Bury and Dribble Out Big Data Analytics for Useful Results

Buried Dripping big data is best

Big data is best buried and dripping so people can make use of it in a tasty way. Photo © Damdeeso Dreamstime

The hype around big data analytics has reached epic proportions. It does have the potential to radically improve not only business processes, but outcomes. It can be a tasty addition to the IT menu. Unfortunately, for many manufacturers, the value has not matched up to the promise.

Why Big Data Analytics Is Challenging

Why aren’t production companies getting full value from big data?

  • Most people don’t fully grasp big data, so finding the right project can be a challenge.
  • Not everyone understands and invests in the variety of software capabilities big data demands.
  • Manufacturing is more complex than many other big data analytics applications, so many solution providers are not all well-suited.
  • Not all of the data is typically available, and unless the system is constructed carefully, incomplete data can lead to incorrect analysis and prescriptions for action.
  • Employees simply don’t know what to do with the results.
  • The business processes are not designed to use analyzed big data effectively, or don’t have time allocated to parse the meaning and create action plans.
  • People don’t want another set of data to add into their daily set of systems to use.

Making Big Data Useful

In recent conversations with Radhika Subramanian of Emcien and Chris Knerr of Marēana, companies that offer big data analytics, I got some new ideas on how to overcome those. They may be a bit counter-intuitive.

  • Automate: Both of these companies focus on automating as much as possible. “Results were really great when the downstream was an automated system. However, when a human was in the loop, it failed most of the time,” according to Subramanian. Rather than having a person need to understand exactly what to ask, and how to review a model and results, use automated data ingestion and context assignment, search, pattern discovery, and workflows.
  • Applications: If the solution set is configured to solve a particular type of business problem, rather than a generic big data analytics toolkit, it can more easily deliver the results you need.
  • Bury: It may be difficult to hear, but hide it from most employees. Don’t make big data analytics another major software system to learn – flow the analyzed results into other systems that they already use. Make those jobs better in a subtle way.
  • Dribble it out: Do not expect people to understand how to use a full set of results, even if it is right. In the ideal world, the results dribble out to drive individual tasks each person is doing in those existing plant and enterprise systems.

Two Different Big Data Companies

These two companies have very different profiles and approaches, so beyond those areas in common, some other topics that might lead to greater success arise:

Marēana is industry specialized for life sciences manufacturing. Serving only the highly regulated pharmaceutical, biotech and medical devices (where founder Knerr was for decades) means that industry expertise is baked into the product set. Their solutions are in search, liquidity and optimization. Marēana also offers proof of concepts as a starting point for most customers.

Emcien also starts with what the customer most wants and needs. Since they are buried, alerts from the system appear in a sidebar within other applications. It has sparked the new path toward task-sized bites of output that goes even beyond, to tell a person what to do based on its discoveries. The predictive machine learning enables prevention actions also.

Small Bites of Big Data

Success in this model requires big data analytics software that is at once very complete, and ready to be buried behind the scenes. The amazing big models in which patterns can be detected are the secret, but don’t use them as a flashy sauce on top, rather as the hidden gem inside.

If you have been struggling to make headway with a big data analytics project, consider how you might use these ideas to move forward with greater success. Let me know how it goes!

Tags: , , , , , , , , , , , , , , , , ,

No comments yet.

Leave a Reply