Fort quickly determined that the company needed to monitor — proactively and persistently — all the services and features that it offered customers. By doing so, it could get right on top of any outages and, more important, "get out in front of those so that they don't even occur," he says.

Fort assessed his own organization and realized that he didn't have the resources to check every kiosk touchscreen. He wanted to monitor the uptime of all critical components, which included in-store technology and all the backbone components from IT, such as key servers, point-of-sale (POS) servers, and the registers right on down to the kiosks themselves.

A critical element of the monitoring strategy is its business focus. When a problem occurs, that problem is immediately tied to a business process. If a key network component fails, the IT department immediately becomes aware of which business processes are affected, such as credit processing or kiosks. Then it notifies the stores involved about the issue and takes corrective actions to prevent any effect on customers. A kiosk's failure at the same time as a credit failure is deemed to be less critical than the credit failure, so IT prioritizes its responses and resources appropriately.

As the IT team embarked on its service monitoring effort, Fort realized that he needed to determine the most critical components of the Megastore infrastructure. He drew a series of concentric circles to help him organize his thoughts around these components. The innermost circle included core critical servers, network, and voice over IP (VoIP) components. The next circle contained the POS servers, which would bring all sales at a store to a screeching halt if they went down. This circle also contained credit services, which depend on the network to get customers' credit cards validated. (A store could continue to accept credit cards if the network was down, of course, but it would assume greater risk.) The next circle out included the registers, the circle beyond that one included the kiosks . . . you get the picture.

The failure of a single register or a single kiosk is less important than the failure of the whole system, but it still leaves a negative impression with customers. Typically, one register would fail and wouldn't be reported to the help desk. When the help desk finally got a call, the report was that multiple registers were down and associated monitors weren't functioning.

Overarching everything in the retail space is the Payment Card Industry (PCI) initiative, which includes mandates from card companies such as Visa MasterCard on how to protect customer data. Fort knew that he had to monitor the credit process very tightly, making sure that IT was aware of any outages or anomalies so that the company was PCI-compliant.

The team implemented a commercial monitoring application that could answer questions like these:

The software also monitors application, database, and operating systems. In all, the team monitors more than 2,000 components. Any number of components make up a business process, and a failure in any physical component could affect one or more business processes. The team also can monitor individual business transactions such as credit card purchases, which helps Virgin remain compliant with PCI.

Fort is quick to point out that although the company is monitoring these components at a technical level, it receives information about the impact of events at a business level. At the technical level, you can imagine that the networks at each of the stores has a key router and switch to connect to the network cloud; behind those routers and switches are the POS servers and the individual registers, kiosks, media players, and digital signs.

At the technical level, someone may see the problem as being a router failure. At a business level, however, someone may see the problem as being an inability to process credit or sales. When a router goes down, the monitoring team not only sees red flags, but also knows which series of business processes is being affected, which helps the team determine the level of priority.

When IT knows the business impact, it can make the appropriate decision. A problem that might affect $100,000 worth of sales that day (if the registers go down, for example) gets top priority; it needs to be fixed no matter how much the repair costs. If, on the other hand, the issue might affect $20 worth of sales (a digital sign isn't working, for example), expediting a fix for it isn't worth a huge expense.

Automated monitoring meant a change for the network engineering team, which was used to a firefighting mentality. If something went wrong, team members got involved. It was a bit of a shift to look at a dashboard that lit up and a sent e-mail to report that something was amiss. The alert e-mail didn't necessarily say that the system had crashed — it might simply say that a database table increased by more than 10 percent in a single day — but it could predict a problem that needed investigation. The idea was to prevent the fire from coming.

This change affected many engineers' personal measures of success, because no one would necessarily know that he'd averted a potential disaster. In fact, several engineers left the organization to move to areas where they could start building new networks.

Fort attributes this situation to a maturity curve, saying that some personality types may find it unrewarding to monitor smoke instead of fighting raging fires. He imagines it this way: On the left side of a firefighting icon are the data-center-minded people, who are very process-oriented types who believe in governance and don't particularly care for fires. On the other side of the icon are the people who love fires, are always off to something new, and love to push boundaries. Fort believes that every organization going through service management will eventually move from one side of the icon to the other.