Advanced Industries Practice

Smart quality in
advanced industries
Advanced-industries companies have the unique opportunity
to implement an approach to quality that goes well beyond execution
and compliance.

by Álvaro Carpintero, Ulrich Huber, Evgeniya Makarova, and Heiko Nick

© Tashi-Delek/Getty Images

March 2021
 Technological advancements have enabled a best smart-quality use cases and technologies
fundamentally new way of delivering quality. will vary by sector. When companies in advanced
Under this approach, companies view the quality industries follow the smart-quality approach,
function as a partner and coach that delivers leaders must adapt it to suit more complex and
value, not just a business cost. This perspective deep supply chains. They should also place greater
helps them integrate quality and compliance emphasis on improving discrete manufacturing,
into regular operations while enabling speed building global service networks that address
and effectiveness. quality in the field, and complying with the
regulators and standardization bodies that govern
The smart-quality approach can be applied across advanced industries. New technologies could
multiple sectors, and a recent article described1 reshape day-to-day quality operations across the
how some pharmaceutical and medical technology value chain of advanced-industries players (see
companies have implemented this framework. The Exhibit 1 for select examples).

Exhibit 1

Smart
Smartquality
qualityintegrates
integratesnew
newtechnologies
technologiesacross
acrossthe entire
the value
entire chain.
value chain.
Smart-quality approach across stages

Develop Plan and source Make Deliver and serve

Agile and digital in Digital supplier quality Digital quality-assurance Real-time product monitoring
NPD/NPI1 and risk management processes enhanced with AA ● Sensor-enabled product-
with AA2
● Digital-maturity ● Digital standard work and quality monitoring and
management ● Real-time supplier batch records review defect prevention
● Automated quality oversight ● AI3-based problem ● Automated recall
gates ● Incoming inspection solving with natural- preparation and execution
● Software testing reduced or eliminated language processing Digitized issue and repair
and validation ● AA for joint Automated quality and management
Real-time customer root-cause process control using AI
● ● AA for field-issue
feedback problem solving ● Automated visual inspection identification and resolution
● Automated in-line and ● AA for optimized repair
end-of-line testing with procedures
robots/cobots4 ● AI-based compliance
● RFT5 improvement and tracking of service points
defect prevention with AA and dealerships

Agile digital quality-management system

Agile and digital in NPD/NPI1
● End-to-end connected product data and digital twin for product lifecycle management
● Automated quality-performance tracking and review
● Digital quality training and learning management system
1
New product development/new product introduction. 2Advanced analytics. 3Artificial intelligence. 4Collaborative robots. 5Right-first-time.

1
Álvaro Carpintero, Tacy Foster, Evgeniya Makarova, and Vanya Telpis, “Smart quality: Reimagining the way quality works,” January 25, 2021,
McKinsey.com.

2 Smart quality in advanced industries

Generating value through five days of training with approximately 1,000
smart quality anomaly-free images. The algorithm delivered
Companies that have adopted the smart-quality human-interpretable pictures of the anomaly
approach have demonstrated that it improves without requiring an exhaustive training data set
quality far more than traditional measures. For with all possible defects. For each image analyzed,
example, smart quality has enabled organizations the model calculated a confidence score, allowing
to reduce their total cost of quality by up to employees to focus their efforts on images with
50 percent. The smart-quality approach has five potential anomalies.
building blocks, all of which can generate value
(Exhibit 2). As more of the building blocks get Companies may also improve anomaly detection by
implemented, the impact compounds. using cobots—robots intended for direct interaction
with humans—that automatically handle parts and
Here are representative case examples for pass them to the next operator after successful
advanced industries for the first four building blocks. inspection. McKinsey uses such hybrid systems
The fifth building block primarily applies to pharma in its Digital Capability Centers and experienced
and medtech companies. payback periods of less than six months.

