Best Practices for Selecting the Right IoT Platform
Best Practices for Selecting the Right IoT Platform
Looking for the best practices when choosing the right IoT platform for your project? Check out this post to learn more about these top three considerations.
Join the DZone community and get the full member experience.Join For Free
IoT Platform Selection: The Importance of Getting it Right
Several thousands of enterprises are currently performing their first IoT projects – whether its smart factory, connected car, or smart energy – most of them now consider using a modular IoT software platform to manage the data flows, support application development, and provide basic analytics. This promises lower cost, faster time to market, and more professional support. The IoT Platform selection process is not easy, however. There are hundreds of platforms to choose from with prominent examples being Microsoft Azure IoT, Amazon AWS IoT, PTC Thingworx, Siemens Mindsphere, GE Predix, Software AG Cumulocity IoT, or C3 IoT Platform.
“Some IoT platforms do not live up to the task”
However, several instances have now been reported where the IoT Platform did not live up to the task. In one instance, the platforms data model proved to be limited to one industry-specific scenario and could not be scaled to other use cases for the same company. In another instance, support for multi-cloud backends was not available and hindered the end-user to exchange important data sets between different platforms.
It is becoming crucial, for most companies even strategic, to ensure they are choosing the right IoT Platform that will seamlessly manage their IoT data flows. Not only for one use case but for most, if not all, of the planned use cases in the years to come.
IoT Analytics recently performed research and interviewed a number of customers of IoT platforms on how they choose their platform. These are three of the best practices that emerged from the analysis:
Best Practice 1: Pre-Select Platform Vendors Based on Qualitative Criteria
Almost all of the customers interviewed use some kind of criteria catalog for pre-selecting vendors. In one extreme instance, a customer used a catalog of 130 different dimensions.
“17 selection criteria are common across all organizations”
Most interviewed customers used a combination of 17 distinct criteria.
Eight of these can be classified as customer-specific. These are the criteria that every customer rates differently based on their own needs and perception.
The other nine criteria are platform-specific. These are criteria that are independent of the customers' needs and are directly related to the overall performance of the platform)
Customer-Specific Purchasing Criteria
Some of the customer-specific criteria are common for any technology decision-making in a large organization (e.g., Value for money, previous experience). Others are specific to the nature of the Internet of Things (e.g. support availability), because the IoT setup can be extremely complex, Strength of Partnership ecosystem – because IoT requires a lot of different systems to work together)
Platform-Specific Purchasing Criteria
Vendors can have a greater influence on the 9 Platform specific purchasing criteria (listed in Table 2) as these aspects are directly linked to the IoT platform’s capabilities. These Platform specific criteria have been ranked by customers regarding their importance in the IoT Platform selection process (based on our interviews).
The three top criteria are:
- End-to-end security (i.e. how secure the platform is including the data exchange and user and device authentification)
- Scalability (i.e. how easy it is to go from 100 to 100,000 connected endpoints)
- Usability (i.e. how user-friendly the interface is)
It is worth noting that the pre-selection process can differ on a company-by-company basis, ranging from assessing a handful of vendors to tens of vendors. For example, while some customers mentioned they only looked at two vendors, one specific customer even had 50 IoT Platform providers on their initial list of potential platforms.
Clearly, the available budget and timing playing a major factor here in the decision process as not all customers have the resources to assess 50 platforms. In any case, the aim is to narrow the options to a few chosen vendors before completing a proof of concept.
It is somewhat surprising to see that the selection criteria vary so much between different companies. For example, one company put a heavy emphasis on open source while another company totally neglected that aspect but instead focused heavily on end-to-end security.
Best Practice 2: Set Up a Real-World Proof-of-Concept (PoC) Testbed
After short-listing a number of platform vendors, most companies run real-life PoCs to test the platform performance and capabilities using different technical criteria (e.g. amount and quality of functionalities, tools, and services, usability, security, interoperability, etc.). Often, other factors like cost and platform ecosystem are taken into account in this step.
“It is essential to test the platform capabilities with real data sets”
Instead of asking vendors to present some hypothetical capabilities, IoT platform customers oftentimes set up a real-world testbed which resembles their actual product or production process and contains actual data sets from their products or factories.
