The Long tail of IoT

[Cross Post from Mr Param Singh' Blog ]

...and how to spot non-replicable use cases

We have all heard of the concept from Chris Anderson’s book ‘The Long Tail’, as it was applied to websites- the hypothesis being a majority of individuals visit only a few sites regularly and that the vast majority of sites appeal to smaller number of specialized audiences - at least my summary interpretation of it 🙂

There are parallels to be drawn between that concept and what we are seeing in IoT. As pointed out in the prior article ‘Is IoT a viable market?’ the success of POC / pilots has been limited. More interestingly, if one was to map out these use cases one is likely to see a similar distribution of use cases (of course this will vary by consumer vs industrial or a particular company’s domain expertise). With a few proven markets dominating the landscape and then a very long tail of uses that use data...

In my own analysis, the markets / segments that tend to be on the left side of the ‘skewed histogram' are the uses from existing enterprise, commercial or industrial applications and legacy market segments. What characterises these segments is that there is no need to conduct pilots to proven the value. There is Documented ROI and scores of customers who have been paying for thee solutions for years or even decades. Enterprise Asset Management is a good example where factory owners have been tracking the lifecycle of their assets.

It is far easier to convince customers of the additional value from the data from sensor streams when it is in the current context of their existing assets, such as location, machine status, machine etc.

This shortening of the sales cycle is critical for both large companies looking to scale revenue as well as IoT startups that need to gain a foothold and prove value to their investors. Some of the other market segments where sensor data can more readily add value, include :

  • Enterprise Production Monitoring.
  • Manufacturing Intelligence.
  • Operational Efficiency.
  • Field Service Management (Service Cloud).
  • Building Management Systems.
  • Fleet Management Systems.
  • You get the picture…

Another segment that is fascinating is Field Service Management. An colleague is installing sensors across a few city blocks with the full range of use cases: air quality, traffic, physical . When an event is detected from sensor data, that ‘alert’ is pushed to one of the leading Field Service Management cloud service. Which in turn triggers a service truck roll to resolve the issue. This solution did not start this way, the first solution from a vendor had their own reference dashboards to view the alerts with no integration with the customer’s chosen service cloud offering - it did not go very far. Most customers want the data, events, alerts to integrate into existing systems and apps.

Finally are some use cases such as ‘track and trace’ that are not quite market segments but can represent a significant opportunity for sensor data. The issue here is that the architecture needs significant modifications based on which market segment / assets one needs to track- perishable frozen palettes, high value servers, trucks, containers etc.

Just as everyone had an idea for a website in the late 90s or everyone has a movie script in Hollywood, most folks have an idea for things that can be connected. Most of these would fall on the extreme right of the long tail- where sensor data is theoretically interesting, but the value is far from being proven. The implication of this, for large IoT companies, is that this translates into non-replicable sales, insufficient gross and contribution margins and slow growth.

I have resisted the temptation to provide a list of examples (there are so many), instead here are some some common characteristics of these use cases require:

  • A very high degree of customization is required.
  • Use cases are very specific and appeal to a very unique set of potential customers.
  • The solutions are not well integrated into the existing systems.
  • Customers are not self-referencing.
  • of the pilot (, software) it too high for the value achieved.
  • The returns are over multiple years.
  • Business models have not evolved from perpetual licensing to subscription or delivered as a service.
  • The ownership of the budget is not clear (between say between OT and IT).
  • Even successful pilots in IoT face unique issues in scaling to production. The protocols, compute, data orchestration do not scale linearly and often require a re-architecture of the entire solution.
  • Other factors that are specific to an industry / use - that prevent the use case from being replicable.

One of the most insightful comments from an system integrator (SI) colleague of mine is that most IoT vendors / startups overlook the total cost of ownership (TCO) of the full solution. Often the TCO is prohibitive for most customers even if a specific technology piece is reasonable.

It would be great to hear from folks on which segments / use case they are seeing significant traction in...

Leave a Reply