News, Insights, and more on Industrial IoT
Having a data challenge? You're not alone.
Sign up for our blog and receive a free eBook on the 10 data challenges in Industrial IoT and how to overcome them.
What do we mean by missing data? Basically, it’s valuable information about our people or assets that just isn’t being collected because the perception is that doing so would cost too much or take too long. To be successful with digital transformation, however, industrial businesses must first understand how to monitor assets and collect that is simply missing.
What causes missing data?
Businesses often choose not to collect what they need because the costs of doing so are too high. Given that many industrial companies have struggled to generate meaningful value from IoT deployments thus far, it’s hard for operators considering digital projects to make new investments to plug their data gaps.
Some resort to machine learning algorithms or data simulations to try to forecast outcomes when data is missing. However, these alternatives can’t fully replace the objective truth, especially if it’s knowable. While forecasting capabilities are valuable, they don’t fix the underlying problem.
Getting the most out of the IoT requires that organizations collect massive amounts of high-quality data. If operators don’t do this, they leave significant intelligence and opportunity on the table. Analysts can’t make informed decisions based on real-world events, and businesses are unable to identify potential efficiencies within their operations.
Who is affected by missing data?
Missing data has significant upstream and downstream impacts on industrial companies.
Those immediately affected are field workers, who have to manually collect what smart devices don’t. For obvious reasons, this is a time-consuming and poor use of resources, especially in industrial markets where assets are dispersed, like oil and gas or mining. Field personnel who don't know how to monitor remote assets effectively can result in bad or incomplete data, and executives can’t expect their field employees to observe and measure every process in real-time, particularly for those involving dangerous applications.
Even if workers in the field do understand how to monitor their assets, they are unlikely to get data in time for processing. They may also miss anomalies that sensors would catch or even misrepresent data in some way. As a result, downstream professionals, including superintendents, engineers, directors, and executives, may draw incorrect conclusions.
What do we need to consider?
When it comes to addressing the missing data problem, businesses should consider two opposing questions:
Calculating how much it would cost to collect missing data isn’t difficult. Project architects simply need to determine what they don’t have and how they would go about getting it. Between choosing the right edge devices, orchestration apps, and related services, deployment experts should be able to accurately quantify what investments their companies would need to make to this end.
For example, a digital project manager or a superintendent could get an all-inclusive quote on an IIoT platform from a full-stack vendor, and get very precise cost data, down to the dollar-per-byte level. Important note: be careful of those hidden costs that some vendors don’t mention, including installation, maintenance, connectivity upgrades, or premium support.
On the other hand, calculating the cost of not collecting missing data is harder. Analysts have to make assumptions and build models based on comparable applications or historical data, a process that inherently carries some risk. This isn’t just an issue of evaluating the costs of equipment or connectivity. Rather, it’s an assessment of the inherent costs of various business processes and how they impact various aspects of the profit and loss statements of a company.
There may be other, more subjective costs that digital executives and managers must consider: Are our competitors collecting this data and edging us out on new opportunities? Does our workforce expect (or demand) that we use technology to keep them safer? Does growing concern over climate change mean new regulations will require more efficient data collection?
Put simply: the cost of not collecting the missing asset data is the opportunity cost of not achieving a new business model which might improve the value of the business.
Fortunately, determining this cost is becoming easier. Machine learning and artificial intelligence capabilities are equipping analysts to better forecast “what ifs” related to missing or hypothetical data. More and more companies are taking the leap to double down on digital transformation, trading some financial risk for the opportunity of revolutionizing their business.
All the while, solving the missing data problem is becoming more important. Macroeconomic pressures, such as competitive dynamics, regulatory policies, and disruptive models, are pushing businesses to pursue digital transformation. Those who innovate and learn how to monitor remote assets with digital technology will solve the missing data problem quickly and position themselves for long-term success.
Based on our assessment working with hundreds of industrial companies, it’s most likely that the cost of not having IoT data exceeds the cost of collecting it, especially as the price of edge computing and network connectivity continues to drop. Technical or environmental limitations can certainly make some data collection economically unviable, but we do know that making this discovery is near impossible if you aren’t even aware of the overarching issue.
What’s at stake for your business?
With this context in mind, consider the following questions:
These questions can jumpstart a vital conversation within your organization around how to get more out of future IoT deployments.
At WellAware, we’re passionate about solving the industrial data problem. We understand the nuanced challenges of missing data, as well as how to address them effectively for any company in any industry. To learn more about our approach, contact us today.
Also, stay tuned for our next Data Problem blog series post coming soon: “Bad” Data!