One manufacturing customer revealed, “We’re essentially at the beginning of our digital journey, so we don’t want to commit to a specific platform from the start – instead we’re trying out different platforms before committing to one or more for a long time”. In essence, many customers are using PoCs to explore and “play” with the platforms to see how they work and what advantages they can get from them based on real usage scenarios while keeping their eyes open for other options.
The number of PoCs undertaken with different vendors can depend on the customer resources (e.g., financial) available. However, PoCs are a common denominator with all vendor selections. Two respondents shared an in-depth perspective of how they went through the IoT Platform selection process:
Example 1: Machinery Manufacturer
The following example presents the IoT platform selection and evaluation process of a machinery manufacturer in Germany looking to build a Smart Factory.
First, the company built a qualitative capability map, which included over 130 features based on the company’s IoT solution needs.
Secondly, the company investigated a range of vendors on the market and came up with a list of around 50 potential vendors.
Thirdly, these vendors were ranked based on their capability map resulting in 14 vendors being shortlisted for further evaluation against 9 key criteria:
- Functional coverage
- Expected cost
- Amount of vendor lock-in
- Partner ecosystem
- Reselling options
- Third-party integration
- Control points
Fourthly, the company selected three vendors to run PoCs onsite. Based on the PoC outcomes, recommendations were made to the board by a large evaluation team, and then the board (including the CEO) made the final decision on IoT platform selection.
Example 2: Automotive OEM
The following example presents the IoT platform selection and evaluation process of an Automotive OEM looking for a smart factory platform (not a connected car platform).
First, based on their internal analysis, the company pre-selected 10 vendors that were invited to perform 2-day PoC’s in a real production setting.
During the PoC’s the competing platforms were tested on usability, interoperability, and other technical factors to assess their suitability.
Following this stage, the automotive OEM kicked out vendors who did not deliver what they needed – some of which didn’t even deliver what they were promising. The remaining 2 best vendors were selected for a further 3-month PoC for in-depth assessment in the production environment.
A steering committee (made up of technical profiles and managers) made a recommendation to the management team (including the CEO) based on the PoC performance and cost. However, in this case the final IoT platform selection decision was taken by the Head of Plant Automation.
Best Practice 3: Ensure the CEO Signs Off on the Final Platform Decision
As choosing an IoT platform is an important strategic decision for any organization, in most cases, the final decision is taken by the CEO and a management team or a board of directors including the CEO.
“CEOs that sign off on the platform also drive its adoption”
One might think that the reason for involving the CEO is that IoT platform-based solutions are a large investment for any organization and should thus be signed-off from the top. This, however, is not the main reason. As one interviewee put it: “Moving from an analog to the digital world is a major undertaking. Our CEO understood that the IoT Platform is an important element of becoming a truly digital champion. Therefore he prioritized the IoT Platform selection and now drives its adoption across the organization.”
The CEO and board will be well informed following recommendations made by a team of technical experts, often based on the implemented tests and PoCs. Results from respondents are presented in Table 1 where participants were asked who is the final decision maker on IoT platform selection:
Customers are challenged to choose the right IoT platform for their needs that is fit-for-purpose now and future-proof in-line with their technology roadmap. Learning from others and following these 3 best practices can help choose the most appropriate IoT platform.
“Some firms reach different conclusions on the same purchasing criteria”
It is surprising to see that even if the criteria are somewhat similar, the outcomes are different for different companies.
Example 1: One company chose to go with a vendor stating that “even though there were gaps in their product portfolio, they seemed to be the most believable firm to close those gaps and be serious about future partnerships." Another firm kicked out the same vendor early-on in the selection process due to “lack of performance”.
Example 2: One company raved about the capabilities of an IoT cloud platform while another one lamented the lack of domain knowledge this provider would bring along.
Using several different platforms in the exploratory phases can help customers compare and understand which platform works best for their needs. In some cases, companies even build up a test lab with real data – then invite the shortlisted platform vendors to connect their platforms and show what they can do. But ultimately, it’s a management decision and the CEO typically has the final say — so choose wisely.
Published at DZone with permission of Padraig Scully , DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.