Member preview

IoT: Monetisation with Analytics (the ‘How?’)

In part 1, I have discussed;

  • State of Commercial IoT is not at the desired stage, despite the big hype and the inflated expectations behind it.
  • It’s a result of multiple reasons, but mainly the challenges in integrations and the difficulty in identifying ROI have contributed.
  • Considering IoT as ‘Monetisation through Commoditisation of a technology’ provides us with a way to take a business value-centric approach to IoT.
  • The key element in monetisation is Analytics.

So in this second part, I will discuss ;

  • How Analytics become the key monetisation element and looking at the big picture, why IoT sensors are only one tool of a bigger solution driven by analytics.
  • To be able to deliver true business value, quickly, some key steps I have identified with our experience.
  • How we can leverage additional (existing) data sources together with IoT/sensor data to deliver exponential value, fast.
  • Discuss some key capabilities the analytics tool need to possess, in our experience. (Caution: no silver bullets, or OSFA)

So what do we analyse?

While not rocket science, I have seen many organisations often missing the fact, that IoT isn’t all about sensors. While Sensors are a key component in IoT, if you look at from the business outcomes viewpoint, sensors are just another data source that acts as a feed into an Analytics outcome delivering operational or business insight.I.e. IoT sensors are not the solution, its a tool. The solution is the outcomes delivered via analytics (which of course a combination of many components)

In my experience, in most use cases (almost all I would say), to deliver true business value (which sometimes is multifaceted) in a reliable fashion, you require more than just sensor data. And some use-cases you don’t even need any new sensors as such. Combination of existing datasets already have answers you want. The common additional data sources we have seen in OT/IoT related projects are;

  • Existing Business Applications (CRM, ERP)
  • IT systems (corporate websites, customer portals, e-commerce)
  • Mobile Apps (Navigation, Social media, Messaging)
  • WiFi/Location systems
  • Industrial Assets/OT systems
  • Electronic Security Systems (CCTV, RFID, Fire & Intrusion)
  • Other industry-specific operational systems (e.g. Production lines in manufacturing, Nurse-call systems in hospitals)

The essence is that the business value of OT/IoT use cases will grow exponentially when you analyse these multiple datasets together, without getting limited to just sensor data.

So in delivering exponential value from data, there are few things I have seen work for us, which I thought is worth sharing.

Start without “IoT”:

While it may sound funny, by experience, you are more likely to achieve quick success by capturing the operational or business pain points, known challenges, or any areas for improvement, without focusing on IoT (or any technology for that matter). It will also help you open up the stakeholders’ minds to share the challenges without them having to get limited by having to worry if it is relevant to IoT.

Go for the quick wins:

Once you figure out the pain points and challenges, then picking the quick wins will be critical. There are few key aspects to consider;

Level of business/operational impact: Sometimes it can be easily attached to a $ figure, or to resolving/optimizing a day-to-day operational issue depicting resource/time savings. Sometimes it may not be directly visible. Sometimes the idea may not be as grand….

We once had an inquiry from a city council on an initiative to trial ‘IoT’. During discussions where many grand use cases were thrown in, one suggestion was to monitor BBQs. At first, the idea was more like for fun, but going into details and understanding a bit more depth of the council’s operation, the current challenges, and potential cost savings were clear. The outcome was a solution which only required measuring one metric/parameter delivering 3 use cases;
  1. Usage-based maintenance of the BBQs: Saves money by a) avoid permanent damage to BBQs due to overuse b) reduce the maintenance contractor’s charges by servicing only the ones required.
  2. Near real-time BBQ usage/availability data (in future to make available for residents): Improving council living standards
  3. Monitoring BBQ operation costs & trends: Energy costs and maintenance costs

Below is one of the dashboards(anonymised) we developed for the PoC.

Alignment with business goals: While the business impact will take precedence, especially when it’s not clear, finding out how well aligned the use case(solving the issue, improvement) with the organization’s business goals, or vision. Because sometimes it can be something not directly attached to a $ value.

Ease/Availability of data: While not to lead with, but an essential aspect is to check if the data you require is already available, if not how easy(technically & commercially) to collect. As I said earlier, IoT isn’t always about having to deploy new sensors.

