What Makes IoT Enterprises Click Technology Backend to Support Deployment and Field Troubleshooting as the Third Critical Requirement
In my first two posts, I emphasised on “hardware robustness and extensibility” and “integrated software and hardware development” as two requirements critical for success of IoT enterprises. In this post, I will touch upon the third and the last critical piece for successful IoT enterprises — Technology Backend to Support Deployment and Field Troubleshooting.
You have worked day and night to develop your hardware, and ensured it passes the “reasonable robustness” test before you plan to deploy it even at a scale of a few hundred. All this while, you have ensured that the software (both the firmware running on the hardware and the one running on the cloud) is developed in an integrated manner with the hardware so as to ensure overall system reliability. What’s left now? Since IoT is an upcoming domain, aiming to collect data that is not seen before (and ensure that analysis thereof can provide useful actionable insights which eventually become the differentiating factor for your venture), it is very likely that specialised skills will be required for deployment (and now I am getting specific to B2B IoT and this may not be very critical for consumer IoT) and maintenance of your hardware. As your organization scales, one option is to proportionately scale up this specialised workforce. The other (and the smarter) option is to start leveraging technology to ensure that field logistics (deployment and maintenance) can be catered to at a large scale without proportionately throwing people at this problem. After all, as an IoT enterprise, technology is in your DNA.
The first stage where technology should be put to use to scale up operations is the pre-installation testing. Organisations should learn from different field scenarios in which system have failed before and ensure that all those dimensions are well tested before the system is marked ready for future deployments. It is important for this feedback loop (careful categorisation of field issues, abstracting out product features required to address these issues, dividing the resolution of these issues into hardware/software modifications, developing the corresponding modifications and then enhancing the testing procedure to incorporate testing of new modifications) to be as quick as possible. Several open source and off-the-shelf tools can be leveraged to put in place processes so as to ensure timely reaction all the way through the feedback loop in order to quickly put into practice the lessons learned from the field. Rest of the things — actual product modifications and enhancing testing procedures have to be part of your core technology strategy.
Testing procedures need to go through continuous iterations as they evolve based on issues addressed in the field. All this while, it should be ensured that the whole process is automated to whatever extent possible — How else can you think of testing and deploying 10s of thousands of devices (once a reasonable scale is achieved, this division for assembly and testing at both unit hardware as well as system level can likely run as an independent unit — sourcing and predicting future requirements and maintaining a tested inventory accordingly)! Once this process is setup, you can even contribute to the overall ecosystem — allowing your non-competing IoT enterprises to leverage your setup and get to a reasonable scale (while learning all the way before they setup their own such process or before there is large enough demand from multiple IoT enterprises such that it make sense to just run this whole setup as a separate independent entity, catering to a significantly larger volume and thus providing economies of scale to everyone). In addition to developing the current assembly and testing methodology to cater to automated and robust testing at the scale, this idea of supporting the IoT ecosystem is one of the things on our mind to carefully evaluate and expand upon at an appropriate time in the future.
Building a team of experts for installation and maintenance is again a journey that has to be lived through and there is no shortcut to it (and the eventual holy grail is to develop a separate training arm which can leverage the large workforce and train them for the specialised work — in accordance with the “Skill India” dream). At best, you will get people who are reasonably good in their own domain — Electrical, IT etc. Training programs are to be setup to ensure that they get trained for the specific nuances of your system. Even then, they will have their own share of errors in the field. Technology should be developed to catch and address these errors at a time which is least expensive (which is, at the time when the person is still at the installation site). Specific to our energy monitoring and control solution, we have developed a mobile application for field troubleshooting. Our electrical people (who install the system) need to just run the app and perform a few clicks. At each click, a specific functionality of the system gets tested (through the scripts developed accordingly and running on the system — another example of utility of integrated hardware-software development) and a pass/fail response is provided. If the test has failed, the system knows what specific connections have gone wrong or which piece of hardware is not working and accordingly directs the electrical person to address the specific concerns. The deployment is considered complete only when all tests pass and a report is automatically sent for our records (these reports are buffered on phone for electrical people to upload them after they return back from a busy schedule of several days doing multiple deployments every day to a place where they have access to free/cheap and high bandwidth internet connection). The application also captures the GPS coordinates for reference of field support team to ensure they do not waste time in locating the place they need to reach for addressing a specific problem.
Technology should also enable troubleshooting without a physical visit, wherever possible. One such way is to ensure that the systems have remote over-the-air access. All our systems are accessible over our own VPN created using OpenVPN. However, this access is possible only if the system is connected on internet. There will be times when the system gets disconnected for no fault of yours — as we have been experiencing recently with 2G networks of Airtel showing significant performance degradation across the country. At these times, a backup option of connecting to the system via remote access of a laptop at the installation site allows us to troubleshoot many of the problems without paying a field visit. Quick problem resolution often results in high customer satisfaction as well. In addition, we have built in some health parameters into our system. When the system is not accessible, an automated system goes through these health streams to understand the exact cause of the problem. If the network has been intermittent over the past few days, one of the health streams will show it and hence we can just take the action accordingly, even without taking remote access to the system. Eventual idea is to learn from all the field failures to build in enough redundancies as well as an automated troubleshooting engine that can form the backbone for scaling up field maintenance without scaling up the field force.
I have now covered all three aspects that I feel are critical for success of any IoT enterprise (at least in the near term). Just to summarise, they are:
In my last and final article, I will give a high level overview of how these three come together into creating successful use cases on which the stickiness and novelty of the IoT solutions hinge — Stay Tuned!
Originally published at bwdisrupt.businessworld.in.
