Proactive and Preventive:
Scalable Analytics for Predictive Maintenance


Data Generation Is Not The Problem

The opportunity for scalable advanced analytics in predictive maintenance.


predictive maintenance on cloud or on-prem

The ability to manage a variety of asset types and avoid unplanned downtime is no trivial task. An entire operation can come to a halt if just one critical process, asset, or machine fails. But too often organizations are forced to rely on outdated, incomplete, or inaccurate data to make decisions resulting in inefficient operations.

predictive maintenance for service issues

The #1 driver for investments in predictive maintenance among service leaders is the need to have a faster response to product quality and service issues.

predictive maintenance support

As more products and equipment are connected providing a wealth of data points, data availability is no longer the primary problem impacting organizations’ ability to improve maintenance, service, and support.

data analytics for predictive maintenance image

Siloed processes and data, lack of a scalable, high-performance data analysis platform, and analytic methods focused only on past performance hinder the ability to shift from reactive maintenance to more proactive, predictive, and prescriptive models.

Source: IDC Manufacturing Insights Product and Service Innovation Survey, 2021

Shortcomings of Break/Fix and Reactive Service Models

Unplanned downtime is more than just a nuisance to facilities, plant, and field service operators


dot-pattern
0%

of manufacturers still characterize their service, operational, and maintenance approach as break/fix and reactive.

0%

of manufacturers find it very difficult to address challenges resulting from a lack of process automation, inefficiencies in workflow (e.g., manual, duplicative, siloed work), and delays in transforming data from information to knowledge.

0%

Only 14% of manufacturers rate themselves as excellent at surfacing actionable information to all users in the flow of work.

The lack of insight, based on data, to predict failures and to plan for downtimes and outages can cost an enterprise millions of dollars and negatively impact customer experience.


Sources: IDC’s Future of Intelligence Survey, 2021 and IDC Manufacturing Insights Product and Service Innovation Survey, 2021


more on-database-analytics-for-predictive-maintenance

The Promise of Predictive Maintenance

Turn smarter equipment, assets, and products into better outcomes


The Right Outcomes Need Relevant and Timely Insights

Q: What are the top five drivers for your organization’s service life-cycle management efforts?

(%of responses)

56%

Metrics

Improve key customer metrics like NPS, customer effort, and CSAT

49%

Knowledge

Capture and make accessible service knowledge and best practices

59%

Response Time

Faster Response to product quality & service issues

51%

Collaboration

Improve collaboration between cross-functional teams

52%

Performance

Use information from actual product and asset performance

57%

Capabilities

Establish more capabilities around remote service, collaboration, and resolution

Faster Response to product quality & service issues

59%

Establish more capabilities around remote service, collaboration, and resolution

57%

Improve key customer metrics like NPS, customer effort, and CSAT

56%

Use information from actual product and asset performance

52%

Improve collaboration between cross-functional teams

51%

Capture and make accessible service knowledge and best practices

49%

Source: IDC Manufacturing Insights Product and Service Innovation Survey, 2021


dot-pattern

The Right Outcomes Need Relevant and Timely Insights

Q. What percentage of products your organization currently manufacturers are considered “connected” (i.e., have a unique IP address and software within them to enable service and product performance information to be communicated over a wireless network)? What will that be in three years?

44%

Today

60%

3 years
from now


Improve service and maintenance through data-driven, near-real-time insights

31%

of manufacturers
are seeing a shift to:

More predictive and preventative service

60%

of manufacturers
are seeing a shift to:

More prescriptive service and maintenance


The future of maintenance is:

The ability to predict a failure and just as importantly trigger the appropriate response while allocating the right skills, parts, and tools to ensure preventative resolution


A Modern, Scalable, Data Analytics Platform

Support for data exploration and decision automation

Decision Making, Data, and Analytics Usage Patterns

Data exploration
and investigation

Key driver identification

Guided root-cause analysis

Data exploration and investigation is about helping users understand and explain what happened and why it happened.

Decision
Automation

Conditional decision automation

Algorithmic decision automation

Decision automation is about automating tactical decision making in the flow of operations.


Consider the following predictive maintenance decision variables:

Scope

The breadth of the impact of a given decision. Does it impact a single asset or many; a single activity or one whole process or multiple processes?

Latency

The time window or time internal within which a decision needs to be made or an issue needs to be resolved.

Variability

The extent an issue is pre-defined vs. ad-hoc. Is this a regularly or consistently reoccurring decision or one that needs to be made rarely?

Ambiguity

The extent to which the issue at hand is open ended. How open to interpretation is data needed to make the decision?

Risk

The monetary value at risk of the decision.



Requirements of a modern data analytics platform for predictive maintenance

Minimization of data movement

Pre-built support for commonly used analytics, including support for AI/ML algorithms

Ability to extend analytic capabilities with customized and unique algorithms using data scientists’ preferred languages and tools

Availability of cloud storage APIs

Support for, and integration between relational data warehousing and data lakes, including those based on open-source software

Support for standard development languages and skills (e.g., SQL, Java, C++, Python, R)

Support for real-time service level agreements

Separation of compute and storage to enable flexibility in matching technology resources and costs to variability in analytic workloads

Support for Big Data processing requirements, including terabytes per second ingest/egest rate, and exabyte storage capacity.


Source: IDC, 2022


What steps should organizations take to solve the problem or seize the opportunity?

Insights at Scale

Complexity of Maintenance Operations Doesn’t Require a Complex Response

predictive maintenance data and analytics strategy

Develop a long-term data and analytics strategy that considers various decision-making patterns and related data and analytics IT requirements – both for streaming and batch data processing and analysis at scale.

predictive maintenance to assess data quality

Assess data quality and availability guaranteeing data-driven decisions can be made.

predictive maintenace partners

Consider IT partners that provide a modern data, analytics and AI platform that is extensible and leverages a broad partner ecosystem as no single vendor can do it all.

Don’t expect a single technology to address all requirements. SQL-based, columnar, MPP analytic databases have a role, so do data lakes and streaming data processing software, and a range of upstream and downstream data integration and data analysis and visualization tools.

predictive maintenance value

Select appropriate data and analytics technology that is not just about finding solutions with the most compute power or storage capacity (and flexibility); also consider security, support from solution provider, and overall total cost of ownership.

Source: IDC, 2022


View PDF of infographic created by IDC Insights.


How can predictive maintenance help your team avoid unplanned downtime?

Learn how a modern, scalable data analytics platform can remove the guesswork and unnecessary labor and expense of a traditional, scheduled maintenance approach.

Vertica is the analytical database with the best value for the highest performance on any data analytics, at any scale, anywhere.

Best Value

Compressed storage, efficient processing, fungible license

Any Analytics

BI, time series, IoT, geospatial, machine learning

Any Scale

Terabytes to petabytes

Anywhere

On-prem, clouds, hybrid, containerized

Learn more about using Vertica as part of a modern, scalable, data analytics platform for predictive maintenance