How to launch IoT devices – Part 1: Why it takes so long

nilayshrugged

on 14 January 2020

(This blog post is part of a 5 part series, titled “How to launch IoT devices”. It will cover the key choices and concerns when turning bright IoT ideas into a product in the market. Check out this white-paper on IoT app stores for background reading.)

You have a budget and a bright idea. How do you turn that into a revenue generating and market capturing product? How do you escape “Pilot Purgatory” and the volume of decisions that need to be made about IoT before day-0? 

This blog series distills the learnings from over 30 Canonical project summaries and case studies from IoT launches. You’ll learn about the key design decisions and choices to consider. The goal is to increase the velocity of your technical decision making. It will show how Ubuntu Core and its suite of products are designed to turn IoT launches into a 5-step journey that can take as little as 2-weeks.

Need for speed: How e-commerce launches fast

This decade, IoT will become a trillion dollar industry. Currently it sits between $200 and $250 million, so there is a lot of planned growth. Best of all, the projected value is still up for grabs by new players in the market, i.e. you.

Contrast this sentiment with McKinsey’s finding that 85% of IoT products are still in pilot, one year from starting. 25% take more than 2 years to release. To find out why, let’s compare IoT with another “modern” business model: e-commerce and retail. 

A generic launch sequence
An e-commerce launch, positioned against generic launch sequence

Generally speaking*, the steps to launch a retail product are outlined above. The internet’s impact is that each stage is, at best automated (Amazon, Sourcify, Swipe) and at worst commoditised (Ali Baba, East-Asian manufacturing and wholesale, 3PL). The result is a de-risked, faster and cheaper retail product launch.

So why are so many IoT business ideas, stuck in Pilot Purgatory?

Automated and Commoditised launch of IoT devices

To answer this question, I looked at Canonical’s internal database of projects, to see how successful IoT devices are launched, as well as the problem statements customers came to us with. This covers +30 business cases, project summaries and case studies. 

A few things stood out as different (but similar) to e-commerce and retail, as seen below.

An IoT launch, positioned against generic launch sequence

Sourcing

A decision on what IoT hardware to use becomes a decision on what your entire software stack needs to be. Parts of that stack might not work together or with the operating system, apps or your current CI/CD while remaining reliable, robust and secure.

Best case scenario: Select hardware that is known to work with a full-featured IoT stack, now and in the future. 

Customer Interaction

Making a portal is not about finding a GUI that a human user/customer can easily access (as is the case with retail). IoT devices are typically unmanned for much of a device’s life. While fleet management vendors exist, it is difficult to get this to work with a) the hardware selected and b) the operating system and software stack selected. 

Best case scenario: Select a software distribution method that allows a device manager full control over what is running on a device, with minimal downtime and in-field engineering costs when apps change.

Value exchange

Write good software that solves a use-case, and integrate this so it works well with hardware and the operating system. This requires deep knowledge of embedded development and the hardware, operating system and kernel. 

Best case scenario: Leverage your current code base and DevOps while writing code that will definitely work on your device

Distribution

Finally, to ship a device, it needs to be configured specifically for each customer. Letting the customer configure a device, is a large source of friction. Configuring a device before shipping reduces first touch friction for the customer at the cost of engineering time.

Best case scenario: Provision a device in your factory, but without needing extensive engineering support and costs.

Summary

Many modern businesses have commoditised or automated steps which make taking a product to market faster and easier. IoT does not have this yet. To get the answer to these questions now, contact Canonical and read this white-paper on IoT app stores for a bit more background. 

The following blogs in this series will show you how to reach the “best case scenario” in all of the above steps. Next, we will look at selecting hardware. You should sign up to our IoT newsletter on the right hand side of this page to make sure you read it when it comes out.

*Specifically, you can start with Chinese wholesalers like Alibaba and get the raw materials to sell. Amazon provides access to +250 million customers with a store front, or Shopify can be used for the DIY enthusiast. Then, East-Asian manufacturing can be leveraged at any scale, for example by using Sourcify. This adds value to raw materials and turn it into a unique product, while Stripe, Paypal, or banks take payments. Finally, third-party logistics gets the product from warehouse to consumer.

Internet of Things

From home control to drones, robots and industrial systems, Ubuntu Core and Snaps provide robust security, app stores and reliable updates for all your IoT devices.

Newsletter signup

Select topics you’re
interested in

In submitting this form, I confirm that I have read and agree to Canonical’s Privacy Notice and Privacy Policy.

Related posts

Roadmap for Ubuntu official support for the Raspberry Pi 4

With 19.10 release of Ubuntu Server, Canonical announced official support for the Raspberry Pi 4. The latest board from the Raspberry Pi Foundation sports a...

Standardising software to quickly deliver a smart city platform for millions

New York has the highest population density of any city in the United States with over 27,000 people per square mile. The average New Yorker has a commute of...

Getting started with Ubuntu Core – streaming video from a Raspberry Pi

Artificial intelligence relies on machine vision just as much as human intelligence relies on vision. Image sensors are, therefore, crucial for AI...