Most organisations are sitting on treasure troves of untapped data, which if mapped correctly with the use cases, can readily deliver value without having to deploy a single new sensor. To give you some common examples;

  • CCTV cameras: the smartest, (data) richest and most matured IoT sensor in the market and almost all organisations got them already. A valuable source for use cases such as people counting, crowd analytics, retail/shopping centre visitor analytics, Car park, kerbside parking analytics, safe city use cases, critical assets monitoring.
Below example was using CCTV to monitor kerbside parking (for councils).
  • WiFi Access Points: They are everywhere and they collect more data than you think. While the capabilities and available data could vary from vendor to vendor, it ranges, from retail/visitor analytics, public WiFi, location analytics, asset tracking, facility usage tracking, predictive facility maintenance
  • Electronic Security Systems: In addition to CCTV, Access Control(RFID) systems, Intrusion & Safety Alarm systems, Intercom systems are few other sources which commonly exist in organisations.
  • Operational Technology (OT) assets and systems: Industrial IoT is considered as a domain on its own. And specialty in that domain is most sensors required are already in place within the process systems or plants, and data is being gathered for operational purposes. Hence its a matter of tapping to the datasets to deliver most of the use cases in the Industrial domain. But sometimes people get discouraged thinking the data is hard to access or require complicated plumbing. Not the case always;
Below is an example I usually keep bragging about. Data source was a SCADA data dump, and it only took indexing 2 years worth of data, and a 30 mins train-ride before finding the root cause for previously known issue, lingering for almost a year. I have blanked out the asset names, but those two charts together pointed us to the root; triggered by humidity condition in a heat exchanger
  • Then there are OT systems that fall in between OT & enterprise, e.g. Nursecall/Duress alarm systems in Hospitals, Voice pagers, POS machines.
This example is about Nursecall data, just during the PoC phase analysing data from an Aged Care nurse-call system, we made a revealing discovery. One-third of the calls werent responded within SLA time period!

So the point I wanted to emphasis is that there is a lot of value you can deliver just by tapping into existing data sources where the use cases can be categorised as IoT.

Identify new data sources… carefully

In understanding availability/ease of collection of data, it’s very important to do a thorough data source analysis per use case, to make sure you capture the best fit sources and methods. There can be multiple methods to achieve one use case, but considering technical and commercial factors relevant to the use case, you can take a decision on what fits the use case most; not to pick the ‘superior technology’ and fit the use case into it.

Take vehicle(GPS) tracking as an example. While concept/technology remains same the use case can have a big difference. There are GPS trackers operate on 3G, and then there are newer ‘IoT Trackers’ that work on Low power networks. But if the use case is fleet safety involving motorway travel, 1 second at 110km/h is 30m distance. Which means you would require sub second data capture and large/continuous transfers. And in such cases relying on the old tech can be more practical.
Different type of an example ; from an Air Quality Monitoring use case for a Smart City, where, instead of investing on expensive gases sensors to measure all parameters, we have used a combination of field sensors and Bureau of Meteorology data feeds (for SO2 and Ozone) of closest locations.

Analytics != Visualisation

I have mentioned this before, that sometimes organisations get lost in detail and losing direction. So it’s important to keep your end goals clear;

  • What is the use-case, insights expected
  • How is it going to resolve a problem, or improve the operation
  • How are you going to measure the success

Then it’s important to pick a right platform that can deliver today’s outcomes while having the flexibility & scalability to deliver future outcomes and use cases. In deciding this the key factors to consider would be;

  • Ability to easily ingest all required data sources(Business, IT, OT, IoT) with minimum tinkering, and breadth of ingestion capability for future needs.
  • Scalability to expand in capacity and use cases
  • Proven analytics capabilities, ideally being applied in multiple use-cases expanding across Business, IT, OT, IoT (not just IoT)
  • Flexible deployment architecture; On-prem, Private/public cloud, distributed analytics etc.
  • Product maturity, customer base, and market credibility
  • Skills & expertise availability (partner, tech community adoption)
  • Overhead (staff, training, infrastructure costs required to operate)

Essentially, in my view, Analytics being the most critical element, and what impacts the monetisation directly, it’s critical to select the right analytics platform selected for the right reasons.

Closing note

Thank you for reading this article. Hope you foundthis article useful. Would love to hear your feedback, thoughts or even any critisisms (ideally constructive :)).

Kushar Perera, Head of Innovations @ DNAConnect