Solutions

The road to success with IoT

Building a strong business case for IoT

Where business leads, technology follows

How can you avoid common mistakes in your own IoT projects? Start by asking the right questions about what you want to do with your data. These questions include:

What goal do we want to achieve with the information collected via IoT devices and other sources?
What meaningful answers do we want from this collection of data?
What business insights do we want to learn?
What business problem do we want to solve?
What type of value will such answers bring to the business?
How will we measure the value of such insights and their commensurate cost?
Does our organization have the foundation to support adding the new volumes and varieties of data necessary for IoT real-time or near-real-time insights?
Can IoT data help answer our questions? If so, how?
What other types of data do we need to receive this type of business insight?
How hard will it be to obtain and process this data?
How much will it cost to extract the necessary meaning and insight from IoT data?
How hard will it be to protect, secure, and manage the IoT and edge devices we need to collect and process this data, along with the data itself?
How easily can we scale the IoT project beyond a single, LOB-level use case or its initial PoC?

Answers to questions like these can act as the foundation to build a more solid (and successful) business case for your own IoT project.

The strategic journey from IoT data to wisdom

Leading with an integrated data fabric approach

You can make answering early business-side questions easier by applying a more strategic, business-driven analytics framework to your next IoT use case.

This framework requires a more strategic lens from which to view Barrett’s original IoT Data-to-Wisdom pyramid. Through this lens, organizations can focus on the specific disciplines and tasks required to achieve agility at various levels of their own data-to-wisdom evolution.

A more strategic lens also includes an accompanying, integrated data fabric which manages data and analytics together with automating pipeline processes, such as ingestion, preparation, and processing. By combining strategy with more automated processes, you can obtain more valuable insights from your next IoT project while also minimizing data silos and more effectively managing data across the enterprise.

[Source: Stratascale.]

Before you embark on your next IoT project or pilot, you should be able to answer questions about how you can apply disciplines like analytics, business intelligence, ML, and AI to extract meaningful insight from the IoT.

[Source: Stratascale.]

Each layer of the data fabric pyramid shown here involves various tasks. At this stage, some organizations may find it helpful to seek guidance defining their specific project goals and objectives, then mapping those to the different data intelligence layers in the framework.

(Sometimes, even an inverse pyramid may show insights originating from data on IoT sensors. One case in point: This i-SCOOP article shows Big Data insights and implications about ocean currents first generated from a funnel of sorts that began with a network of IoT sensors on moored buoys.)

Building device-level considerations into your IoT strategy

Look to answer these key questions

To help solidify your IoT strategy, you need to ask questions about the IoT devices themselves, such as:

Which IoT devices will we deploy?
Who decides which devices we deploy? How do they decide?
How do we track the efficacy of deployed IoT devices?
What process do we follow to remove, improve, or add devices? What about patch management? Which team is responsible for these areas?
How will we manage the data and which teams will be responsible for managing it?
How will data be collected or transmitted in the event of Internet or network disruption?
How do we ensure the devices adequately support the needs of the business?
Which platform team will own the IoT rollout?
What policies do the IoT devices need to follow?
Who defines security and other policies for these devices? How do they define them?

Note: These questions borrow from Stratascale’s Cybersecurity team and its Horizon Report on Device Assurance. See this report—along with a related Insights article on IoT security—for more guidance to properly secure your devices.

If you plan to use IoT devices or edge servers for ML computations, a Stanford lecture on Machine Learning Design suggests answering the following types of questions as well:

How much power does the device or system need to process the ML computations?
How much memory does the device or system need “to store ML models and load them into memory”?
What type of battery or power source do you need to support the application and its computations?

For more IoT hardware, software, and connectivity-specific questions, see the RFP Programme’s RFP/RFI (fee-based) templates provided by the IoT M2M Council (IMC).