Active Notification: Let your systems do the talking

The most frustrating aspect of the five-minute horizon is that all the data you need to make timely decisions is literally at your fingertips in your ERP and MES systems. Unfortunately, it is nearly impossible to find the exact right information at exactly the right time. Supervisors and managers are far too busy running a plant to spend hours behind a computer looking for incremental improvements, yet the incremental improvements often mean the difference between having a good and a bad day.

I think that the incremental improvements should come looking for you. Since all this information is already in the system, your software just needs to be smart enough to identify an issue and then have the capability to notify you. Ideally, active notification allows you to shift from fire fighting into opportunity spotting because you can address an issue before it becomes a crisis.

Fortunately, there is an inexpensive and established technology to put active notification in place: email. Email is the universal currency for intersystem communication. Most software can generate or be easily modified to generate email. Email is ideal because everyone can receive email at their desktop or even to their alpha-pagers and cell phones. Getting active notification working for your plant can be as simple as a clever use of system generated email.

For example 

One of my clients is using notification to improve the performance of their MES system. I added a simple program, called the watchdog, that monitors the MES database for critical performance factors like system failures, downtime, and out of limits conditions. When these conditions occur, the program sends email to alpha-pagers worn by floor supervisors and engineers. The text of the message tells the people on pager duty exactly what caused the problem and where to look for it. For example: a page might read "ALERT: Machines A10, B6 Down. Machines A1, E12 Out of Limits." The alert messages are specific enough that the right people know what to do: engineers are not chasing production problems and production coordinators do not have to find an engineer for every process problem. In addition, when the problems are solved, the watchdog issues an "all-clear" notice so everyone knows the system has stabilized.

In another case, my client had problems with the program that moved information from the real-time system to their historical system. Occasionally, incomplete data in the real-time system creates holes in the historical data. This recurring problem, although easily corrected, reduced management acceptance of the system due to perceived data integrity issues. Since we could not address the actual cause of the problem, I added a simple check step in the data generation process. This check step looks for holes and then sends an email problem report to the system administrator. The email contains specific details about the exact date and machine that needs attention. When management is confident about data reliability and integrity, acceptance of the system at all levels improves.

Getting the message right 

Putting active notification in place requires some planning and forethought. In the past, I have been guilty of over zealous use of the technology and inadvertently "spammed" system users with hundreds of redundant or cryptic alerts. With email this is just a nuisance, but commercial alpha-pagers start charging by the page if you cross a threshold of about 200 pages per month. Here are some lessons learned by hard experience.

Content is critical. 

• You do not want a computer to cry "wolf." Instead, it should cry "rabid gray furred canine approaching work cell 15, emergency medical alerts initiated, HR and lawyers notified by email." 
• Your messages should have intelligent subject lines. The subject line is critical because that is all users see on most email systems and pagers. If the subject line is not clear, your message may be overlooked. 
• Time stamping lets users know how timely a message is right away. Users will use time stamps to determine if they are already too late or when messages are being delayed.

Manage your messages 

• Do not rely on just alpha-pagers for active notification or get locked into just one way of delivering a message. Some messages are not important enough to send a page and should just be sent as regular email. Email is cheaper and easier to filter than notices to pagers. In addition, some people who need active notification may not carry a pager. Do not forget to be creative: you can send even email to a salesman's or manager's digital cell phone to achieve nationwide paging. 
• Use email groups for notification instead of hard-coding notices to specific people. Groups are easy to manage and allow you to easily arrange vacation coverage. 
• Incorporate an escalation scheme so issues that are not resolved generate notifications up the corporate ladder. This improves issue response time and keeps managers in the loop. 
• Sending an "all clear" notification helps coordinate the people responding to the messages. If you are also using escalation, make certain that send the "all clear" to everyone who received a notification. 
• Log messages in a separate file or email account if possible. A notification log can be invaluable when tracking down repeat problems, resolving communication delays, or just troubleshooting system performance. I can learn more about actual system performance, including user responsiveness, from the email logs than any other factor by comparing alert sent to all-clear timing.

Stay in control 

• Pool your notices when possible. Pooling messages means that you combine notices from multiple sources into a single message. For example, you may send minor out of tolerance messages only once an hour so many can be collected together but a severe warning would trigger an immediate notification. Pooling allows you to reduce notification overload, improves system performance, and makes messages more usable because they are partially filtered.
• Local area paging is more reliable and faster than commercial services for inside the plant notification. Current systems are inexpensive, have a good range, and can be deployed for a wide variety of uses beyond active notification. Visit www.h-consulting.com for links to pager manufacturers. 
• Make the system easy to turn off or include sleep cycles for close of business hours. Once of my first active notification programs with escalation often detected system failures at 2 AM. The escalation feature made the IT managers sort of grouchy the next morning so I made it easier to the people on call to suspend further notifications until they could brew coffee and start working on the problem. 
• Do not forget to install heartbeats or checks on your notification engines. It is easy to become dependant on active notification so you need to take steps to make sure the system keeps running.

Reach out and touch someone

Even though the technology for active notification is within your grasp, do not overlook the effort involved. To make these ideas work, you must make good decisions about all aspects of the notification. Think carefully about what to send, to whom to send it to, what are the delivery media, and when to send it? You do not need to jump in with hundreds of messages and a fleet of alpha-pagers: start small to prove the idea to yourself and your employees. After all, if you can save the right five minutes, you can have a big impact.

Active Notification Sidebar1: 

The cost of sending your message

Active Notification Sidebar2: 

Links to Pager Resources

• Local Area Paging: JTech Communications
• Cell Phone Paging: Spring PCS
• Dedicated Paging: PageNet