How to capture maximum value from industrial IoT solutions

Data is everywhere and we take this for granted. After all, we create more than 2.5 exabytes (EB) of it every day, and researchers predict that there will be 163 zettabytes (ZB) (1 ZB = 1000 EB) of the stuff by 2025.

Nevertheless, do you know:

  • A typical oil rig has around 30,000 sensors but managers examine only 1% of the data.
  • Often, 80% of data scientists’ time is spent simply in preparing data for analysis.
  • Less than half of the structured data of an organization is actively used in decision-making. And less than 1% of its unstructured data is analyzed or used at all.

Do we take that for granted?

Whether we accept it or not, data is only numbers if it cannot be used beneficially to modify future actions. Logic dictates that a company should develop and nurture the process that distills data into actionable, useful information. Such data-driven information provides the basis for informed business decisions and the creation of value. That’s the logical objective. Let’s see how your company can capture the value from an Internet of Things (IoT) setup.

We have defined a loop to capture the value from IoT applications.

Common business use cases:

Irrespective of the use case, the principles for capturing values from an IoT project are very similar.

Let’s say you would like to apply a predictive maintenance solution in your facilities or assets. In that case, let’s see how the loop mentioned above can ascertain value capturing.

Identify the outcome

As the name suggests, identify the desired result, in this case, predictive maintenance. The catch here, however, is in the understanding that you must be able to take action on what you predict. Otherwise, the prediction, as a result, is worthless. For example, predicting that a heating and cooling unit will fail in the next few days is not useful if you can do nothing to prevent it.

Hence, whether you would like to:

  • Improve uptime
  • Reduce maintenance costs
  • Extend lifetime of an aging asset, or
  • Minimize safety, health, environmental and safety risks.

Make sure you have everything when it comes to acting on it so you get the best results. After deciding the outcomes and ensuring that all arrangements are handy, the next step is to define the data sources with why.

Select the data sources

It takes serious efforts to identify appropriate data sources that are perfectly suited to address your business problems. These efforts can be channelized with the following four components:

  • Originality:Whether you work with internal or external data, always make sure it comes from the primary source. You will never know if the methods were faulty, if the sample size was too small or if the questionnaire was biased without reviewing the primary data yourself.
  • Comprehensiveness:Your data source should be sufficiently informative to obtain a big picture and a suitable context. It would, therefore, be helpful to match your data with your historical data.
  • Timeliness:Data changes rapidly. So always use the latest available data.
  • Reliability:Check that your selected source is relevant, legitimate and as impartial as possible.
    You may need at least the following data for predictive maintenance:

    • Condition of an asset
    • Usage of the asset
    • Maintenance history
    • Environmental data

    Regardless of the sources you select, you should check if you have the original, comprehensive, current and reliable data.

Data capture and combine

Connect all your data to a single location, whether from the expected behavior or failure logs.

You are now ready to lay the basis for predictive analysis. This involves:

  • Condition of an assetSince data can live in many different places, connecting it to a single system is a crucial step. Data may need to be moved in some cases, but it is often a matter of connecting a data source to an analytical system. Since you are likely to deal with large volumes of data, an analysis tool that can handle big data is essential.
  • Normalization of data: Normalization of data can take time, but it is crucial, especially if you rely on anecdotal information from your repair teams. Normalization of data also helps to improve your analysis, accuracy and validity.

Model implementation and validation

Your model can be implemented by using machine learning techniques with understanding about how much in advance your maintenance team needs to respond to a prediction. Use stack-rank models to determine which model is best here for your HVAC unit failure timing.

Validate your model using a data subset. This guides your thinking about what outcomes can be accomplished and gives you a basis for evaluating the analytical results. Overall, validate your model with conditions that indicate future equipment problems.

Model train and retrain

Use your model on live data and see how it works in real-world conditions. Based on the data collected during the real-world pilot, refine your approach.

Train your model with possible scenarios and conditions like when to trigger an alert or action (e.g. automatic order of a replacement part).


Adjust the maintenance processes, systems and resources to operationalize the model to respond to new insights. Make continuous improvements with those insights.

You are now ready for wider implementation and capture of value. Integrate this predictive maintenance approach into your existing facilities or equipment to increase efficiency, reduce costs and develop your business.

In summary, if your IoT value capture strategies do not meet your expectations or if you are uncertain about how to capture value from an IoT implementation – our IoT readiness workshop can help you.

The what, why and how of IoT business models

With the increase in adoption of the Internet of Things (IoT), companies from all industries now have a question – which IoT business model and application is right for them? Check out this whitepaper to know what, why and how of IoT business models.

IoT Readiness Workshop Program Guide

In this whitepaper, we have covered some important elements you should consider when looking to transforming your business model.