The Internet of Things (IoT) is rapidly moving into the mainstream of business and IT. Fueling the trend are software platforms that introduce a framework for building networks and systems.

A typical IoT platform encompasses infrastructure, managed edge, managed data, security, integration, optimization, operations, applications, and reporting. The importance of those functions makes platforms crucial to the success of IoT projects, according to a new report from BI Intelligence, The IoT Platforms Report: How software is helping the Internet of Things evolve.

“Platforms are the glue that holds the IoT together,” the report notes. “[They] allow the IoT to achieve its transformational potential, letting businesses manage devices, analyze data, and automate the workflow.”

For anyone offering services and solutions in the IoT marketplace, it’s critical to sort through this rapidly changing space. According to a report from German research company IoT Analytics, more than 450 companies offered IoT platforms worldwide in July 2017, a 25 percent increase over the previous year. Complicating things further, IoT platform vendors toss out an assortment of claims and often define the space differently.

Make Connections Count

Yet the IoT platform selection process can make or break an initiative, according to Benson Chan, senior partner at research, acceleration, and advisory firm Strategy of Things. “With so many companies streaming into the market, and no end in sight to the growth, understanding what a company does and what it offers is critical,” he says, adding that some platform vendors are feeding on the confusion by delivering exaggerated claims or subpar products.

“There are issues and challenges specific to different industries and the business, including the types of data sets used and compliance regulations. It’s important to not be fooled by marketing claims,” he adds.

Platform capabilities vary greatly across an array of factors, from data management and delivery speed to security and privacy. Even the best platform, moreover, may not be adequate on its own. “Building a robust solution requires multiple vendors that may reach across horizontal platforms from Amazon, Google, or Microsoft; or vertical platforms for specific industries or types of initiatives, such as smart cities or industrial IoT,” Chan says.

To make matters harder, not every organization has the required skill sets to build applications using all the platforms they require, or the tools that are provided with them. Plus, the same set of tools can produce very different results for different businesses, including gaps in security and privacy.

Ecosystems Matter

Sorting through platform options is no simple task now, and it’s set to get even more difficult. According to the BI Intelligence report, the “IoT platforms market is set to expand rapidly in the years to come, with current leading platforms expanding and others entering the space.”

Chan says a good starting point for your selection process is to identify each platform’s value proposition and then use that to determine which ones are best aligned with a client’s mission and goals.

Then map out the platform’s infrastructure elements. This includes the types of devices it supports (such as robots, drones, sensors, and vehicles), its connectivity and gateway components (such as Wi-Fi, Bluetooth, Zigbee, Z-Wave, Sigfox, and LoRa), and its messaging protocols (like HTTP, MQTT, WebSocket, and more). Also consider what presentation components it uses (such as mobile OS, Web browser, augmented reality, virtual reality, voice interface, or wearable interface), and in what industry the solution will be deployed (such as manufacturing, oil and gas, smart city, healthcare, automobile, or enterprise).

A mismatch in any of those areas can lead to an array of problems, including latency, data collection and management challenges, or disaster recovery obstacles, Chan says.

Identify each platform’s value proposition and then use that to determine which ones are best aligned with a client’s mission and goals.

Be sure to investigate where a platform operator’s data centers are located. A business may not want certain data stored outside the U.S. or another country, and regulations may dictate that data is or isn’t stored in a particular place. Some platforms or platform combinations might not deliver the necessary flexibility to conform with these needs.

Finally, Chan advises, evaluate compatibility and interoperability across platforms. A report from IoT Analytics, 5 Things To Know About The IoT Platform Ecosystem, states that “vast differences” exist in platforms claiming to do the same things.

In the end, though, more goes into the success of an IoT deployment than which platform you use. The underlying network matters too. If it isn’t designed for maximum performance and efficiency, your platform—and the resulting IoT deployment—may fall short of expectations.

Samuel Greengard is a business and technology writer based in West Linn, Ore. He is author of the Internet of Things (MIT Press, 2015).