How data management will make (or break) your IoT strategy

Your data lives in ways that can significantly impact your business. And because data fundamentally takes on a life on its own, there are questions you need to address from the very start of your IoT journey. Questions like:

• How much data are you going to capture on the Edge?
• How much data are you sending and transmitting to the cloud?
• What are you going to do with your data?
• How long are you going to keep it?
• Should you archive it when you no longer need it?

Below, we explore why data management is key to a viable IoT strategy, and how you can take the first steps in managing the data tidal wave. (For more information on the overall IoT Journey, please read my other blog post, Four steps to get the most value from your IoT data.)

Often, companies experiment with their IoT strategy before they launch full-scale efforts. A proof of concept (POC) is a low-cost, low-risk approach that can help you refine your strategy. However, many enterprises that have implemented a successful IoT POC or pilot study are surprised as they shift into production model. Difficulties arise the project scales up. The data streaming from a few connected devices may be manageable, storage can become an issue as more devices come online.

It isn’t feasible to simply hang on to all the data generated by your connected devices indefinitely. For one thing, storage costs would soon spiral out of control. For another, there’s little point in capturing data just for the sake of having it.

Instead, your data needs to help you achieve something specific. Like, solve a problem, improve operational efficiency, streamline equipment maintenance, or reduce waste. If you really want your data to work for you – to help you identify patterns, trends and areas for improvement – you have to understand how to manage that data.

Imagine, for example, that you want to understand how a building’s lighting consumes energy. Instrumenting your building is a good first step. But a connected building collects real-time data about lots of factors – from ambient pressure, to altitude to temperature. Most of that information isn’t critical to understanding how lighting impacts your energy bills.

However, what would be useful to know is whether or not the lights are constantly on in empty rooms. Or whether they could be dimmed on a bright day. Data management is the process of taking the overall available data and refining it down to a few, specific metrics. An IoT platform should help you to set and maintain these parameters, and store data accordingly.

When you are trying to solve a specific problem, as with our lighting example above, you need to understand how different data sets correlate to one another.

For example, you need real-time data from your IoT devices to tell you whether the lights are on or off. You can also learn if a room is empty or full, or if the natural light in a room is sufficient on its own. That information belongs in short-term storage, where it can be easily accessed.

You also need access to some historical data to help you spot patterns. When you experimented with automatic timers on the lights six months ago, what was the impact on your energy costs? When is your facility most busy, and when are the quiet periods? Is this change seasonal? Depending on your industry, you may only want to dip into these data points every few months. So it makes sense to store them in a longer-term analytics warehouse. There, costs are lower than a short-term facility.

Performing advanced analytics on both real-time and historical data means you can understand more holistically how your facility is performing. It also enables you to proactively implement solutions before problems arise.

Splitting your data in this way also gives you the best of both worlds. As you connect your devices, your IoT data comes to a short-term store. There, you have the speed and performance you need to handle this real-time information. Other, less immediate data can go to long-term storage, with more capacity, and at less cost.

When you’re developing your IoT use case, you have to consider your data strategy. That includes choosing the right platform, too. You need a solution that enables you to optimize data storage, and I recommend finding a partner with a multi-tier data approach. That way you can optimize costs and achieve a better return on your IoT investment. You can also

If you’re interested learning more about platforms, I invite you to read a recent Forrester report: The Forrester Wave™: Industrial IoT Software Platforms, Q3 2018. It’s a 24-criteria evaluation of industrial Internet of Things (IIoT) software platform providers.

About the author: An engineer by training and a lifelong technology enthusiast, Jiani Zhang is the Program Director for Offering Management for the IBM Watson IoT Platform. In this role, she helps lead customer engagements and guides the development of the Platform technology, both of which help clients realize business results. Previous to this role, Jiani led an offering strategy and management team focused on Industrial IoT. And to round out her IoT expertise, she also served as an original member of the IBM IoT leadership team. Her technology expertise runs from product design and development, to management and consulting.

Jiani holds a B.S. in Electrical Engineering and Computer Science from University of California, Berkeley and an M.B.A. from UCLA Anderson with emphasis in Technology Management.