If you do one thing well, you are an expert. If you do two things well you are a prodigy. If you do three things well, you are amazing, but if you do everything well, you are a mountebank. Mike Brown is a mountebank. He does everything he touches well. "I'm hearing a recurring theme that we have to pay more attention to the profitability of automation. But my cynical bone keeps wondering if this is just marketing," he said. The challenge, he went on, is to measure the bottom line ROI of automation initiatives. Users and suppliers are beginning to rethink the role of automation and demanding to know more about how automation contributes to plant profitability and the core business drivers. We've gone through several phases: in the 70s, we invested in automation technology for technologies' sake. Then we moved to a network centric approach, with open systems. Then we moved to an application centric model with MPC, advanced control, and other approaches that are standards centric as well as open. Now, we're being told that our systems need to be business centric. "I confess to being confused," Brown said, "and I don't really know how to do that." Brown and I discussed this same thing yesterday afternoon, and we came up with an image of a data stream being viewed through a prism, so that everybody gets the view of the same data they need for their jobs. "Sometimes we try to let KPIs drive the system, where the KPIs aren't relevant to the person doing the job." Linking Automation Information to Core Business Drivers Given that 70% of our processing plant assets are under some form of control, do we know what impact automation is having or not? "Data Transformation:" an alarm goes off. The operator silences acknwoedges and tries to understand the process issue. The automation engineer finds out that that sensor has generated 52 alarms in the last shift. The area has experienced 3 alarm floods and has an average of 12 pending alarms. The operations manager can understand the impact of the alarms on the operators. It is all the same piece of data, but it means different things. We can contextualize that data, so that it is relevant to every person in the organization as you move through the enterprise. This is very important for the process engineer, as well as the automation engineer. Think about how that data can drive your organization as a whole. Think about the need to transform the data so that it is intrinsically useful for the process engineer and the plant manager. We are struggling today to get very experienced very capable people. We have enough technology to skin the cat, but we don't have the people to use the technology always. People are buying automation on ROI and profitability, but the highest values to consumers of automation technologies are product expertise and industry expertise. People are critical. Where are you going to increase outsourcing? performance management and optimization solutions...the delivery of automation that ties to the core business drivers. How can you use this technology to make your people more effective in getting ROI out of the automation initiatives you do? Control monitoring technology has the ability to improve resource efficiency by up to 90%. Now how does that directly translate to ROI? Again, it is about data transformation. But where is the innovation coming from? We must now take advantage of these enabling systems to tackle the real challenges in our plants and transform work processes. For example, 5% of refiners say that they can't run their plants witout APC, and 70% of refiners feel they can't be competitive without APC, but over 50% of refiners say their APC applications experience significant degradation. The change comes when people adopted a new standard methodology for sustaining APC...the Tai-Ji APC has made it possible for APC to become inexpensive enough to be used. So we need to bring this back to the humble PID controller. Too many people have a tuning package that is sitting on the shelf. We're going to change the game and re-define how you are going to maintain your assets by making automated testing and tuning actually work. And this task is going to be easy enough that it can be launched by an operator, not an automation engineer.