The Internet of Things (“IoT”) has created a buzz over the last couple of years. Mainstream media, such as the New York Times and the Wall Street Journal, are reporting on it. Both established companies, such as Atmel, Broadcom, Intel, Microsoft and Google, as well as start-ups have introduced a number of new IoT-related products. Industrial companies such as GE and Bosch were among the early adopters. Others are beginning to formulate their IoT strategy. M&A is also spicing up the market, such as Samsung’s $100 million acquisition of 2 year old start-up SmartThings.
What is the Internet of Things?
IoT is an evolution of a technology that in its previous version was called machine-to-machine. In simple terms, IoT consists of sensors and actuators that are connected to each other and the Internet using a network. Sensors include light, temperature, humidity and pressure meters. Actuators consists of motors, software triggers or anything else that triggers an action.
Today’s Internet of Things, however, is far more powerful and useful than its predecessor technology. There is a lot more “intelligence” or processing power in the system. The sensors are attached to high speed processors, making them “smart sensors.” Such processors are getting faster. They have a smaller footprint and can be installed easily. They require lower energy and thus last longer. And they are far cheaper, costing just a few dollars. Data from such sensors are then transferred using a Wi-Fi or cellular network to a cloud or another collection and analysis system. Analyzed data can not only be used for executive-level decision making but also to trigger other devices or Internet-related actions within microseconds.
The real power of IoT is felt when these sensors and actuators are networked with information on the Internet, such as weather, city or company data. Alternatively, they may networked in such a way that information is displayed on a smartphone or the devices and processes can be controlled using the smartphone. In addition, since diverse data streams can be meshed and analyzed, such as equipment performance with ambient temperature and dust levels, we can now see patterns in the data using machine learning which were difficult to see earlier. And we can use such patterns to make recommendations or predictions, such as predicting machine failure.
Let’s consider a simple example of a sprinkler system in a botanical garden. Sprinkler systems spray a known quantity of water uniformly. Consider attaching a sensor and actuator to each sprinkler. You can now control the amount of water you are spraying on a Eucalyptus tree versus a cactus. Better yet, why sprinkle any water if it is going to rain? The sprinkler system checks Yahoo weather and does not activate if it is going to rain. Let’s also add soil monitors. Each sensor can now send data back to a central location regarding soil conditions and how the plant is growing. A botanist analyzes this data and can determine if a plant needs some special attention. You have now gotten the expert handling of a botanist, regardless of where he or she sits, for thousands of plants located in a city. Many parts of this system are already in place in parks in Barcelona, which in 2014 was recognized as the #1 “smart city” in the world.
Why IoT now?
Several factors are accelerating the growth of IoT.
- New sensors are being introduced to take lots of different measurements. They are becoming faster, smaller, more accurate, more energy efficient and cheaper.
- Processors are also becoming faster, smaller, more energy efficient and cheaper.
- Network connectivity using Wi-Fi and cellular is now ubiquitous.
- With smartphones, which are common, information can be viewed and actions can be taken by the device in your pocket – significantly enhancing access and usefulness.
- Data storage is inexpensive and getting cheaper. For example, Google now charges just $0.01/GB/month.
- Pattern recognition through machine learning is becoming more accurate, faster and cheaper.
- The open source movement is accelerating hardware and software innovation without the need for technology to be controlled by any given consortium.
Who is using IoT today?
A wide variety of companies are using IoT. The following are three applications are indicative of industrial and service settings.
GE Aircraft Engines1
For airlines, ‘time on wing,’ or the amount of time the plane is flying versus sitting in the maintenance shop, is a key performance indicator. Instead of doing maintenance every certain number of hours, airlines using GE’s engines can predict when maintenance is actually needed – thus increasing time on wing. Engine maintenance – typically an engine wash – is needed when particulate matter (dust) builds up. As dust increases, the operating temperature of the engine increases. Sensors monitor the temperature and send data to GE continuously while the plane is flying and also in “batch” when the plane is on the ground. By looking at the trend of temperature increase versus normal operating temperature, airlines can predict when a plane is due for maintenance. Maintenance schedules differ not only by airline but also by aircraft, based on departure and arrival cities. Planes operating in cities with high levels of air pollution require more frequent maintenance.
