As more and more organisations invest in IoT and look to either buy or build IoT platforms to handle their data and manage their devices and services, it is becoming more complex to identify the right approach from the slew of IoT platforms, now numbering in the hundreds, that are on the market. In addition, as IoT services scale up from connecting a small and easy-to-manage to a large number of devices, the decision between buying or building your own platform becomes less clear cut. Dima Tokar, co-founder and head of research at IoT analyst firm MachNation, and Bill Dykas, IoT Platform product manager at Telit, an IoT solutions provider, tell IoT Now managing editor, George Malim, how businesses can assess both the quantitative and qualitative merits of IoT platforms to determine the best approach and the most appropriate product selection
George Malim: What are some of the top challenges plaguing IoT implementation today?
Dima Tokar: From a business standpoint there are ultimately three key challenges keeping IoT implementations from going forward and going into production. These are: figuring out the return on investment (ROI)-positive solutions, getting executive buy-in and generating business value from the data collected.
It’s easy enough to come up with visionary ideas but the tricky part is identifying which of those solutions make sense and drive revenue, increase profit or build better relationships with customers.
Once you’ve done that and figured out what you’re going to be doing, deploying an IoT platform is quite a transformational thing. It impacts many parts of the business. These are obvious changes in IT and engineering, but IoT solutions also affect marketing, sales and finance so all executives need to have an understanding of what’s going to happen and be all-in on that. Finally, once you are well on your journey and have collected data, knowing what to do with it is not always easy. It typically requires help from people who understand data science and analytics to get to the best outcomes. Not all organisations have teams of data scientists available at their disposal, so this is potentially a lot more challenging and costly than people expect.
Bill Dykas: What we’ve observed from most sophisticated customers is that it’s not always about only IoT, it’s about integrating IoT with the existing business process. Changing these can be fraught with complexities and, in addition, to ensuring the executive buy-in that Dima mentioned, IoT solutions involve distribution networks, have to address where a product is located and work with resellers and dealers. Ensuring support for predictive maintenance is vital but it’s not necessarily the only technical challenge.
GM: What IoT platform specific dilemmas exist?
DT: There are hundreds of self-described IoT platforms and MachNation has identified that there are also many different types of IoT platforms. This makes it difficult to create an apples-to-apples comparison to find a platform that checks all the boxes across a set of business and technical requirements. This is a difficult place for enterprises to start from and most will seek expert advice before moving forward.
Compounding the challenge, once you’ve found the platform that meets your needs, you’ll then find someone internally who asks, “Why are we buying a platform when we could build one?” This is a common scenario in many IoT platform selection processes today.
BD: Whether you build or buy, there’s a scale and scope question to be considered. If you’re doing something with tens of devices, it’s quite straightforward but what happens if tens become thousands of devices?
In addition, you’re rarely attaching to one type or piece of equipment. We’re supporting companies that have to attach to the likes of Siemens, Rockwell, Mitsubishi and a variety of other devices that don’t have standard protocols. Further to this, when you build a platform it takes focus away from IoT innovation because your efforts are directed towards IoT platform development, not IoT service development.
GM: What are some of key considerations for buying vs. building an IoT platform?
DT: The build vs buy decision is a common conversation we have with enterprises. We try to break apart quantitative versus qualitative metrics to address this. Quantitative metrics are clearer, we suggest enterprises consider the cost of hardware, infrastructure and software-as-aservice (SaaS) and, if they’re building their own platform, the cost of labour. Build your own will require a significant investment in labour whereas on the buy side you’ll typically be dealing with SaaS pricing. It’s important to put some numbers down to understand each alternative.
On the qualitative side, it’s more complex because it’s harder to quantify time-to-market, the level of customisation required and your organisational core competencies. For instance, is your organisation in the business of best-in-class software platform development, compliance or security?
GM: How can an off-the-shelf IoT platform help accelerate deployment?
BD: Building a platform requires technical research, building application programming interfaces (APIs), and implementing device specific protocols every time you want to connect a device to the platform. Off-the-shelf platforms built for large-scale deployments already come with APIs and device protocols. My experience with some of our customers is that the process of getting started is much faster with an off-the shelf platform, allowing them to spend much more time on the development of their services.
DT: Another benefit of an off-the-shelf platform is that it allows you to shop around. This is truly important in a world of IoT platforms, where marketing is not the reality. All platforms are not created equal so going with an off-the-shelf approach enables you to engage with a couple of vendors and to test drive each, so to speak. If you build your own platform, you’ll be 12-18 months down the line before you discover what you may have missed at the start.
GM: What are the most important capabilities organisations should look for in an IoT platform?
BD: At a base level, key capabilities like device management and data management are needed but you also need to think of the edge. When you start doing that the device management might involve firmware updates but also edge model updates which could deal with event logic or event streaming processes and, in some cases, artificial intelligence or machine learning moving to the edge.
Data management is also critical because, once you’ve onboarded a device you can’t just fire and forget. The device is now part of a critical business process. You therefore have to understand what connectivity is used, what device is used and how to normalise the data streams so useful data can be digested.
GM: What resources are available to help in the IoT platform decision process?
BD: We have worked with MachNation to create a very practical IoT Platform Build Versus Buy Decision Guide. This guide includes a free, easyto- use total cost of ownership (TCO) calculator. Telit also provides professional services to help company’s scope out their requirements and aid in the decision process.
DT: When it comes to platforms there are lots out there, but most tend to excel in a niche area. There are a lot fewer that can support your device management, your data management and your edge needs. Look to those three fundamentals to create a shortlist. We profile platform vendors in our ScoreCards but keep in mind that hands-on testing is key. Check out our IoT platform test environment, MIT-E, where we test drive these platforms ourselves so enterprises can pick up from there.
IoT Platform Build Versus Buy Decision Guide: http://info.telit.com/iot-build-vs-buy