mm

 

Written By

Aaron Kamphuis

Share

Subscribe

Stay up-to-date with OST blog posts.

July 28, 2016

Part V: Selecting the Right Platform is Critical to the Success of an IoT Product

mm

Written By

Aaron Kamphuis

This post is Part V of the “IoT… are we all doing it wrong?” series. Check out Part I,  Part II, Part III, and Part IV then come back next week for Part VI, Which platform type is right for your IoT product?

Imagine this scenario…

A company wants to take a product which has been relatively unchanged for 20 years and combine it with a full digital experience. They feel confident that they have uncovered a new business model that will revolutionize their industry. At this moment, they have a critical technical decision to make: what is the best way to connect this product?

There are a wide variety of IoT platforms that are available for the product team to build their solution on.  The options are nearly endless. The number of IoT platforms that have emerged in the recent years is absolutely amazing, which makes the challenge of settling on an IoT platform extremely daunting.  The vast number of IoT platforms each have their own strengths (and weaknesses) and can each play out very differently.

Ultimately, the business should drive the platform selection.  However, how do you close the gap between business requirements and features offered by a technical platform? It is so easy to get lost in the details, miss the priorities and have misalignment across the team.

Establish guiding principles

It is best to start high-level, first by establishing guiding principles, then dig deeper.  Guiding principles are a broad set of very high level functional and non-functional requirements for your IoT product.

You would see requirements such as:

  •        Build versus buy
  •        Ability to scale to projected device count
  •        Security
  •        Time to market
  •        Platform lifecycle / long term viability
  •        Opex spend versus capex spend
  •        Technical fit (do we have the skills to build on the given platform?)

… and many more!

Guiding principles feed into the criteria in a scoring matrix. Be sure to organize, prioritize and eliminate any non-starters as soon as you can.  Properly evaluating a platform in an evaluation matrix takes a deep understanding of the chosen platform and is time consuming. By being able to quickly narrow down the players, the team can spend more time focusing on the platforms still in contention that meet the top priority criteria.

Understanding cost models up front is key

Having the highest score on the matrix is not enough. Applying financial modeling to understand how each platform would affect your capex and opex spend over time is critical.   You need to make sure that the platform you choose does not blow up your short term or long term business case. You will find that many of the leading IoT platforms have very diverse cost models.  Modeling costs will not be trivial. Some platforms you can get into at lower capital cost, but will come in at a higher operational cost as you connect more products. Others have higher upfront capital costs, but as you scale out, the operational cost model is much lower.

Capital costs are typically seen in the following areas:

  •        Hardware / infrastructure costs
  •        Purchase based licensing
  •        Engineering / development costs

Operational costs are typically seen in the following areas:

  •        Consumption based pricing
  •        Monitoring / support activities
  •        Network communications (cell / other)

Validate the chosen platform early to avoid increased risks later

Once a platform has been selected, identify a core set of features that will validate the platform / architecture and any assumptions made during the selection phase.  Proving that it will work and updating your assumptions is key in a project so that you can reduce your risk of failure before going too far into your budget/timeline.

Pivoting to a different platform can be done in the IoT space. However, a re-platform of an IoT product during it’s lifetime comes at a much higher degree of risk than an average IT project. Many factors, such as these, drive up the complexity/cost:

  •        Durable/Edge devices can number in the thousands to millions
  •        Durable/Edge devices can be globally distributed and are located in uncontrolled environments
  •        Platform integration is usually done at the hardware or firmware layer
  •        Durable/Edge device lifetimes may exceed 10 years
  •        Durable/Edge devices typically have limited resources and capabilities

Platform lock-in is a real thing in the IoT space. Picking the right IoT platform upfront demands careful consideration and should not be entered into lightly.  A team’s selection process should be driven by the business model, business requirements, as well as, the platform’s ability to continue to support a company’s long term IoT strategy.

OST has extensive experience assisting customers in strategy, user experience design, platform implementation, mobile/application implementation and data analytics in the IoT space.

This broad and deep experience has been leveraged to help many customers through this process of defining the requirements, navigating cost implications, and making well informed platform decisions.

OST is one of the few AWS IoT Competency holders in the world. This distinction points to our unique expertise in connected products.

Share

Subscribe

Stay up-to-date with OST blog posts.

About the Author

Aaron Kamphuis has spent 20+ years in data analytics, application development, cloud architectures and software testing with a background leading development teams in the use of cutting-edge technologies to satisfy unique business/end user requirements.

Aaron spent several years with Sagestone Consulting Inc., where he worked with his partners to build out 65+ people for a multi-million dollar application development services organization.

At OST, Aaron and his teammates have worked with clients to build global scale IoT solutions, Data Analytics solutions for packaged software, and companies that range from large multi-national enterprises to small businesses.