Certain swimming pool equipment manufacturers are moving from a one-time sale to having a recurring relationship with their customers. Currently, pool temperature and lighting can be controlled by a smartphone. Innovative manufacturers are now taking over the harder part of making sure the mix of chemicals in the pool is in the right balance. Sensors monitor the chemical composition and send data back to a maintenance center. The manufacturers’ experts ensure appropriate mix by controlling the dispensation remotely. The manufacturer not only benefits from monthly recurring revenue but also from deeper ties to customers.
John Deere – connected farms
Farm equipment manufacturer John Deere installs sensors in its equipment which send data back to the farmer as the equipment is running. They have also successfully created a service, called myJohnDeere.com2, where farmers can manage equipment data, production data and farm operations. On their tablet, they can see near real-time data from their machine including average yield, total yield, average moisture and seeding variety. Since soil can vary across a single field, farmers can get a “prescription” of what seeds they should plant in which part of the field based on soil conditions to maximize yield. Recommendations are based on statistical analysis of data – in effect, farmers are sharing best practices with each other. One farmer is getting a recommendation based on experiences of other farmers with similar soil and environmental conditions. But more importantly, farmers can communicate with each other, get advice, give advice and form a community. The relationship increases the likelihood that the next equipment sale is from John Deere as well.
Such services by equipment manufacturers are going to become a competitive differentiator. Per IDC, by 2017, 70% of discrete manufacturers will offer connected products3. In fact, manufacturers who do not offer such services are going to be at a competitive disadvantage. We mentioned earlier how John Deere is offering a service to farmers. Monsanto, an agriculture company specializing in seeds, acquired The Climate Corporation4 for $930 million in October 2013 to offer similar services5.
Manufacturers can incorporate IoT technologies with marginal incremental costs. The upfront cost of sensors, actuators and processors is small. Typically, the largest cost component is that of the cloud computing network. These costs have rapidly declined. Moreover, the tech industry has also adopted monthly subscription models and is offering “infrastructure as a service” and “software as a service.” Tech vendors, such as Amazon AWS and Microsoft Azure, bear the upfront expense, and more importantly, take over the responsibility of managing the network, upgrading, scaling and providing security. Their customers benefit from a “pay as you go” model. For the manufacturer, the monthly costs can be more than offset by the monthly revenue. Adopting IoT is a profit enhancing proposition.
Security is sometimes expressed as a concern in IoT. Security can be broken down into two components – security of information moving from sensors to the cloud, and security of information in the cloud. The security of information in the cloud provided by reputed companies can be made about the same as information security provided by banks. Security of information moving from sensors to the cloud does exist currently. Current technology is being improved upon by network security companies and IoT companies. There are some proprietary security solutions offered. We expect to see considerable improvements over the next 24 months.
Advantages of using IoT in systems and processes
For equipment manufacturers, there are several advantages of integrating IoT into their machines.
- Get access to data even when the machine is at the customer site. Once sensors send data into a cloud system, manufacturer can get access to a customer’s machine performance data even if the equipment is at the customer’s factory. In fact, a New Jersey manufacturer can get access to data even if the equipment is being used in Enugu, Nigeria. Such access to data overcomes a major hurdle when it comes to providing customer satisfaction and continuous improvement.
- Monitor performance and offer suggestions to improve performance. Once you can access data, you can monitor performance. Even better, you can provide your customers advice regarding fine tuning machines and optimal machine performance given by your best engineers regardless of where they are located.
- Predict failure or recommend a maintenance schedule. Predict parts failure in time to take corrective action by observing the deviation of the performance from the normal trend line. Alternatively, use the performance data to fine tune the maintenance schedule.
