The Power of Data or how to get back on your feet with IOT

IOT, big data and intelligent usage of it (especially artificial) are again among my list of 2019's top buzzwords. Most of my contacts here have multiple touchpoints and many for sure have amazing and value adding business use-cases for utilizing associated technologies. However, from an IOT hardware perspective to the creation of more or less bigger sets of data to the intelligent and even autonomous drawing of conclusions based on it, I often see complexity being overrated and sometimes exaggerated when it comes to utilization in business. I personally do feel it often is not that complex if you break it down and apply it consequently. In this - my first - LinkedIn article, I will share some personal IOT learnings that did impress me and I try to revert those lessons learned back into a business context.

1.5 years ago I severely broke my knee on a skateboard accident (Yeah , I know - in my age and so on, but as for most DIY electric mobility attempts: No pain no gain - maybe I will write about those experiences in a second article). Torn ligaments and multiple crushed bones in combination with not being an active sports person (you don't consider riding an electric skateboard if you enjoy pushing it manually) gave me not the most positive recovery perspective. So, situation was clear to me and I needed to improve on various associated areas to get back on my feet. Identification of the need to improve should be the first step combined with DEFINING your target state! Mine was to walk again like "normal" (or at least like before my accident). Relating to business: Identifications of operational- or any other types of deficits that need drastic improvement is one of the first steps and should be fairly simple. E.g. just take any of your existing logistics processes - as soon as you are moving things, any lack of transparency will most likely lead to inefficiencies, losses or damages at one point or the other. Improving your specific current situation by transparently observing it anytime you want to is one of the most common use cases within IOT, so I wont elaborate on that further.

After defining my target, I took a stock take. Where do I actually stand today and what are the KPIs to MEASURE my progress against? For my challenge, this was quite some simple list of measurements. It basically did burn down to directly related parameters like maximum running distance and maximum running speed to the point my knee would stop me from proceeding (Starting from the point where running wasn't possible at all). Directly and indirectly influencing parameters have been average steps per day, stairs climbed per day, my weight, body fat, heartbeat (min, max, development over time) and some more. While in my case this was self explaining, business wise this can be the first challenge to tackle. Sticking with your originally defined target is key. The measurements you could select for e.g. a last mile delivery optimization are often far more numerous and measuring them often is interfering with other use case requirements like battery runtime, robustness, hardware mounting options and much, much more. Even if additional measures do open up new improvement opportunities, they very often can dilute your initial objectives. Sticking with your initial objectives can be hard, but to my experience it's worth it. Sometimes, adding upside potential through additional IOT sensors for later, already anticipated use cases might be fine, but in case of very tight project budgets or risk averse investment policies, this could kill your IOT case before you even started. At least this is what I observed in my professional life. On my "consumer case" I was lucky to already have been equipped with a real-time sensor on my wrist - a Garmin fenix3 HR - constantly measuring many of the parameters I was curious about. Since I could use a standard notebook for data analysis and because there was no need for setting up connectivity (e.g. GSM or GPRS based) data connections, I was good to go.

In business context, I experienced the ANALYZE phase as being very impressive. If you all of the sudden are able to view a whole supply chain transparently in real-time from end to end, this creates really eye opening events. For example, you can immediately see, why you are always lacking drivers starting half an hour before you have a shift change. Or, all of a sudden you are able to identify spots within your supply chain, where (geographically), when (timewise) and why (and here the fun starts) you really are in control of important process parameters like temperature, vibrations or even just process ownership. Optimizing routing of any of your assets becomes very clear if you can simply visualize movements as well as locations of stock in real-time. A deeper analysis to find underlying root causes in many of my business IOT applications wasn't even required at all. In my "knee case" I had a deviating experience, which also can be due to the fact, that I do not have any clue about sports. It took me quite some test runs and painful recoveries to learn that I need to be more patient and that I should consider indicated recovery times more seriously. On the other hand, analyzing my individual performances and comparing them with prior data sets did help me to better understand when I made progress and what things did influence this progress positively as well as negatively. So, one simple, exemplary IOT finding from my various analysis attempts was that data showed me some direct correlation between maximum running distance (until pain kicks in) and number of days between my last strong exercise as well as the weight trained on my knee during this exercise.

