How Actionable Insights from IoT Data Analytics Can Give You Positive Business Benefits

How Actionable Insights from IoT Data Analytics Can Give You Positive Business Benefits
How Actionable Insights from IoT Data Analytics Can Give You Positive Business Benefits

This is the first part of our series of posts about how the IoT can help to reinvent your business. In this part of the series, I will show you:

Is there a problem with IoT and Big Data?

Going back to our earlier tale about the school and the Smart Board, if we analyze what went wrong we find this. When your organization makes the decision to begin an IoT initiative, you generally look at a lot of things. You look at the features of the devices and the deployment, as well as the operational implications of using those devices. This, in itself, is an important decision and one that requires a lot of expertise to make well.

But, IoT projects are about more than deploying smart sensors. These devices also generate vast amounts of data. Understanding how to use this data is as important as choosing the right devices or deploying and operating them well.

Creating competitive advantage through data

Many business leaders think of IoT initiatives as setting up devices to help them make better business decisions. However, the real competitive advantage does not come from deploying and maintaining devices. True innovation is in the effective use of the right data generated by these IoT devices. To create a real competitive advantage, you must:

  • integrate those IoT data with your other data sources and thus create a comprehensive, real-time overview of your business,
  • translate the insights gathered with the solution to everyday business operational tasks,
  • adapt your way of thinking about making business decisions.

The key is not to use your IoT solutions as a”low price, simple projector”! Modern big data platforms provide a vast amount of features on which you can build your innovative business decisions – use them!

And, this is where “actionable insights” enter the IoT equation. Actionable insights offer us a powerful way to get a great ROI by turning data into effective actions. Utilizing IoT data to drive decisions is about generating “actionable insights” from these data.

What is Meant by Actionable Insights?

This is, in my opinion, a crucial part of delivering an effective IoT project.

Gerenating actionable insights through the IoT

Creating actionable insights, in theory, is really straightforward. The solution should:

  • ingest the data,
  • apply data analytics on it and then
  • you just have to process and build those results back into your everyday work.

But when it comes to building such solutions, challenges have to be faced. This includes, how do I:

  • Scale my solution as I scale my IoT initiatives; operate such a solution?
  • Decide on data and how long to store?
  • Integrate my data sources and analyze all these data?

This is especially true when it comes to real-time data analytics.

Other questions include:

  • Do I need Machine Learning know-how in my projects and if yes, how do I access such know-how?
  • How do I make decisions based on the analyzed results?
  • How do I activate and use the decision to drive my everyday business operations?

Three Steps to Actionable Insights from Your IoT Project

According to analysts, Forrester, 74 percent of businesses want to be data-driven. To do this well, they need actionable insights. However, most of the businesses struggle with realizing the real return on investment in their data-driven projects. That only comes from unlocking the power from generating actionable insights.

There are three steps used to generate actionable insights from collected data. These allow you to achieve an ROI on data-driven projects, thus on your IoT projects as well:

Three steps to actionable insights

  1. Ingest and Transform: Preparation of the data is the first step to precise data analysis. This can be a challenge in an IoT project. As you’ve probably experienced, IoT continues to evolve and morph with no true standards among the various technologies and vendors. So this variety adds a factor of complication to the whole ingestion process. The quality assurance of the ingested data and the transformation of it create the right format required by the next stage, which is –
  2. Analyze and Visualize: Large data sets add complication to data analysis and affect precision. Machine learning is a technique that is used to predict outcomes from data and find hidden patterns in data. Visualization tools help human beings to understand the results of the data analysis by presenting them in a graphical and easier to understand way.
  3. Process and activate: Visualization is a great start in separating out, at-a-glance, those insights that are truly actionable, and those that are not. Insights that you process as ‘actionable’ are valuable. Therefore, optimizing your insights during a processing part of the cycle will drive better business decisions. An actionable insight must have certain requisites for it to ‘tick the box of actionable’. Criteria that point to an action include:
    1. Relevant: to your business – the insight is only of value if it fits with your business strategy (or if it can, at least, offer insight in a strategy upgrade) or helps to improve your everyday operations’ efficiency
    2. Contextual: Is there an obvious anomaly in the data? Do you need to collect more data within a specific context?
    3. Fit: Hence, does the insight fit with the environment in which it was generated?
    4. And there is a 4th criteria which is optional, but if achieved, that could make a real advantage: it is not only based on understanding what has happened in the past, but is based on what could happen in the future.

Kicking Actionable Insights Up a Notch

Above all, being insightful with IoT generated data will lead to better and more relevant business decisions. However, the speed of achieving relevant insights is also important. Analyst Gartner has stated that payback timelines for IoT initiatives need to be, ideally, less than one year and anything over five years enter a “danger zone” of never achieving pay-back.

Let’s suppose you were able to improve your business efficiency everyday with 0,1% on average. We have in Hungary 250 working days in 2019. So this could lead of an improvement of 28% in your efficiency during next year. This is where actionable insights show their enormous power.  On the other hand if your business fails 0,1% daily it may lead to 23% decline by the end of the day.

But therefore you need to change the traditional way of thinking on how you come to business decisions. In my opinion, in such a highly changeable environment there are no good or bad decisions at a certain point in the time, aside from the ones which are explicitly good or bad. There are only decisions that are based on a mix of environment/context information, data, and intuitions. One can wait until all the information is collected and analyzed for long periods of time and try to decide safely, but especially in the decisions where data can be fast obsolete this is too slow.

What’s much more important is to make good enough decisions and then iterate – this brings you to that 0,1% improvement of business efficiency described above. Measure the impact and if needed adjust your business operations with new, good enough decisions already on the second day of your previous decision. To be able to do this, you need the right platform, the right information and the right decision making processes. And just not forget to have a team that is agile enough to adjust its operation to the new experiments. A team that is willing to take small risks frequently and learn from mistakes.

Looking ahead

In the second part of this series, we will take a look what leads to delays in scaling an IoT big data initiative and how to mitigate those delays to overcome 2 out of the above 3 challenges: the right platform and the right information availability for your actionable insights.

Interested in learning more about Big Data?

5 Good Reasons to Move to a Cloud-based Data Warehouse
Data Warehouse in BigQuery — Tracking Changes In Dimensions
How to Boost BigQuery Performance
Take Control of Your BigQuery Costs