- Monetize service level agreements. Many manufacturers already monetize maintenance through service agreements. With performance data, manufacturers can now be confident about machine performance and monetize this confidence by guaranteeing service level (such as uptime), which may be a competitive advantage.
- Use data to improve parts performance. Manufacturers routinely use statistical analysis of parts performance to improve the quality of various components. With real time monitoring, such analysis can be done much faster. It can also be more accurate with sample sizes being larger, since data can be measured frequently and stored easily.
- Provide recommendations. Form a community and lasting customer relationships. With connectivity, manufacturers are encouraging communities around their products. Customers interact with each other, answer questions and share experiences. Tighter relationships means that it is less likely the customer will go to a competitor for the next purchase.
Exploring new business models
Performance monitoring can lead to incremental service revenue, higher customer satisfaction, less downtime, and eventual improvement in quality. The greatest impact of IoT, however, could be in changing the revenue model to a usage-based pricing model – i.e. selling a product as a service rather than just a product.
Consider the typical pricing dilemma a manufacturer faces. Let’s consider a large air compressor manufacturer. Per microeconomics, a company maximizes profits as long as marginal revenue from selling a unit is greater than the marginal cost of the unit. Airlines, for example, routinely sell similar seats at different prices to maximize profits. Manufacturers often try to do the same with a system of different features in machines, or rebates and discounts. However, there may be a more elegant way of measuring value derived by a customer and pricing accordingly.
Let’s consider a simple example of a compressor manufacturer whose marginal cost of manufacturing and selling a compressor unit is $30,000. Furthermore, let’s say 20 large customers derive $100,000 or more of value from a unit and 30 smaller customers derive $50,000 of value. Profit from only selling to large customers would be $70,000/unit * 20 units = $1.4 million. If the manufacturer wants to sell to everyone, then the price per unit has to be decreased to $50,000. Overall, profits then become $20,000/unit * 50 units = $1.0 million. Clearly, it is in the manufacturer’s interests to keep prices higher and sell only to the large customer. However, money is left on the table.
Now consider a scenario where pricing was based on usage, say quantity of air compressed, time the machine is run or some other relevant measure. Prices of this unit could be set in such a way where the revenue from the large customers is still $100,000 and from the small customers is $50,000. This time, though, the manufacturer can have all 50 customers. Profits would be large customers ($70,000/unit * 20 units = $1.4 million), PLUS smaller customers ($20,000/unit * 30 units = $600,000), or an aggregate of $2.0 million.
Pricing the product as a service leads to a recurring, and therefore more predictable, sources of revenue. Certain technology companies have moved or are in the process of shifting to usage-based pricing. Salesforce, Amazon AWS, Adobe and Microsoft are prime examples. For instance, Amazon AWS’s server farms are used not only by large corporations like Netflix but also by start-ups who would rather outsource specialized skills of cloud management. Amazon AWS’s usage-based pricing makes it affordable for large and small corporations.
The Internet of Things is a powerful new technology for both manufacturing and service businesses. Its time has come. We expect to see rapid adoption ahead. For manufacturers there are tremendous benefits of using this technology to improve customer satisfaction and performance quality. IoT can also be used to gain incremental revenue streams such as subscription revenue for reports, recommendations and community resources. The most significant impact of IoT, however, may be in changing the revenue model to “pay as you go”, which would not only make revenues recurring, and therefore more predictable, but also maximize profitability.
1. https://www.gesoftware.com/sites/default/files/seizing_competitive_edge_with_the_industrial_internet_by_josh_greenbaum.pdfhttps://www.gesoftware.com/sites/default/files/seizing_competitive_edge_with_the_industrial_internet_by_josh_greenbaum.pdf, pg. 7
3. Heather Ashton, IDC, “Smart, Connected Products in Manufacturing”, June 2015.