By 2020, research shows that the number of connected industrial devices will grow 285% from 2015. If you’re like some industrial leaders who feel IIoT is more hype than real, you’re taking a wait-and-see approach. But if you’re like many of the leaders we speak to, you recognize it’s not if you need to digitally transform your business, but when … and even more fundamentally, how.
Knowing how to get started and the right direction to take requires defining the right outcomes. This is especially true when industrial organizations look to data and predictive analytics at a component level to provide better insights. But data in and of itself does not impact an outcome on its own. Instead, it’s important to focus first on outcomes, and use data to help deliver those outcomes.
Starting With a Scenario
Let me illustrate. ACME, a fictitious business, has a very sophisticated manufacturing process where a large amount of parts are produced around the clock. Every operation on every piece on the shop floor is tagged from raw material to work in progress (WIP) to finished goods inventory (FGI). There are over 40 disparate data sources with varying level of details. The data sources deliver information related to parts, operations, manufacturing process, and parts cost.
One possible path ACME could take in its digital journey is the data-first approach — connect all 40 disparate data sources, stitch the data together, find a way to use all the data and develop analytics to extract value from the data. This is a bit like throwing a bunch of ingredients into a blender without a recipe and ending up with something that could taste awful.
However, a better approach is to start by creating a recipe — identifying and defining the outcome ACME is trying to impact. Having the right outcome first allows the organization to bring clarity to its transformation goals, and be able to more accurately measure ROI.
But even with a specific problem identified, ACME doesn’t know what data and/or analytics solutions it needs to solve the problem. One way it could tackle this is through a concept called Proof of Value (PoV), which is different from creating a Proof of Concept or a prototype. A PoV allows an organization to focus on finding the value within the data and the analytics solutions that leverage the data to drive desired outcomes.
The most effective type of outcome is well defined, and can show that an impact can be made and—most importantly—measured. For instance, defining an outcome such as “increasing capacity of a manufacturing plant by x%” is too high level since capacity can be influenced by many factors. A more effective outcome could be something like “increasing capacity of a single line for top two product categories by x%”. Or, “reducing downtime for a specific equipment to 1%.”
What we see in this scenario is that a data-first approach might sound good (“be data-driven” we’re constantly told) but may end up providing little value. On the other hand, an outcome-first approach allows ACME to crisply define the problem to be solved and create measurable metrics to map to, and then structure the data to support those outcomes. This means they can apply analytics more strategically to enhance business performance.
I will continue this discussion with another blog soon, where I’ll take a deeper look at defining the outcomes for ACME, and what you can learn from them.
Meanwhile, you might find this infographic from LNS Research helpful. It provides a summary of key insights into adoption of digital technologies to improve process improvements, production forecasts, continuous asset performance, and other metrics.