What is "pilot purgatory"?

It is the stall point where a prototype works on the bench but never scales to production. Research suggests as many as 84% of IoT deployments get stuck there, usually because operational readiness was never validated.

What do EVT, DVT, and PVT stand for?

Engineering Validation Test, Design Validation Test, and Production Validation Test. EVT checks that the design functions, DVT checks that it survives the real world, and PVT checks that the factory can build it reliably at scale.

Why is a watchdog timer considered non-negotiable?

Because it automatically restarts a device if its firmware freezes, recovering the unit without a technician visit. For fleets in the field, that prevents expensive service calls.

Why are prototyping decisions so hard to change later?

Once a device is deployed in the thousands, hardware modifications are economically impossible. Choices about the processor, radio, and firmware architecture effectively become permanent at scale.

courses > Cellular IoT Foundations > IoT product lifecycle overview

chapter 07

IoT product lifecycle overview

7.1From idea and prototype to production and deployment 7.2Manufacturing and certification basics 7.3Fleet deployment overview 7.4Support and maintenance lifecycle

7.1

From idea and prototype to production and deployment

Q: What is a Minimum Scalable Unit (MSU)?

It is the smallest fleet size that still exercises every part of your production system that includes provisioning, monitoring, support, and billing. Deploying an MSU during the Pilot gate confirms your operations scale, not just your hardware.

TL;DR

Key takeaways

Start with the job, not the chip.

Define the business problem and functional requirements before selecting any hardware or platform, or you will pay for it downstream.

Ideation has concrete deliverables.

A value proposition, ROI model, technical scan, certification roadmap, and user journey map should exist before you write firmware.

Prototyping locks in hard-to-change decisions.

Right-size the processor, connectivity module, and firmware architecture early, and design for the physical environment from the start.

Four gates protect you from scaling mistakes.

EVT proves function, DVT proves durability, PVT proves the factory, and the Pilot proves that operations scale.

The MSU breaks pilot purgatory.

Deploying a minimum scalable unit tests provisioning, monitoring, support, and billing together, not just the device.

Deployment is the starting line.

Production hands off to operations, the longest and most demanding phase of the product's life.

Every connected product, whether a smart meter, a shipping-container tracker, a farm-field sensor, followed the same journey: someone had an idea, built a prototype, validated it, manufactured it, and deployed it. Research suggests as many as 84% of IoT deployments get stuck in "pilot purgatory," the limbo where a prototype works on the bench but never reaches production scale. By the end of this lesson you will understand each stage of that journey, the key decisions at every gate, and how to move through them deliberately.

What you'll learn

Why every IoT product starts with a business problem, not a technology choice

Successful products begin by answering a single question: what job is this product doing for the end user? A "Jobs-to-Be-Done" framework helps you identify the friction point your product eliminates and translate it into precise functional requirements before you pick a cellular module or a cloud platform.

What deliverables to produce during ideation and requirements

At the ideation stage you should produce five deliverables:

Value proposition statement: What pain point does this solve, and for whom?
Financial ROI model: Market sizing including Total Addressable Market (TAM), Serviceable Addressable Market (SAM), and Serviceable Obtainable Market (SOM).
Initial technical scan: Can the physics of range, power, and connectivity support this use case?
Certification roadmap: Which regions will you sell in, and what approvals do you need?
User journey mapping: How will operators, installers, and end users interact with the product day to day?

What hardware and connectivity decisions to lock in during prototyping

Prototyping is where you make choices that are extremely difficult to reverse, because once thousands of devices ship, hardware changes become economically impossible. Right-size three things:

The processor: Enough compute for your application logic, without wasting power and cost.
The connectivity module: Choose the right radio for the use case. Cellular options such as LTE-M and NB-IoT give kilometer-range coverage at medium power, ideal for logistics, smart metering, and asset tracking; LoRaWAN, Wi-Fi, and satellite serve different tradeoffs.
The firmware architecture: Build modular firmware from the start, separating the networking stack from application logic so you can update one without touching the other. Include a watchdog timer, which automatically restarts a frozen device and prevents costly field service calls.

You should also harden the design for its environment. Account for temperature shifts, humidity, and vibration, and for factory-floor or outdoor deployments use industrial-grade components rated to IP68 or NEMA enclosure standards.

How the four validation gates move a prototype to volume production

In software you can push a patch; in hardware, a mistake at scale is hard to recover from. Four formal gates catch problems while they are still cheap to fix:

EVT (Engineering Validation Test) confirms function. Does the hardware work, can the firmware reach the cellular network, does the sensor read correctly? You are verifying the design works at all, not testing reliability.
DVT (Design Validation Test) confirms durability through thermal cycling, drop testing, and vibration testing. Can the product be built reliably and survive its environment?
PVT (Production Validation Test) shifts focus to the factory. Can you build efficiently at scale, what is your yield rate, and are quality-control checks catching defects?
Pilot deploys a Minimum Scalable Unit (MSU), the smallest fleet size that exercises your entire production system: provisioning, monitoring, support, and billing. The MSU methodology forces you to validate that operations scale, not just that hardware works.

Why production and deployment are the beginning, not the end

Passing all four gates moves you into volume production and deployment, but a validated prototype is only the starting line. Production and deployment open the longest phase of the lifecycle and introduce new variables in how you build units in volume and how you get legal clearance to sell them.

quiz

Lesson 7.1 Self-assessment

1. According to the lesson, what should every successful IoT product start with?

A. A choice of cellular module

B. A business problem

C. A cloud platform selection

2. What do the four validation gates stand for, in order?

A. EVT, DVT, PVT, and Pilot

B. Prototype, Production, Pilot, and Product

C. Design, Build, Test, and Ship

3. What is a Minimum Scalable Unit (MSU) used for during the Pilot gate?

A. Testing a single device on the bench

B. Deploying the smallest fleet that exercises the entire production system

C. Estimating the total addressable market

faq

Frequently asked questions

next lesson

Manufacturing and certification basics

Learn how to scale your IoT device production. Review component sourcing, factory security, and essential certifications like FCC, CE, UL, and RoHS/REACH.

Start lesson