The Internet of Things (IoT) is becoming a hot-button issue for the C-suite. Many senior leaders are concerned whether implementing IoT-based technology and processes is worth the time, money, and effort involved. Others worry they risk being left behind and losing their customers to competitors if they avoid it altogether.
In fact, Gartner found that 64% of large enterprises plan to implement a Big Data projects. Sounds promising, right? Not really – considering 85% of them will be unsuccessful.
All of this data needs to go somewhere – but where ?
During his Webcast “Operationalizing IoT Data for Predictive Analytics, Dave Roberts, OSIsoft fellow specializing in cities and industrial clusters, observed that complexity in the IoT is attributed to technologies that support a wide variety of standards that seem to be emerging in the IoT space. “There’s AllJoyn, Thread, IEEE, Open Internet Consortium (OIC), Industrial Internet Consortium (IIC), among others that are promoting different standards on how sensors and assets will communicate with gateways and routers,” he explained.
As a result, businesses are spending 50%–80% of their development time just prepping their data – collecting, cleansing, shaping, backfilling, and timestamping all of this information. Some are even going as far as creating a new job title, “data engineer,” responsible for getting this data together and shaping it so data scientists can accurately answer critical business questions with this information.
10 things you should consider when operationalizing the Internet of Things
- The value of the IoT is not technology. The real value lies in the creation of new value propositions and potential revenue streams. The key is taking this technology and move them towards new business models and services that will help realize them. According to leading analysts and thought leaders, the growth potential is significant. IDC believes the IoT market will hit $7.1 trillion in revenue by 2020. Gartner foresees the IoT install base growing to 26 billion units by 2020. And Cisco predicts that the IoT is poised to become a $19 trillion market.
- IoT data will be more democratic than SCADA data. Historically, SCADA data has been locked away in somebody’s process control network. To access this information, update it, and revalidate it, people needed a miracle. With IoT, you can freely and quickly bring up this information when and where you need it. This one aspect is revolutionizing business models, allowing businesses to enhance their services in real time.
- Businesses outside of your industry may know something you don’t. For example, highly powerful tools developed for clickstream analysis, fraud detection, cyber security, and genome sequencing are now coming to process industries. Don’t snub other industries, thinking that you are different from them. They may have a few tricks in their pocket that you need.
- Standardization leads to repeatability. The more comparable assets are in your organization; the better your forecasts will be. Machine learning is better with more, similar data. Anything less leads to misconstrued information and inefficiency.
- IT and OT are converging – deal with it. Data engineering can take significant time and resources. However, it shouldn’t stop you from moving forward with IoT initiatives. Instrumentation and controls engineers from the world of operational technology (OT) have to bridge the gap between the analytics and IT communities.
- Sensors will not live forever. In other words, cheap sensors are not going to be 100% reliable, 100% of the time. Physical damage during normal maintenance and operation in hostile industrial environments (such as dust, vibration, water, and caustic materials) will occur. Even sensor batteries can discharge. Ultimately, all sensors fail either instantaneously or slowly degrade. Processes must be established to make sure sensors are fully operational and deliver correct data.
- Your information is as good as your sensors. Reliability of predictions is only as good as the data feeding them. If you are going to run analytics based on sensor data, you better make sure that the sensor is in good working order. At times, you even have to go as far as validate the sensor data before it is reported or analyzed to answer critical business questions.
- Data needs context. To develop a model that forecasts behavior, data scientists require context and time-series data. Otherwise, it becomes very hard to consume this information and truly see what happening now and in the future. To make the best possible decisions, people need real-time data. With pervasive monitoring, this information is captured and delivered for business intelligence analysis.
- The IoT brings a tsunami of data. IoT rollouts bring a proliferation of cheap, distributed sensors – resulting in a huge volume of data in a short amount of time. Is your infrastructure ready to support it ?
- Don’t forget what powers the IoT. Data integration and actionable information are the heart of collection and analysis of IoT data. Invest in the technologies, expertise, and processes that support integration, reporting, decision making, and action – and maintain them well.
Start making the Internet of Things a reality for your business
Watch the entire Webcast series “Internet of Things (IoT) Community Webcast Series” presented by ASUG and SAP. Topics include:
Replay available: Making the Internet of Things Real
Replay available: Transform Business Operations and Reimagine Business
Replay available: From Big Data to Smart Data (IoT)
Replay available: IoT and Industrie 4.0 – European Perspective, Examples, and Business Models
Replay available: Operationalizing IoT Data for Predictive Analytics
Replay available: IoT – From Vision to Value
June 30: Transforming Your Business with IoT – SAP Partner Perspectives
July 9: Accelerating IoT in your Organization: Introducing SAP’s IoT Development Platform