You Don’t Know What You Don’t Know

September 04, 2019 by Otmar Foehner

My day job for the past 30 years has been helping people with their sensor data. In my time off I tinker with and build PCs, try my hand at some home improvement projects and stay on top of new technology developments through the consumption of copious amounts of online tech news website articles and YouTube videos. What I have learned—and keep being reminded of time and time again—is that there are plenty of things that I don’t know. And, that sometimes I’m not even aware of things that I don’t know. Let me explain.

I recently did a little DIY project for my wheelchair-using stepdaughter—adding a grab bar to a bathroom wall so she would be able to have a little more independence when using the shower in her bathroom. The installation required me to drill a pretty sizable hole into the drywall to install the required hardware, as the grab bar needs to hold the full body weight of an adult person.

Now, I have drilled plenty of holes into walls in my time and I know that I always need to check for water pipes, power lines, etc. I even have a pretty nifty “stud finder” that helps with making sure I don’t accidentally mess up. I also have a good drill, a great collection of drill bits, and a suite of other tools that are all good at doing their respective tasks. In short, I have the right tools to do the job.

I made sure to “measure twice”, look for water pipes, power cables, etc. and started drilling “the hole”. “The hole” was great - perfectly round and the right size for the installation of the hardware for the grab bar. What I didn’t know was that there is some requirement in the local building code that basically had a hollow space behind the wall, and then another wooden structure behind that hollow space. What I also didn’t know, setting out into this small project, was that the hardware wouldn’t fit in the hollow space behind the outer and inner wall. I’m not a builder and neither have the knowledge or the experience of how these structures are built. In the end, with one big hole already in the wall, I decided to hire a professional to install the grab bar, fix the sizable hole I had created, add the texture back on top of the hole and make everything look professional – I wasn’t going to drill more holes without knowing what I would find behind those wall. The guy I hired (a professional) had everything done in 30 minutes, and the grab bar is installed and secure for my stepdaughter to use.

You may be a better handyman than I am, but the fact remains that we often don’t know what we don’t know when we dive into a project. When we start a new project, we make assumptions based on previous experience (remember how I had successfully drilled holes in walls before) and don’t take into consideration things we don’t know yet (due to lack of experience or expertise). That approach can work for small projects (learning by doing), but it can easily get expensive and time consuming to fix mistakes when this happens in larger projects. For important and investment heavy projects, it makes financial and business sense to hire someone “who knows what I don’t know”.

So, why is this story on a blog for a company that provides solutions in the sensor data management business?

Sensor Data Management is a very complex topic that requires expertise and knowledge in multiple areas: 

It requires an understanding of the engineering process, the challenges that engineers face with their testing and their data and the activities involved in turning that data into useable information. 

It also requires a good understanding of IT processes and technology, issues like compute power, network bandwidth usage, scalability of processes, database technologies, user and data security, etc. 

And finally, it needs a good understanding of the software development process, everything from creating an architecture that allows for great performance and easy expandability to long-term maintenance, technical support and training. 

A Sensor Data Management system is a long-term investment that will be in use for many years, so any solution should be deployed with future changes, requirements and expansions in mind. These changes involve hardware infrastructure (back in 2009 we had PCs or laptops with 2 processor cores and 2GB of RAM), software infrastructure (remember 10 years ago: We were using Windows XP, celebrated the release of Windows 7, then Windows 8, Windows 8.1 and currently Windows 10) and of course the requirements of the userbase. 

Who is going to make sure that the system runs on the hardware and software we’ll have access to 10 years from now? Who is responsible for system maintenance and bug fixing? Who will add features in the future that are requested by the user base or demanded by new processes? Who will ensure that the system scales with future needs?

A Sensor Data Management System is complex enough to design and build for people who have years of experience and deep expertise in the subject – it’s not something that a company can DIY in a few weeks/months and hope to benefit from in the long-term. It requires a long-term strategy to stay useful, as well as insights into where technology is going and how it will benefit specific applications. 

There are measurable and undeniable benefits in having a sensor data management system. Anything from creating clean and verified data, to sharing peer-reviewed processes to reproducibility of analytics and reporting results. Studies have shown that customers who use sensor data management systems reduce their time-to-insight dramatically, while saving millions of dollars over time due to better test equipment utilization, faster analytics and reduced time to market through improved testing efficiency. Elsewhere on this website you can find actual examples of companies that have benefitted greatly from their sensor data management systems, both financially and by reducing product development cycle time.

So basically, the lesson here is to get professional help involved before you drill “the hole” in the wall – the wall was a relatively quick and cheap item to fix; fixing an incorrectly designed and implemented sensor data management system will be a much more time consuming and costly endeavor.