While in a business context, define- and measurement phases often are more crucial to get your IOT use cases on the road, for my knee-fitness the IMPROVE phase was the toughest one. Due to all data gathered so easily, frequent reminders and automatically monitored targets, it felt more easy to overcome my "weaker self" . If I felt, that I moved enough for the day/week/month, a short glimpse on the actual data often proved me wrong and I learned to improve my perception over time. Next to this essential learning on better estimating the current state the review of achieved milestones as well as overall target achievements really felt motivating to keep on track - especially if the often expected, immediate impact on my physical mobility /or fitness did not show up that quickly. In a business IOT implementation, my experience is that due to the often high investment business cases, the actions required for the improve phase are more singular-/one time than regular /frequent actions and therefore more easy and fast to be realized. Also, the potential for automation seems to often have a higher cost/effort impact here, too.

So, all set? Nope - assuming you reached your initial business targets you have to stay in CONTROL, too. Maintaining the technical setup, replacing broken hardware, taking care of IT platform or listener associated updates as well as spontaneous firefighting actions in case a volcano eruption significantly disturbs your country wide visibility are just one side of the story. Since this requires expertise and additional investments all throughout your business case, you should have an eye on those in the beginning. Recently I finally achieved the last of my 2019 "Knee-sub targets" (among others running more than 200km, riding bicycle for 1800 km, etc.) and I finally was able to survive a 10km run. Though my defined aim was to be able to move my knee again, there also had been a multitude of observed side effects like losing weight, reducing my idle heart rate or reducing body fat. Now the only thing left is to stay in control, continue to monitor my KPIs (latest until to the next test ride of my electric skateboard v2.0 that might throw me back to start again :-).

Closing thoughts: 

• Motivation and willingness to stick to initial targets was surely one crucial success factors for me and this also seems valid for any business IOT implementation as I learned in my professional life. If you give up too early on any IOT Implementation - e.g. because there are issues with the received data, your ability to interpret and analyze it or even with the reliability of the hardware - you will not reach the sweet spot of when your use cases starts kicking in.
• Methodology rules! As engineer and six sigma guy I might be biased on this one. Nonetheless, I was involved in many IOT cases and all those that succeeded have been using clearly structured development approaches and methodologies within them. While some cases did work well based on DMAIC, others went even better based on the classic waterfall development processes with very precisely documented requirements. Even if it feels that "agile" is still on vogue for any kind of project, don't attempt it in big IOT implementations! Rapid developed sprints might make sense for part of those integrations. Due to the high costs of developing/finding a good solution setup (based on combinations of hardware, software, telco and processes), costs for creating required MVPs and managing the scrum of interdependent requirements striving for your attention will create more than a headache - at least this is my experience.
• Reliable working hardware and data transfer is a prerequisite for all IOT cases I got involved in. The associated high investments as well as interdependencies between hardware configuration and communication intervals will make intensive testing mandatory. I can't emphasize enough on this one - if you do not have a hardware/communication setup that matches your requirements, your whole IOT case will blow up more sooner than later. For my "knee case", I was lucky with the reliable high quality and impressive long battery runtime of my existing smart wearable and there was no need for me to upgrade. Today there are even more advanced smart wearables available like the Garmin MarQ Driver. Those even support you in improving on your lap times on a race track like the Nürburgring - maybe an interesting option for fine tuning my electric skateboard v2.0 :-) So, thanks @ Garmin for helping me getting there!

So, what are your experiences on consumer IOT product related use cases and how they might transfer to applications in business context?

Looking forward to your comments and wishing you all a fantastic, happy and especially healthy start into the new year 2020 with lots of upcoming successful IOT implementations!!!

Disclaimer:

The views and opinions expressed in this article are those of the author Dr. Frank Josefiak and do not necessarily reflect the official policy or position of his employer Deutsche Post AG. Examples of analysis performed within this article are only examples. They should not be utilized in real-world analytic products as they are based only on limited and dated private source information. Assumptions made within the analysis are not reflective of the position of any Deutsche Post DHL entity. The author is also neither associated by any means with Garmin or any other party nor payed for this article in any way by anyone! Links to external pages are used as references to provide further information and the author cannot ensure availability nor correctness of the links.