View the original post.
Customers don’t buy the Internet of Things. Customers buy a solution to a specific problem where Internet of Things (IoT) components are part of that solution. This could be a well known problem where the customer is actively seeking a solution or it could be longstanding problem that a customer didn’t realize could be fixed. An IoT solution typical requires an ecosystem of partners that spans from sensors to applications.
The Internet of Things (IoT) landscape is littered with startups. The challenge for any company in the Internet of Things (IoT) market is to define what problem it can solve and how much of the overall problem it can reasonably solve. . While it might be necessary for a single vendor to provide and service an end to end solution, the vendor that supplies it frequently can’t develop all of the components by themselves. Most IoT solutions require an ecosystem of partners.
It was with these challenges in mind that I visited Helium headquarters for a technology analyst roundtable. I had the opportunity to meet with the company’s CEO Amir Haleem, its Vice President Strategic Sales Ned Taylor and Rob Chandhok the company’s President and COO. Helium is an integrated platform for enterprise and industry to monitor, learn and capture insights from the existing physical things in their environment. It offers a solution that ranges from wireless connected sensors to complex event processing. Helium states its mission as bringing data from the physical world into the business systems that create value for its customers. It wants to be the go-to source for things that are IOT requirements but offer limited differentiation at the solutions layer.
It starts by making things smart by combining sensors, connectivity and software. The main difference between this and the M2M concept of the past is the focus on smart versus connected. Of course the goal of M2M was to deliver intelligence by connecting items. But in reality, M2M largely focused on device connectivity and reporting data from those devices that resembled historical business intelligence versus real-time insight. Today, the market is focused more on near real-time analysis, insight and action rather than reporting. Chandhok said “The idea is to paint the world in sensing but to have a way to quickly make sense of that data.” There are several factors that make this easier today than it was in the past, but mostly it comes down to cost and availability. Almost every function that makes smart sensing networks happen is significantly cheaper than it was five years ago.
Chandhok noted newer IoT companies use modern software, modern protocols (for example a call Twillio or iMessage or other APIs), and run in the cloud. The latest solutions can be easily accessed on mobile devices (tablets, smartwatches, glasses) and via the Web. These software-driven startups can rapidly iterate, which makes it difficult for IoT companies with proprietary protocols to compete. It’s also easier for a programmer to modify this software without breaking the system. In speaking with Chandhok, I found some valuable lessons for both IoT startups and companies alike. Companies should:
- Find a specific problem to solve that is quantifiable. While there’s value in being a platform company, it’s important to have a specific use case that a customer can understand and purchase today. In the case of Helium, one of its anchor products is monitoring refrigeration units. The system constantly monitors temperature to ensure that medicine, food, and other temperature sensitive items are always stored at the proper temperature and logs are maintained for compliance. If there is an issue, the system can contact designated personnel. The value of maintaining the right temperature is evident and loss due to equipment failure is quantifiable. In addition to specific use case, Helium can also make money buy selling its platform to startups or other IoT vendors. For example, vendors that want to build baby monitors and hydroponic systems don’t want to construct the entire system. These companies want to focus on their core diffentiation (which are typically at the smart sensor and the application solutions layer) and purchase the network and cloud functions from companies like Helium.
- Define an edge processing strategy. Originally, many firms planned to instrument devices with sensors, gather that device data, and send everything back to the cloud for processing. The problem with this strategy is that working in the physical world requires you to perform at least a certain amount of processing at the edge for various reasons. You may need to process and act on data immediately or you may not want to send large volumes of data over the network (e.g. terabits of gas turbine data). As a result, companies need to build a strategy for when to analyze data at the edge and when it needs to be returned to headquarters.
- Look for solutions that can do correlation and analytics. IoT needs a certain amount of intelligence to be valuable. In addition to defining where data is processed, companies also must select equipment with edge processing in mind. For example, taking a temperature reading and sending it back to HQ every five minutes will drain the battery but remembering the temperature and send every it 12 hours may be too long to wait if there’s an issue. You need to add logic and processing at the edge that might say read, remember but don’t send unless it’s abnormal. The best systems will offer machine learning that looks for patterns and informs the user of what they need to know. One of the ways that IoT is different today is that the user experience, while not seamless, is much easier than it had been in the past. For example, an employee may have had access to equipment alerts before but the volume was overwhelming and it wasn’t easy to understand alerts or customize when you could be alerted. Today’s IoT solutions can make data access simpler, easier to consume and act upon.
- Make sure the solutions focuses on preserving battery. Hardwiring an electrical line to each sensor is cost prohibitive but battery-powered sensing is useless if you have to replace batteries every 6 months. This means the end sensing device must have software that is intelligent about how it processes data and delivers data to optimize battery life. Chandhok notes building a robust IoT solution requires designing the right flexibility and analytics into the software. This will allow customers to get multiple use cases from one from one sensor type and will allow even more use cases as the number of sensors scale.
- Develop a broader mindset for IoT security. Encryption is the first line of defense for IoT but companies should also evaluate how security best practices in other areas such as identity and authentication will offer additional safeguards. A company should know if the network, the hardware and the person accessing the data are trusted. Also given the rapidly changing threat landscape, IoT solutions needs to be upgradable in the field to prevent future threats.
Today many sensors are special purpose. In the very near term, we’ll live in a world where it’s cheap enough to combine many sensors into one unit. In this way, a company could have richer environmental sensing that includes temperature, humidly, ambient light and motion. Companies should be evaluating how this new information can be integrated into existing applications or if processes need to be rewritten with new data in mind. As companies look to mobile-enable existing applications and purchase Software as a Service (SaaS) applications, there’s also an opportunity to consider how these new IoT data streams at value or changes application requirements. While the term may sound trite, I do believe IoT will create a paradigm shift, making certain businesses smarter and burying others that fail to adapt.