Many of today’s data analytics processes are cumbersome at best. You have a system collecting data, a tool to translate that data into a visual report, an analyst that interprets that report, and an executive that then makes a decision based on that interpretation. Naturally, this approach takes time, and can often delay a company from making rapid optimisations that could save them both time and money.
For data to be truly valuable, teams should be able to quickly interpret it and make changes accordingly. This doesn’t just mean putting it into visually appealing reports and interactive charts — it means making the data actionable so that you can optimize customer experience and improve business processes.
Let’s take a step back and put this into perspective. When we first start working with our clients, and they ask us to create an analytics architecture that captures and processes specific data sets, we ask them a question. “What do you want to use this data for?” Basically, we want to understand what actions the data will inform or prompt.
Sometimes, our clients don’t have an answer, and that’s usually an indication that they don’t actually need to be collecting that data — or that they need help connecting the dots between their data and their operations. Meanwhile, those that do have an answer tend to limit themselves in terms of how much they think they can get from their data.
The truth is, the right data can be a crucial part of designing automated systems that save you time, money, and headaches. Instead of just looking at data on a chart and coming up with interpretations, you can use machine learning to raise red flags and conduct predictive analytics that keep you one step ahead at all times.
For instance, one of our customers had IoT sensors on their HVAC systems, and they wanted our help to create a dashboard so that they could see real-time performance data and take action as needed.
Working with them, we uncovered a better, more actionable use of the data that would provide emergent insights on when a machine was broken or overheated — and send a notification to the appropriate parties. This means that instead of spending time checking the data to see if something was wrong, teams can focus on mission critical tasks and only respond when notified by the system.
Looking into the future, the potential for actionable data is endless. We expect to see fewer analysts, but more insights that drive automated action. Moreover, it’s likely that companies will be able to rely on external, openly available data sets and pair these with their own internal data to predict outcomes and work proactively.
Let’s go back to the HVAC example. A building or real estate complex with multiple HVAC systems could be continuously analysing data from the machines themselves, from weather forecasts, and from other HVAC systems in nearby buildings. These data points would all feed into their analytics platform, providing automated decisions on when to run the HVAC system, when to store its energy, and when it might be underperforming or in need of maintenance.
As we move in this direction, it’ll be important to strike the right balance and not collect more data than we need. With data becoming more and more available, this is a risk — and it could put a burden on operations if not done properly.
These are exciting times. Organizations have an opportunity to build and execute robust data analytics strategies that optimise their operations in a way that really wasn’t possible five years ago. At InTWO, we’re excited to work with our clients to define what actionable data looks like and how it can shape their business models.