Smart-quality controls Smart-quality assurance

An aluminum-coil producer achieved 95 percent One automotive OEM used advanced analytics to
accuracy in visual-anomaly detection after just refine its entire product-development pathway

Web 2021
Smart quality in advanced industries
Exhibit
Exhibit 2 of22

Smart quality drives

drives measurable impact across
measurable impact acrossall
all building
building blocks.
Building-block type and value generated by implementation

Direct sources of value Enablers

Product Smart- Smart-

Smart-quality Smart-quality
and process quality ways compliance
controls assurance
mastery of working foundation

50–100% increase 25–40% increase 20–30% increase 25–35% increase Better quality
in productivity in productivity in right-first- in development through
and speed with and speed of key time rate efficiency and collaborative
optimized testing, quality-assurance 25% decrease speed to launch quality-
paperless labs, processes in service and 10–15% decrease management-
process automation, Higher employee repair costs in warranty costs as system redesign
and shift to the satisfaction from a result of digitized Faster new-
shop floor 25–45% decrease
user-friendly quality in time needed to risk and maturity product and
processes and tools identify quality issues management change approvals
with open access to
20–90% decrease data and records
in time required to
solve issues

Smart quality in advanced industries 3

 Smart quality has enabled organizations
to reduce their total cost of quality by
up to 50 percent.

and improve the right-first-time indicator for newly potential field-quality issues.2 The company also
developed parts to more than 80 percent. The considered other leading indicators, such as those
analytics model included data from 24 different based on data from heavy-duty cycle machines,
sources on all factors—both human behavior and and expert knowledge from specific dealers to
mechanical processes—that could have caused anticipate issues earlier and acknowledge critical
delays. Information from three years of product- issues immediately after their first occurrence. The
development projects and analysis of 1,500 accelerated issue identification translated into a
individual car components allowed the company to warranty-cost reduction of about 5 percent. The
train the underlying algorithms. It then implemented company will achieve additional value by using
several improvement initiatives, including those that insights from the algorithm to prevent issues during
involved scrutinizing open issues after virtual builds, product development.
enforcing stricter rules for virtual-build-process
adherence, improving bill-of-materials management, One steel supplier reduced its technical reject rate
identifying suppliers with low risk scores, and for a high-value product from 20 percent to nearly
providing more comprehensive onboarding for new zero within four weeks. The company gathered
team members. and structured coil-process data, meter by meter,
for more than 300 variables and 13 months of
Product and process mastery production in an end-to-end approach. It then
By applying advanced analytics to field-quality applied several analytics methods, including
information, an agriculture-equipment manufacturer decision trees, random forests, and gradient-boost
reduced the time needed to identify systemic models, to identify the three process parameters in
field issues by 45 percent. The analytics engine the rolling mill and annealing furnaces that caused
calculated daily updates of several algorithms the issues. Heat maps, area charts, and correlation
involving approximately 450,000 warranty claims analyses allowed the company to visualize the
over the past five years, as well as other data, such insights gained and prevent issues from reoccurring.
as service tickets drawn up in dealer support. By
conducting Weibull analyses, the company could Smart-quality ways of working
extrapolate expected failure rates for nearly 7,000 An agriculture-equipment manufacturer digitized
critical-part numbers daily, allowing it to identify and automated cross-functional risk and maturity

2
Weibull distributions are often used to assess product reliability and model-failure rates.

4 Smart quality in advanced industries

management in product development to reduce new development, advanced analytics and digitization
product warranty costs by more than 10 percent. allow companies to improve the right-first-time
The digital dashboards automatically compiled data rate, accelerate time to market, and increase
from different sources and visualized all maturity product maturity. These changes lead to higher
KPIs, including the warranty costs forecast at the customer satisfaction and lower warranty costs
component and vehicle level, across functions. By from the start of production. In manufacturing,
using a parameterized resource model for each visual-anomaly detection enables automated
function, the company was able to assign required quality control with significantly increased
project-team capacities at the outset of each effectiveness. In addition, advanced analytics
project and plan capacity requirements for the allow fast and effective resolution of production
portfolio of product-development projects. The problems and prevent their recurrence. Finally,
company applied industry best practices and took artificial intelligence accelerates identification of
a systematic approach to software testing and field-quality issues and automated detection of
validation to help prevent code-based quality issues. their root causes along the value chain. For best
results, companies that embark on the path to
smart quality by pursuing use cases in these areas
must make a sustained commitment to change. As
As the case examples show, a smart-quality they implement the five building blocks, both the
approach can deliver multiple benefits to quality function and the organization as a whole
companies in advanced industries. In product will achieve higher levels of agility and speed.

Álvaro Carpintero is a partner in McKinsey’s Madrid office; Ulrich Huber is a senior expert in the Cologne office, where
Heiko Nick is a partner; Evgeniya Makarova is a partner in the Chicago office.

The authors wish to thank Miguel Angel Morán and Marion Duriez for their contributions to this article.

Designed by McKinsey Global Publishing

Copyright © 2021 McKinsey & Company. All rights reserved.

Smart quality in advanced industries 5