The Four Budget Domains of Enterprise IoT Projects
IoT is rapidly transforming manufacturing, energy, healthcare, and several other industries. The ability to predict asset failure, rapidly diagnosis bottlenecks, and invent new revenue streams are elevating industry standards at a pace that is difficult to keep up with. Executives in these industries have been asking me to give them the basics of what they need to do, and I've been using this simple model to provide guidance.
Sense - Sensors and Microcontrollers
The core question of this domain is: “What valuable information do we not have today that we could have tomorrow?” Adding sensors to assets is relatively easy compared to deeply understanding what information is the most valuable to the business.
- Activities: Supplier selection, prototyping, power and storage experiments, field testing, security and threat assessments, update paths
- Things to buy: Source suppliers for sensors and microcontrollers, installation and maintenance hardware
- People you’ll need: Manufacturing/Electrical Engineers, Software Engineers
The core question of this domain is: “When, where, and how do we get sensor data off the device and into our cloud?” The connected home may be sending data every second to the cloud, whereas the connected offshore windmill is only sending data when service vessels are nearby.
- Activities: Selection of network protocols, north- and southbound protocols, investigate partnerships with telecoms, select cloud platforms and streaming stacks.
- Things to buy: Gateways, potentially licensing bandwidth on cell towers, cloud instances for queuing and streaming data.
- People you’ll need: Data Scientists/Engineers, Network Engineers, and Cloud/DevOps Engineers.
The core question of this domain is: “What actionable insights can be obtained with the massive amount of data pouring into our business?” Predicting asset failure or maintenance needs can be done with simple algebra all the way to artificial neural networks
- Activities: Applying the scientific method to machine learning and statistical algorithms, optimizing early detection and notification architecture, preparing data for consumption by apps and 3rd parties.
- Things to buy: Cloud instances for long term storage, in-memory computing, and APIs. Licenses for machine learning and statistical analysis software.
- People you’ll need: Data Scientists/Engineers, Cloud/DevOps Engineers, Software Engineers.
The core question of this domain is: “How do people experience, learn, and take action based on our IoT investments?” This is the space where you consider the web, tablet, and phone portfolio and (hopefully) make them the best experiences possible. Sales teams will thank you for investing heavily in providing excellent user experiences of charts, tables, notifications, and other features in your product portfolio.
- Activities: Wireframes and comps, prototypes, user testing, agile development methodologies
- Things to buy: Licenses to any Software Development Kits (SDKs) for charting, tables, and other UI components. IDEs and cloud instances for web and mobile environments, marketing assets, and other common go-to-market investments.
- People you’ll need: Software Engineers, QA Engineers
- As with all investments, deep dive into The Five Whys.
- Be wary of Penny Wise; Pound Foolish - Sample as many sensor and microcontroller suppliers as possible, and field test their products to exhaustion.
- Never underestimate the number of security risks you are introducing to the business.
- Unleash marketing, product, and project managers on solidifying a vision and execution plan.