI information by 25 to 30 percent. Three people in Calcutta’s group now support 2,000 users.
This is possible because, as effective as dashboarding is, it’s not rocket science. “Dashboarding isn’t the difficult bit,” Bell says. “The difficult part is at the back end, with integration and building the data warehouse.”
“The real issues to decide are what to measure, how often to update, what data sources you’ll use? What are the bits of data that would be actionable, that prompt you to do something – and not just give you a warm and fuzzy feeling because the light is showing green all day?”
At SaskTel, as at WCB Alberta, business analysts in the departments and divisions – Calcutta refers to them as “power users” – do the relatively hard work of generating reports from the BI system. “All we do,” he says, “is build a pretty GUI front-end, bung that report into the front-end and make sure it works in a quick and easy way. We can pound out a dashboard in half a day.”
It begs the question: why aren’t BI-enabled companies making even more and better use of dashboards? Why is WCB Alberta only now letting front-line workers see dashboards, for example? They’re cheap and relatively easy to implement and deploy once the data warehouse and BI systems are in place. And they deliver solid benefits. It should be a no-brainer.